Protein Targeting, Translocation and Insertion in Escherichia Coli

Home , Magnesium transporter, Protein targeting

Protein targeting, translocation and insertion in Escherichia coli

Proteomic analysis of substrate-pathway relationships

Louise Baars

Stockholm University Doctoral thesis © Louise Baars, Stockholm 2007

ISSN 978-91-7155-481-9, pp 1-55

Printed in Sweden by Universitetsservice AB, Stockholm 2007 Distributor: Stockholm University Library

To my family

Cover illustration

A cross-section of a small part of an Escherichia coli cell. The average distribution of molecules is shown at the proper scale. The outer and the inner membranes, studded with transmembrane proteins, are shown in green. A flagellum extends upwards from the cell surface. The motion of the flagellum is driven by the large flagellar motor that crosses both membranes. The cytoplasmic area is colored blue and purple. The large purple molecules are ribosomes and the small, L-shaped maroon molecules are tRNA, and the white strands are mRNA. Enzymes are shown in blue. The nucleoid region is shown in yellow and orange, with the long DNA circle shown in yellow.

List of Publications

This thesis is based on the following publications:

I. Marani P, Wagner S, Baars L, Genevaux P, de Gier JW, Nilsson I, Casadio R, von Heijne G. New Escherichia coli outer membrane proteins identified through prediction and experimental verification. (2006) Prot Sci 15, 884-9. II. Linda Fröderberg, Edith Houben, Louise Baars, Joen Luirink and Jan- Willem L de Gier. Targeting and translocation of two lipoproteins in Es- cherichia coli via the SRP/Sec/YidC pathway. (2004) J Biol Chem 279, 31026-32. III. Baars L, Ytterberg AJ, Drew D, Wagner S, Thilo C, van Wijk KJ, de Gier JW. Defining the role of the Escherichia coli chaperone SecB using comparative proteomics. (2006) J Biol Chem 281, 10024-34. IV. Louise Baars, Samuel Wagner, David Wickström, Mirjam Klepsch, A. Jimmy Ytterberg, Klaas J. van Wijk and Jan-Willem de Gier. Global analysis of an Escherichia coli SecE depletion strain. Submitted to J Biol Chem

Reprints were made with permission from the publishers.

Additional publications:

Samuel Wagner, Louise Baars, A. Jimmy Ytterberg, Anja Klußmeier, Claudia S. Wagner, Olof Nord, Per-Åke Nygren, Klaas J. van Wijk, Jan- Willem de Gier. Consequences of membrane protein overexpression in Escherichia coli. (2007) Mol Cell Proteomics Drew D, Fröderberg L, Baars L, de Gier JW. Assembly and over expression of membrane proteins in Escherichia coli. (2003) Biochim Biophys Acta 1610, 3-10. Annika Sääf, Louise Baars, Gunnar von Heijne. The two internal repeats in the E.coli YrbG hypothetical Na+/Ca2+ transporter have opposite ori- entations in the inner membrane. (2001) J Biol Chem 276, 18905-7.

Contents

Cover illustration ...... iv

List of Publications ...... v

Abbreviations ...... viii

1. Introduction ...... 1 1.1 The biological membrane...... 2 1.2 Protein structure ...... 3 1.2.1 Globular Proteins ...... 4 1.2.2 Membrane Proteins...... 4 1.3 The model organism - Escherichia coli ...... 6 1.3.1 The cytoplasmic proteome...... 8 1.3.2 The inner membrane proteome ...... 8 1.3.3 The periplasmic proteome ...... 8 1.3.4 The outer membrane proteome ...... 9 1.3.5 The extracellular proteome ...... 9

2. Biogenesis of secretory and integral membrane proteins in E. coli...... 11 2.1 Targeting to the inner membrane...... 14 2.1.1 Trigger Factor ...... 15 2.1.2 The SecB-pathway...... 15 2.1.2 SRP-pathway ...... 18 2.2 Translocation across, and insertion into, the inner membrane ...... 21 2.2.1 The Sec-translocon...... 21 2.2.2 The SecDFYajC-complex ...... 25 2.2.3 YidC ...... 26 2.2.4 The TAT-pathway ...... 27 2.3 Maturation and sorting of lipoproteins...... 28 2.4 Insertion of β-barrel proteins into the outer membrane...... 30

3. Objectives ...... 32

4. Summary of papers...... 33 4.1 Paper I - New Escherichia coli outer membrane proteins identified through prediction and experimental verification...... 33

4.2 Paper II - Targeting and translocation of two lipoproteins in Escherichia coli via the SRP/Sec/YidC pathway...... 34 4.3 Paper III - Defining the role of the Escherichia coli chaperone SecB using comparative proteomics ...... 35 4.4 Paper IV - Effects of SecE depletion on the inner and outer membrane proteomes of E. coli...... 36

5. Concluding remarks ...... 39

6. Sammanfattning på svenska – Hur hittar proteiner rätt i en cell?...... 40

7. Acknowledgements...... 42

8. References...... 44

Abbreviations

1DE one-dimensional electrophoresis 2DE two-dimensional electrophoresis ABC ATP-binding cassette ADP adenosine diphosphate Asp aspartic acid ATP adenosine triphosphate BN blue native C-terminus carboxy terminus Cys cysteine D aspartic acid DNA deoxyribonucleic acid E. coli Escherichia coli GDP guanosine 5'-diphosphate GTP auanosine triphosphate IMP integral inner membrane protein LPS lipopolysaccharide M. jannaschii Methanococcus jannaschii mRNA messenger ribonucleic acid MS mass spectrometry N-terminus amino terminus OMP integral outer membrane protein ORF open reading frame PAGE polyacrylamide gel electrophoresis pmf proton motive force RNA ribonucleic acid RNC ribosome nascent chain complex SRP Signal recognition particle TAT twin-arginine protein transport TF Trigger factor TM transmembrane tRNA transfer ribonucleic acid

1. Introduction

Proteins are biological macromolecules with diverse cellular functions, ranging from enzymatic catalysis to the formation of cytoskeletal structures that gives shape and rigidity to the cell. Once a protein has been synthesized inside the cell, it has to be sorted to the cellular compartment(s) where it is designed to function. Miss-localization of proteins can be harmful to the cell for a number of reasons, one being the formation of toxic protein aggregates. The focus of this thesis is the cellular infrastructure that ensures the efficient sorting of proteins to their final destination. Central questions are how proteins destined for a certain compartment are recognized, and what cellular components are involved in the translocation of proteins across, or insertion into, biological membranes. These questions have been addressed using the Gram-negative bacterium Escherichia coli as a model organism. Importantly, several of the underlying principles that govern protein sorting are evolutionary conserved and discoveries made in a unicellular organism such as E. coli bacteria are often applicable to the specialized cells of multicellular organisms and vice versa.

1 1.1 The biological membrane

Biological membranes are thin, fluid structures that form the boundary between the exterior and interior of the cell, as well as between different cellular compartments. Biological membranes consist of a mixture of lipids and proteins. Membrane lipids assemble into bilayers, illustrated in Figure 1a. A bilayer functions as a permeability barrier that prevents leakage and enables different chemical environments to exist on each side of the membrane. This is a fundamental aspect of life since many biological processes depend on the formation of concentration gradients across a membrane. However, for a cell to survive, it is essential that membranes permit passage of various molecules such as nutrients and ions. Proteins are therefore essential components of the biological membrane since they can function as pores, channels or transporters that allow selective passage across the lipid bilayer (Figure 1a).

Figure 1. The Biological membrane. (a) Schematic representation of a biological membrane consisting of lipids (light grey) and membrane proteins (dark grey) [1]. (b) Chemical structure of a membrane lipid, phosphatidyletanolamin.

Membrane lipids are amphipathic molecules consisting of a hydrophilic head group and a hydrophobic tail (Figure 1b). The lipid tail region consists of two acyl carbon chains that are unable to form hydrogen bonds. To avoid disrupting the hydrogen bonding network that exists between water molecules, membrane lipids therefore self-assemble into a bilayer where the lipid tails point towards the interior, while the head groups face the aqueous exterior on each side of the bilayer. In this way, a 20-30 Å thick hydrophobic core is created in the middle of the membrane, boarded on each side by a 15 Å thick hydrophilic interfacial region [2]. This organization effectively

2 minimizes the energetic cost for lipids to exist in a polar, aqueous environment. The hydrophobic core region makes the bilayer impermeable by blocking passage of polar molecules. Membrane lipids have different properties depending on the chemical composition of their head and tail regions. The head groups vary in size and polarity; they can be neutral, zwitterionic or negatively charged. The acyl chains of the lipid tails are typically 16-18 carbons long and differ in the degree of saturation. Double bonds reduce the flexibility of the hydrocarbon tail and give the lipid a cone shaped conformation, whereas fully saturated lipids are cylindrical. The proportion of different lipids in a bilayer determines the physical properties of the membrane such as thickness, curvature, strength and fluidity [3].

In 1972, Singer and Nicolson presented the fluid mosaic model in which proteins were viewed as icebergs floating freely in a sea of lipids [4]. However, the lateral movement of both lipids and proteins can be restricted. Today, most membranes are believed to contain patches that have a different lipid and protein composition than the surrounding membrane. These patches can vary in thickness and fluidity. One function of membrane patches is to locally increase the concentration of proteins (and lipids) involved in the same biological process [1]. Furthermore, the fluid mosaic model and most text book pictures of biological membranes greatly exaggerate the area occupied by lipids and underestimates the areas occupied by protein [1]. In addition, lipids do not only function as membrane protein solvents, but can sometimes interact specifically with membrane proteins and act as ‘co-factors’. For instance, a number of proteins involved in bioenergetics, such as NADH dehydrogenase, cytochrome bc1, ATP synthase, cytochrome oxidase, and the ADP/ATP carrier require cardiolipin to function properly [5].

1.2 Protein structure

The basic building blocks of all proteins are 20 different amino acids that contain an amino group, a carboxyl group and a side chain that varies in size and polarity. Proteins consist of polypetide chains formed by amino acids joined together by peptide bonds. These bonds are formed in a condensation reaction when the amino group from one amino acid reacts with the carboxyl group of another. The specific combination and linear arrangement of amino acids are unique for each protein and encoded in the DNA. To form a functional protein, the polypeptide chain has to “fold” into a three dimensional structure in which all atoms are organized to form energetically

3 favorable interactions, either with each other or with atoms present in its surrounding. Proteins are therefore designed to fit into a specific environment, e.g., in an aqueous environment, in a complex with other proteins or in a membrane where they interact with lipids. When a protein fails to reach the localization where it is supposed to be, it may form unproductive interactions leading to e.g., protein aggregation. To avoid this, the cell contains ‘chaperones’, proteins that prevent unproductive interactions and facilitate proper folding, but also proteases that degrade miss-folded proteins.

1.2.1 Globular Proteins

Proteins that function in an aqueous environment typically acquire a ‘globular‘ fold. A major driving force in the folding of globular proteins is the hydrophobic effect. The polypeptide folds into a conformation where hydrophobic side chains are buried in the interior, while hydrophilic amino acids typically are localized to the surface of the protein. The structure is stabilized by favorable intramolecular interactions as well as by the positive entropic effect on the surrounding water molecules.

1.2.2 Membrane Proteins

Membrane proteins are defined as ‘integral’ or ‘peripheral’ depending on their mode of interaction with the membrane. Peripheral membrane proteins are associated with the membrane via electrostatic and hydrophobic interactions with membrane lipids and/or other membrane proteins. Addition of chemicals (e.g., high salt or alkaline conditions) that disrupt electrostatic interactions leads to removal of most peripheral proteins from the membrane. Proteins can also be anchored to the membrane by covalently attached hydrophobic moieties. One example is bacterial lipoproteins that are modified with fatty acids that insert into the lipid bilayer. In contrast, integral membrane proteins consist of one or more segments that are fully embedded in the membrane. The folding environment of these transmembrane segments is fundamentally different from the aqueous environment outside the membrane. For a transmembrane segment to exist in the hydrophobic core of the lipid bilayer, the polypeptide chain has to fold into a structure that stabilizes the amino acid side chains and the polar amide and carbonyl groups of the peptide bond. Two different types of integral membrane proteins are known; β-barrel membrane proteins and α-helical membrane proteins.

4 Figure 2. Membrane proteins. Ribbon diagram of (a) a β-barrel membrane protein (PDB 1PRN) and (b) an α-helical membrane protein (PDB 2BRD). The yellow lines indicate the boarders of the membrane.

So far, β-barrel membrane proteins have only been found in the outer membrane of Gram-negative bacteria and mitochondria, and in the outer envelope of chloroplasts. β-barrel membrane proteins are composed of an even number of anti-parallel β-strands arranged in a barrel like structure. The amide and carbonyl groups of the polypeptide backbone are stabilized by hydrogen bonds between neighboring strands. Amino acid side chains of mixed polarity point towards the center of the barrel, while amino acids with apolar side chains project into the lipid bilayer on the outside surface. β- barrel membrane proteins often function as aqueous pores that allow passage of small solutes.

α-helical membrane proteins represent a conserved structural fold that is found in all organisms. In this class of proteins, the transmembrane (TM) domain folds into an α-helix where the amide and carbonyl groups of the polypeptide backbone are stabilized by internal hydrogen bonds. The TM segments are enriched in hydrophobic side chains that can interact favorably with the acyl carbons chains of the lipids. Aromatic residues that can interact with the lipid head groups are enriched at the end of the TM segment [6]. The length of each TM domain is limited by the thickness of the membrane. Typically, a membrane spanning segment consists of approximately 20 amino acids arranged in an α-helix that is perpendicular to the membrane. However, the TM segments can also be tilted, which allows longer α-helices to fit into the membrane. Based on studies of membrane proteins with known structures, it has been estimated that the average number of amino acids in TM α-helices is 26 [7]. Furthermore, an increasing number of high resolution structures has revealed the existence of atypical α-helices that vary greatly in length, tilt angle and position in the membrane [8].

5 Single spanning α-helical membrane proteins have of one TM domain, while multispanning α-helical membrane proteins have of two or more TM domains connected together by hydrophilic loop regions. The loop regions can be short linkers consisting of just a few amino acids or they can consist of a long segment, which folds into globular domain in the aqueous environment outside the membrane. A topological model of a membrane protein is a two-dimensional representation of the protein, showing the number and position of TM domains and loop regions and their orientation relative to the membrane. The simplest way to predict TM domains is to search for 20 amino acid long segments with an overall hydrophobic character. Theoretical and experimental studies have revealed that the orientation of TM domains is determined by the relative distribution of positively charged amino acids in the regions flanking them [9, 10]. In all organisms, the TM flanking regions that are localized on the cytoplasmic side of the membrane are enriched in positively charged residues [11]. This observation, referred to as ‘the positive inside rule’, can be used to predict the orientation of membrane proteins TM domains.

1.3 The model organism - Escherichia coli

The Gram-negative bacterium E. coli is by far the most well studied organism on Earth, and the E. coli genome was one of the first to be completely sequenced. The chromosome consists of a single, circular DNA containing approximately 4.6 million base pairs, harboring around four thousand open reading frames (ORFs) [12]. The chromosome is stored in the cytoplasm, an aqueous compartment encapsulated by the inner membrane (Figure 3b). The E. coli inner membrane is composed of phospholipids (70– 80% phosphatidyl-ethanolamine, 15–20% phosphatidylglycerol and <5% cardiolipin) and membrane proteins [13]. Gram-negative bacteria such as E. coli also have an outer membrane that surrounds the inner membrane (Figure 3b). The composition of the outer membrane is highly asymmetric with phospholipids (enriched in saturated fatty acids and phosphatidyl- ethanolamine) in the inner leaflet, while the main constituent of the outer leaflet is lipopolysaccharide (LPS) [13]. LPS is a unique, complex molecule that is important for the barrier function of the outer membrane. The compartment between the outer membrane and the inner membrane is called the periplasm (Figure 3b). The periplasm contains the peptidoglycan layer, which serves as an extracytoplasmic cytoskeleton that contributes to cell shape and prevents cells from lysing in dilute environments. It is covalently

6 attached to the outer membrane lipoprotein Lpp. The environment surrounding a bacterium is called the extracellular space.

Each E. coli compartment contains its own ‘proteome’, a dynamic mixture of proteins regulated to fit the requirements of the cell. Since all E. coli proteins are synthesized in the cytoplasm, proteins destined for the inner membrane, the periplasm, the outer membrane, or the extracellular space, have to be sorted to their correct compartment.

Figure 3. The E. coli cell. (a) An E. coli cell visualized by electron microscopy. (b) A cross-section of an E. coli cell showing the cytoplasm, the inner membrane, the periplasm containing the peptidoglycan layer, and the outer membrane with LPS in the outer leaflet (adapted from [14]).

7 1.3.1 The cytoplasmic proteome

The E. coli cytoplasm is a crowded compartment [15]. Out of the around four thousand ORFs in the E. coli genome, approximately 70% encodes proteins that are predicted to remain in the cytoplasm [8]. The composition of the cytoplasmic proteome reflects the role of the cytoplasm as the keeper of the genome; it consists of proteins involved in replication and protection of the DNA, gene transcription, protein translation and protein sorting. Other major constituents are proteins involved in metabolic processes. Several of these proteins interact with proteins localized at the inner membrane. The cytoplasmic proteome also includes proteins involved in maintaining cellular homeostasis, e.g., molecular chaperones that prevent protein miss-folding and proteases that degrade proteins that are miss-folded or no longer needed.

1.3.2 The inner membrane proteome

The E. coli inner membrane proteome consists of α-helical integral membrane proteins (from hereafter referred to as IMPs), lipoproteins and peripheral proteins. Based on predictions, it is estimated that approximately 20% of the ORFs in the E. coli genome encode IMPs [16]. As mentioned earlier, one of the main functions of membrane proteins is to allow selective passage of e.g., ions, nutrients and polypeptides across the membrane. Thus, the inner membrane proteome contains a large variety of different channels and transporters. Several IMPs are involved in transduction of information across the membrane. These proteins are often receptors that transmit the information by changing their conformation in response to a specific stimulus, such as the interaction with a messenger molecule or a change in the pressure profile in the membrane. The inner membrane is an important site of action when it comes to bioenergetics and it contains a system of proteins involved in the conservation and utilization of energy stored in the form of a proton gradient across the membrane.

1.3.3 The periplasmic proteome

Around ten percent of the ORFs in the E. coli genome encodes periplasmic proteins [8]. A large proportion of the periplasmic proteome consists of degradative enzymes and binding proteins for amino acids, carbohydrates, ions and vitamins. The periplasm is an oxidizing environment and periplasmic proteins often contain intermolecular disulphide bonds. The oxidation of cysteins into disulphides is catalyzed by chaperons with an

8 oxido-reductase function [17]. The reducing conditions in the cytoplasm, and the absence of these chaperones, prevent folding of periplasmic proteins prior to translocation across the inner membrane.

1.3.4 The outer membrane proteome

The outer membrane proteome consists of outer membrane lipoproteins and integral outer membrane proteins (OMPs) of the β-barrel type. The E. coli genome is predicted to encode approximately 80 outer membrane lipoproteins [18] and around 100 integral OMPs [8]. However, only a fraction of these proteins has been identified in the outer membrane experimentally. In paper I, eight putative OMPs were studied and the outer membrane localization of five of them was confirmed. Proteomics studies, including paper III, have identified approximately 30 additional proteins in the outer membrane. The majority of these are ‘porins’, OMPs that function as aqueous pores and allow passive diffusion of small (<700 Da) molecules through the membrane. One example is the highly abundant outer membrane protein OmpA, a porin with low permeability that allows slow penetration of small solutes [19]. The most abundant E. coli outer membrane protein is the lipoprotein Lpp, used as a model protein in paper II of this thesis. Lpp forms a homotrimer and one of the molecules in the trimer is covalently attached to the peptidoglycan layer in the periplasm. A recent structural study has shown that the outer membrane of E. coli contains at least one integral membrane protein (Wza) that is not of the β-barrel type [20]. Instead Wza is structured as a novel transmembrane α-helical barrel.

1.3.5 The extracellular proteome

Most bacteria export proteins to the extracellular environment. These extracellular proteins can e.g, be involved in digestion of polymers for nutritional purposes, or act as toxins during infection of host cells. It is generally assumed that nonpathogenic E. coli strains cultured for laboratory purposes do not secrete proteins to the extracellular environment. This view has recently been challenged by a proteomics study. In this study, it was shown that several proteins can be precipitated from the culture media of a laboratory E. coli strain [21]. Using two-dimensional electrophoresis, 66 protein spots were visualized and 44 different proteins were identified by mass spectrometry. Several of these proteins were outer membrane proteins. Since nucleic acids could not be detected in the growth medium it was concluded that no significant cell lysis had occurred. Another study found

9 that native vesicles, containing miss-folded outer membrane and periplasmic proteins, are released from the bacterial surface upon envelope stress [22]. Thus, it is possible proteins are exported to the extracellular environment by vesicular secretion.

10 2. Biogenesis of secretory and integral membrane proteins in E. coli

Protein biogenesis refers to the birth of a polypeptide and its maturation into a functional protein. In this thesis, E. coli has been used as a model organism to study one aspect of protein biogenesis - the sorting of proteins to the correct compartment. All E. coli proteins are synthesized by ribosomes in the cytoplasm. Using the messenger RNA (mRNA) as a template, ribosomes translate the genetic information stored in the DNA into the amino acid sequence specific for a particular protein. Proteins destined for translocation across, or integration into, the inner membrane are recognized by targeting factors that directs the proteins to the correct translocation/insertion machinery at the inner membrane [23].

Most secretory proteins (periplasmic proteins, lipoproteins and OMPs) and IMPs are targeted to the Sec-translocon, a protein-conducting channel that mediates the translocation of proteins across, or inserted into, the inner membrane [24]. Targeting of most IMPs is mediated by the signal recognition particle (SRP), while the targeting of secretory proteins can be facilitated by the chaperone SecB. The Sec-translocon is composed of the IMPs SecY, SecE and SecG [25]. The peripheral subunit SecA is an ATPase that powers translocation of secretory proteins and large periplasmic domains of IMPs across the inner membrane. Several accessory proteins; SecD, SecF, YajC and YidC, interact with the Sec-translocon to facilitate different aspects of protein translocation and insertion [24]. Some IMPs have been shown to integrate into the membrane via the Sec-translocon independent YidC pathway [26]. In addition, at least one E. coli IMP appears to require neither YidC nor the Sec-translocon for insertion, suggesting the existence of unknown mechanisms for IMP insertion in E. coli [27]. A subset of proteins are translocated via the Twin-Arginine protein Transport (TAT)-pathway. In contrast to the Sec-translocon, which only translocates proteins in an unfolded state, the TAT-translocon allows passage of folded proteins and even oligomers across the inner membrane [28].

11 The goal of the work presented in this thesis has been to improve our understanding of substrate-pathway relationships in protein targeting and translocation/insertion processes. Simply put, to find out what components are involved in the targeting, translocation and insertion of a specific substrate and ultimately - to understand the basis behind the selection of a certain pathway.

13 2.1 Targeting to the inner membrane

Secretory proteins are synthesized with a ‘signal peptide’, an N-terminal extension that is proteolytically removed upon translocation [29]. Signal peptides share little sequence homology but display several common characteristics. They are approximately 20-25 amino acids long and typically consist of a short, positively charged N-terminal region (n-region), a central, hydrophobic region (h-region), and a polar C-terminal region (c-region) containing the signal peptidase cleavage site [30, 31]. The signal peptide of secretory proteins that are translocated via the TAT-pathway contain a consensus sequence of S/T-R-R-X-F-L-K (where X is any polar amino acid) [32]. The two arginine residues that have given the pathway its name are almost invariant and essential for efficient translocation. TAT signal peptides are longer than Sec signal peptides and consist on average of 38 amino acids [33]. The h-region of a TAT signal peptide is less hydrophobic than the h-region of Sec signal peptides, and the c-region contains a positively charged residue, the so-called ‘Sec-avoidance signal’ that prevents targeting of TAT signal peptides to the Sec-translocon [34].

The vast majority of IMPs in E. coli does not contain a cleavable signal peptide. Instead, the hydrophobic segment representing the first TM domain acts as a signal sequence that directs the protein to the membrane. Depending on the orientation of the TM domain in the membrane it is referred to as a ‘signal-anchor sequence’ or a ‘reverse signal-anchor sequence’. If an IMP does contain a cleavable signal sequence, it is followed by a so-called ‘stop-transfer sequence’, a hydrophobic stretch that halts translocation and gets inserted into the membrane [35].

Proteins are targeted to the inner membrane either ‘co-translationally’ or ‘post-translationally’. Co-translational targeting occurs while the nascent chain is still attached to a translating ribosome. Post-translational targeting takes place after translation is complete, or when a substantial part of the polypetide has been synthesized. As the N-terminal end of a growing polypeptide chain emerges from the ribosome, it is ‘scanned’ by the Signal Recognition Particle (SRP) and Trigger Factor (TF), which are both positioned at the ribosomal exit tunnel [36]. Signal sequences with a hydrophobicity above a certain threshold interact preferentially with the SRP and are selected for co-translational targeting to the Sec-translocon and, in some cases, to YidC [24, 26]. Less hydrophobic signal sequences bypass SRP, and are routed into the post-translational targeting pathway. Post- translational targeting may be facilitated by the ribosome associated TF and

14 the cytoplasmic chaperone SecB [36]. In E. coli, most secretory proteins follow the post-translational targeting route, while most IMPs and a few secretory proteins are co-translationally targeted to the inner membrane.

2.1.1 Trigger Factor

TF is a non-essential chaperone that associates with the ribosome and interacts with low affinity with nascent chains as they emerge from the ribosomal exit tunnel [36]. The role of TF in protein targeting is not clear, but some studies suggest that it prevents translating ribosomes from docking onto the SecY complex [37].

2.1.2 The SecB-pathway

The E. coli protein SecB is a cytoplasmic chaperone involved in the biogenesis of secretory proteins [38]. Homologues of SecB are found in Gram-negative bacteria but not in Gram-positive bacteria, Archaea, and eukaryotes. SecB can maintain newly synthesized preproteins in an unfolded, i.e., translocation-competent state. Furthermore, SecB can target preproteins to the Sec-translocon via a high affinity interaction with the peripheral translocon subunit SecA. Previous studies have shown that SecB is required for efficient translocation of Galactose binding protein, LamB, Maltose binding protein, OmpA, OmpF, OppA and PhoE, while translocation of β-lactamase, PhoA and Ribose binding protein appears to be independent of SecB [38-42]. In paper I, the SecB dependence of five putative outer membrane proteins was tested. In paper II, we investigated the SecB dependence of two lipoproteins, Lpp and BRP. In paper III, we used a proteomics approach to characterize a secB null mutant, which resulted in the identification of an additional 12 SecB substrates.

SecB has no affinity for signal sequences but interacts with multiple sites in the mature part of unfolded preproteins [43-47]. It has been shown that binding of SecB to a preprotein requires that at least 150 amino acid residues are exposed from the ribosome [48]. The SecB protein assembles into a homotetramer with a molecular mass of 69 kDa [40, 49]. The high-resolution structures of SecB from Haemophilus influenzae and E. coli demonstrate that the tetramer is organized as a dimer of dimers [39, 40]. Two long channels, one on each side of the tetramer, have been identified. Each

15 channel contains two putative peptide binding sites, one is a deep cleft surrounded by aromatic residues, while the other is a shallow, open groove with a hydrophobic surface (Figure 6). The SecB tetramer forms a stoichiometric complex with precursors that are presumably wrapped around the tetramer to access the binding sites of both channels [40, 50]. Multiple sites in the unfolded preprotein probably interact simultaneously with SecB, contributing to the high binding affinity (Kd~5-50 nM) [45].

SecB delivers preproteins to the Sec-translocon by interaction with the peripheral translocon subunit SecA. In solution, the Kd of this interaction is ~1-2 μM [51]. The binding affinity increases (Kd 10–30 nM) when SecA is bound to the Sec-translocon [52]. The binding is even tighter (Kd 10 nM) if SecB is loaded with a preprotein [52, 53]. The SecB binding site on SecA is localized primarily in the zink containing C-terminal region of SecA, although additional sites have also been suggested [54-56]. Upon binding to SecB, SecA recognizes and binds the signal sequence of the preprotein with high specificity. The subsequent dissociation of the SecA-SecB complex is coupled to the binding of ATP to SecA. Hydrolysis of ATP by SecA also provides energy for the initial insertion and subsequent translocation of the preprotein through the Sec-translocon [57-59]. It has been suggested that SecB promotes the ATPase activity of SecA directly, by acting as an intermolecular regulator of ATP hydrolysis [60].

How does SecB recognize its substrates? A peptide library screen was used to identify a SecB binding motif that is nine amino acids long and enriched in aromatic and basic residues, whereas acidic residues are disfavored [41]. This motif fits well into the proposed peptide binding site in the SecB structure [40]. However, the SecB binding motif is common in both SecB and independent secretory proteins, as well as in cytoplasmic proteins. The presence or absence of the SecB binding motif can therefore not be used to predict whether a secretory protein is targeted to the Sec-translocon by SecB. It has been suggested that formation of a complex between SecB and its substrates depends on a kinetic partitioning between the rate of substrate folding and aggregation relative to the rate of its association to SecB [38]. SecB has no affinity for native structures and, as a consequence, only binds proteins that are sufficiently slow folding. Signal sequences can sometimes affect SecB binding indirectly; by retarding folding of the precursor, thereby

16 Figure 5. Model for SecB mediated targeting of secretory proteins to the Sec- translocon. The SecB tetramer binds the preprotein post-translationally or at a late stage of translation (1). The SecB-preprotein complex interacts with SecA (2), which binds to the signal sequence of the preprotein. The preprotein is transferred from SecB to SecA (3). SecA delivers the preprotein to the Sec-translocon and drives translocation through the channel by hydrolysis of ATP (4) (adapted from [61]).

Figure 6. The peptide-binding groove in the SecB tetramer. The left panel shows the molecular surface of SecB colored according to the underlying atoms: backbone atoms, white; non-charged polar and charged side-chain atoms, blue; hydrophobic side-chain atoms, yellow. The right panel shows a ribbon drawing of SecB tetramer in the same orientation as in the left panel [48].

17 giving SecB time to associate with the unfolded polypeptide [44, 62, 63]. However, SecB binds nonnative proteins promiscuously and the discrimination between secretory and cytoplasmic proteins is instead handled by SecA.

There appears to be a substantial functional overlap between SecB and other cytoplasmic chaperones [38]. In the absence of the preferred chaperone, nonnative proteins may bind to other components of the cytoplasmic chaperone pool. The accumulation of secretory proteins in the cytoplasm, caused by mutations in the secB or secA genes or by the overproduction of export defective proteins, results in the increased synthesis of σ32-regulated proteins, such as the chaperones DnaK/DnaJ and GroEL/GroES [64, 65]. It has been shown that the DnaK/DnaJ chaperone system can facilitate the export of some SecB dependent proteins in a SecB deficient strain [64]. A recent study shows that the temperature sensitive and aggregation prone phenotypes of a strain lacking both TF and DnaK/DnaJ is suppressed by overproduction of SecB, suggesting that SecB can function as a general cytoplasmic chaperone [66]. Furthermore, SecB expression is increased in GroEL and GroES temperature sensitive mutant strains at non-permissive temperatures [67]. In contrast, the level of SecB expression is unaffected by mutations in the sec genes [67].

2.1.2 SRP-pathway

Most IMPs are co-translationally targeted to the Sec-translocon via the SRP- pathway [36]. The coupling of translation, targeting and insertion of TMs into the membrane is thought to prevent aggregation of IMPs in the cytoplasm. Some secretory proteins also appear to be SRP dependent. SRP docks onto the ribosomal component L23 located near the ribosomal exit tunnel and binds to hydrophobic signal sequences as they emerge from the tunnel [68, 69] The E. coli SRP is a ribonucleoprotein complex consisting of two essential components; the 4.5S RNA and the protein Ffh (Ffh stands for fiftyfour kDa protein homologue after its eukaryotic homologues). Structural studies of SRP suggest that signal sequence binding is mediated primarily by a deep, flexible, hydrophobic groove in Ffh [36]. The flexibility of the residues lining this groove probably facilitates interaction with a wide range of unrelated sequences. The 4.5S RNA probably contributes to the groove, and may help to orient signal sequences by interaction with positively charged residues present both in the regions flanking TM domains and in the n-domain of signal sequences [36].

18 The ribosome nascent-chain complex (RNC) is directed to the Sec- translocon via an interaction between SRP and its receptor FtsY [70]. FtsY exists both free in the cytoplasm and associated with the membrane where it can interact with the Sec-translocon [71, 72]. Only membrane associated FtsY promotes the release of SRP from the RNC [73] and the role of cytoplasmic FtsY is not clear. It has been suggested that FtsY may act as a receptor for ribosomes, although this proposal is highly controversial [74]. The mechanism behind the transfer of the nascent chain from the SRP-FtsY complex to the translocation channel is poorly understood. SRP and FtsY are both GTPases, and GTP binding and hydrolysis play an essential role in the regulation and directionality of SRP-mediated targeting [75]. SRP, L23 and the translocon components SecY and SecE can all be cross-linked to nascent chains at an early stage of translation [24]. It has been shown that SRP and the Sec-translocon use the same contact areas for their interaction with the ribosome, suggesting that their binding is mutually exclusive [69, 76]. SecA does not seem to be required for SRP mediated targeting to the translocon [77].

Figure 7. Model for SRP mediated targeting to the Sec-translocon. SRP recognizes the signal sequence or a hydrophobic transmembrane segment emerging from the ribosome (1). The SRP/ ribosome-nascent-chain complex interact with FtsY (2). The signal peptide is transferred from the SRP to the Sec-translocon (3). Hydrolysis of GTP dissociates SRP from FtsY, allowing SRP to recycle into the cytosol (4). The ribosome nascent chain is now docked to the Sec-translocon and the polypeptide chain continues to elongate until translation is complete (adapted from [61]).

19 Secretory proteins that normally follow the post-translational targeting pathway can be co-translationally targeted by SRP if the hydrophobicity of their signal sequences is artificially increased [78-80]. However, some native signal sequences of periplasmic proteins, like DsbA, also promote cotranslational targeting [81]. When the mature part of DsbA is fused to the signal sequence of proteins that are post-translationally targeted, secretion is severely hampered [82]. It was suggested that co-translational targeting of DsbA might have evolved because this protein is essential for proper folding of proteins in the periplasm. In addition, DsbA may be extra sensitive towards the reducing environment in the cytoplasm [82]. SecM and the autotransporter Hbp are other examples of secretory proteins that are co- translationally targeted by SRP in E. coli [83, 84]. These proteins both have unusually long, but not very hydrophobic, signal sequences. In paper II, we showed that two outer membrane lipoproteins, Lpp and BRP, are also co- translationally targeted to the Sec-translocon in an SRP dependent manner. What distinguishes the signal sequence of co- and post-translationally targeted secretory proteins? Hydrophobicity appears to be an important factor. The signal sequences of DsbA and BRP are very hydrophobic. However, it is clear that other, so far unidentified, features of the signal peptide also play a role in SRP recognition [82-84].

An average cell is estimated to contain 20-30.000 ribosomes, 300-600 Sec- translocon channels, 10.000 copies of FtsY but only 40 SRPs [85]. In addition, although both Ffh and the 4.5S RNA are essential, a substantial fraction of the IMPs appears to insert correctly when cells are strongly depleted of these components [86] (and Wickström, personal communication). Taken together, this suggests that the re-cycling of SRP is extremely rapid and/or that alternative (SRP-independent) targeting pathways for IMPs exist. For instance, ribosomes may support cotranslational targeting independently of SRP through their affinity for the Sec-translocon. Another possibility is targeting of mRNA to ribosomes associated with the Sec-translocon at the inner membrane.

20 2.2 Translocation across, and insertion into, the inner membrane

The E. coli inner membrane contains two main protein-translocation machineries; the Sec-translocon and the TAT-translocon. In addition, YidC functions as an ‘insertase’ for some IMPs.

2.2.1 The Sec-translocon

The SecB and the SRP targeting pathways converge at the Sec-translocon, which functions as a protein conducting channel in the inner membrane [87]. The E. coli Sec-translocon is formed by the IMPs SecY, SecE and SecG [24]. Although only a few model proteins have been studied, it is generally assumed that the Sec-translocon is required for translocation of most secretory proteins across, and insertion of most IMPs into, the inner membrane of E. coli. In paper IV, a proteomics approach was used to study the inner and outer proteomes of cells depleted of the essential Sec- translocon component SecE (see section 4.4).

Translocation of secretory proteins and large periplasmic domains (>60 amino acids) of IMPs through the SecYEG-channel is powered by the ATPase SecA and the proton motive force (pmf) [24, 57, 88]. One round of ATP binding and hydrolysis by SecA is estimated to drive 20-30 amino acids through the membrane [57]. SecA exists both free in the cytoplasm and at the membrane where it binds with high affinity to the SecYEG translocon [89]. It has recently been shown that SecA is involved in insertion of single spanning, Sec-dependent IMPs that lack large periplasmic domains [90].

While SecG is dispensable, SecY, SecE and SecA are all essential for viability [91]. It has been shown that purified SecY, SecE and SecA reconstituted into liposomes are sufficient to promote protein translocation in vitro [92, 93]. The translation of secA is regulated by the periplasmic secretion monitor SecM in response to the translocation status of the cell [94]. The secM gene is located upstream of secA in the secM-secA operon. secM translation is subjected to a transient elongation arrest that is prolonged when the export efficiency of nacent SecM is reduced. The translation arrest disrupts the hairpin structure of the ribosome bound secM-secA mRNA. This leads to exposure of the Shine-Delgarno sequence for secA and consequently, translation of secA is initiated [94]. Little is known about the regulation of SecY, SecE and SecG expression.

Structure of the SecYEG heterotrimer The 3.2 Å resolution X-ray structure of the Sec-translocon from the Archaea Methanococcus jannaschii suggests that the protein conducting channel is formed by a single SecYEβ heterotrimer (equivalent to SecYEG in E. coli) [25]. The structure demonstrates that the ten transmembrane segments of SecY are divided in two halves, one half consists of TM1-5 and the other half of TM6-10, arranged in a clamshell like conformation. A channel is located between the two halves and a lateral gate towards the lipid bilayer is formed at the front of the molecule. The signal sequence is proposed to interact with TM2 and TM7 at the gate [88]. It has been suggested that the signal sequence stays in this position during post-translational translocation, thereby sealing the channel towards the bilayer. In contrast, during the cotranslational insertion of IMPs, continuous opening and closure (i.e., ‘breathing’) of the lateral gate would expose hydrophobic signals/TM domains to the bilayer and allow them to equilibrate into the lipid phase [25, 95]. The tendency of TM domains to partition into the membrane can be predicted from a recently developed biological hydrophobicity scale [96]. The two halves of SecY are clamped together at the back by SecE [25]. It has been shown that SecY is rapidly degraded by the integral membrane protease FtsH in E. coli cells depleted of SecE [97]. The SecE ‘clamp’ is formed by a tilted TM domain and an amphipathic α-helix that lies along the membrane interface. In most species, including M. jannaschii, SecE is a single spanning membrane protein with the N-terminus located in the cytosol. In E. coli, SecE has two additional N-terminal TM domains that are not essential for its function [98]. In the M. jannaschii structure, the α- subunit is located at the periphery of the complex [25]. This subunit shares no obvious sequence similarities with E. coli SecG.

Viewed from the side, SecY creates an hourglass shaped funnel with a central constriction formed by six hydrophobic amino acid residues pointing towards the interior. This constriction in the middle of the pore is thought to act as a seal that prevents leakage of ions and small molecules [25, 99]. In addition, a short α-helix (TM2A) blocks the channel on the extracytoplasmic and side functions as a ‘plug’ that stabilizes the channel in the closed conformation [25, 100]. It has been shown that this plug-domain moves towards the back of the complex during translocation [25, 100]. Plug displacement may be triggered by binding of a signal sequence in the pocket formed by TM2b and TM7 [25]. A gating mechanism based on movement of the plug is supported by cross-linking experiments where cysteines

22 introduced into the plug domain and the TM domain of SecE form disulphide bonds in vivo [101]. Permanently fixing the channel in this ‘open’ conformation leads to a massive flux of ions and water through the channel and is lethal to the cell [99, 101]. Furthermore, combinations of different mutations in the plug-domain and SecE yield synthetic lethal phenotypes [102].

Figure 8. The structure of the Sec-translocon from M. jannaschii solved by van den Berg et. al., 2004. Coloring: the α subunit (SecY) TM1–5 in green and TM6–10 in red; the γ subunit (SecE) in blue; the β subunit in yellow. The ‘plug’ (TM2A) and the gate helices (TM2/7) are indicated. (a) Front view of the translocon in the plane of the membrane. (b) Top view of the translocon. The clam-shaped arrangement of TM1-5 and TM6-10 of the γ subunit is indicated by a red line in the structure [103]

The hydrophobic pore ring in the middle of the channel is too narrow to allow passage of even an unfolded polypeptide, suggesting that the channel has to undergo a conformational change to allow translocation [88]. Based on constraints derived from the X-ray structure, the maximum estimated diameter of the SecY pore in an open conformation is 20x15 Å. This is probably too narrow to fit multiple TM domains. It was therefore suggested that TM domains can not be stored in the channel but rather partition into the membrane one by one [88]. A larger translocation channel with a diameter of 40-60 Å has been suggested based on permeability experiments with fluorescent probes [104]. However, it has been argued that this is an overestimate caused by “snorkeling” of the fluorescent probe and its linker region [76]. A systematic cross-linking scan suggests that the six TMs segments of Aquaporin-4 contact Sec61α (the eukaryotic homologue of SecY) in a strict N- to C-terminal order and that each TM segment leaves the

23 translocon as the next TM segment enters. However, some of the TM domains exhibited distinct secondary cross-linking patterns suggesting that they reside within the channel and/or adjacent to the translocon(s) [105].

Oligomeric state of the SecYEG complex and SecA Several studies using a variety of different experimental techniques have shown that SecYEG heterotrimers can form higher oligomeric structures both in the membrane and in detergent solution (reviewed in [89]). Dimeric, trimeric and tetrameric assemblies of the SecYEG-complex have been reported and some studies suggest dynamic oligomeric arrangements [89]. One possible function for oligomerization of the SecYEG-complex could be to provide binding sites for other factors required for translocation and insertion. It was recently demonstrated that SecA powers the translocation of a polypeptide chain through a single SecYEG hetrotrimer. However, the translocation involved oligomers of the heterotrimer, and it was suggested that SecA interacts through one of its domains with a non-translocating SecYEG complex while it moves the polypeptide chain through a neighboring SecYEG complex [95]. A structure of the E. coli translocon in complex with a translating ribosome has been solved by single-particle cryo- EM [76]. When the SecYEβ-complex from M. jannaschii was modeled into this cryo-EM structure, a dimer in a ‘front-to-front’ conformation gave the best fit. It is conceivable that two SecYEG heterotrimers in a front-to-front arrangement could form a larger channel and cooperate in translocation/insertion [76]. However, the 15 Å resolution of this cryo-EM structure is too low to unambiguously assign individual TM domains in the electron density map. A crystal structure of the E. coli SecYEG-complex determined by EM to a resolution of 8 Å indicates that heterotrimers form dimers in a ‘back-to-back’ arrangement [106]. Modeling of the SecYEβ- complex from M. jannaschii into this E. coli SecYEG structure shows that the TM segments in these two structures are arranged in nearly identical ways [25, 107].

SecA is in a dimer-monomer equilibrium influenced by e.g., ionic strength, temperature and by the presence of translocating ligands and phospholipids [108-110]. Most evidence suggests that SecA is predominantly dimeric in the cytoplasm [89]. Interestingly, comparison of the different SecA crystal structures indicates that several dimeric associations are possible [111-113]. It is a matter of intense debate whether SecA functions as a monomer or a dimer at the membrane. Several studies suggest that SecA dissociates into monomers in the presence of negatively charged lipids alone [109, 110].

24 Furthermore, it has been shown that cross-linking of the SecA dimer can not be obtained in the presence of liposomes containing the SecYEG-complex [109]. It was recently demonstrated that nanodiscs containing one SecYEG heterotrimer embedded in phospholipids promote monomerization of SecA dimers [114]. It is not clear if nanodiscs containing oligomeric SecYEG would have the same effect. In contrast, fluorescence resonance energy transfer experiments indicate that SecA retains its dimeric structure during translocation [115]. This is supported by experiments showing that artificial SecA dimers constructed by cross-linking or by fusing two copies of the SecA gene are functional [116-118]. However, another study demonstrated that a dimer formed by covalently linked SecA is inactive [119].

Protease accessibility experiments suggested that parts of SecA insert deeply into the SecYEG-complex during translocation, reaching the other side of the membrane [58]. This model is at odds with the recent structure of the Sec-translocon since the channel formed by a single hetrotrimer would be too small to accommodate these parts of SecA [88]. One controversial suggestion is that SecA oligomers can insert deeply into the membrane and form an alternative translocation channel independently of the SecYEG- complex. This hypothesis is based on the observation that SecA oligomers in lipid layers form ring-like structures that undergo a structural rearrangement upon exposure to SecB [120].

2.2.2 The SecDFYajC-complex

The SecDFYajC complex can be co-purified with the SecYEG-complex [121] and YidC [122, 123]. SecD and SecF both consist of six transmembrane domains and a large periplasmic domain while YajC is a single spanning IMP with a C-terminal soluble domain located in the cytoplasm. It has been estimated that a cell contains ~30 copies of SecD and SecF [124]. Depletion of SecD and SecF affects protein secretion [125] and the biogenesis of some IMPs (e.g., [126]). However, the exact role of the SecDFYajC-complex in protein translocation/insertion is still enigmatic. It has been proposed that the SecDFYajC-complex is the physical link between the SecYEG-complex and YidC [123]. It has also been suggested that SecD is involved in the release of secretory proteins into the periplasm [127] and that the SecDFYajC complex is involved in the control of the ATPase activity of SecA [128]. However, this is most likely not the main function of the complex since Archaea have homologs of SecD/F but not of SecA [129,

25 130]. YajC is not essential for growth and the only clue to its function is the association with the SecDF complex and the fact that genes encoding these three proteins are in the same operon.

2.2.3 YidC

The essential IMP YidC is involved in the biogenesis of IMPs, either independently or in conjunction with the Sec-translocon [24]. YidC is relatively abundant (~3000 copies per cell) compared to other components involved in IMP biogenesis [131]. Homologs of YidC are found in the cytoplasmic membrane of bacteria, in the mitochondrial inner membrane (Oxa1) and in the thylakoid membrane of chloroplasts (ALB3) [132].

YidC depletion has a strong effect on the insertion of a subset of Sec- independent IMPs. Established substrates of this "YidC-only" pathway include the M13 procoat protein, Pf3 coat protein, subunit c of the F(1)F(o) ATPase, and MscL [126, 133-139]. These proteins are all relatively small and consist of only one or two TM domains. Notably, a similar pathway - the Oxa1 pathway - is operational in yeast mitochondria, which lack equivalents to the E. coli SRP, the SRP receptor, or any of the Sec-components [24].

A role for YidC in the biogenesis of Sec-dependent IMPs was first suggested based on crosslinking experiments. These experiments showed that the hydrophobic signal anchor sequence of the single spanning protein FtsQ interacts not only with the Sec-translocon, but also with YidC and lipids during co-translational insertion [140]. Furthermore, a part of YidC was found to co-purify with the Sec-translocon [140]. YidC depletion only marginally affects the insertion of Sec-dependent IMPs in vivo [24, 141]. In spite of this, several in vitro crosslinking studies suggest a general role for YidC in IMP biogenesis [24]. Collectively, these studies demonstrate that YidC interacts with TM domains of IMPs during co-translational insertion. However, the timing of the YidC interaction is different depending on which model protein is studied. The TM domain of FtsQ was shown to first interact with SecY and then move to a combined YidC/lipid environment upon elongation, suggesting that YidC is involved in the lateral transfer of TM domains from the translocon into the lipid bilayer [142, 143]. In contrast, the first TM of Leader peptidase (Lep) was crosslinked to YidC at a very early stage of translation, before crosslinks to SecY were observed, indicating that YidC may also be involved in the reception of IMPs at the membrane [144].

26 In support of this, subunit II of the cytochrome o oxidase (CyoA) requires SRP and YidC for insertion of its N-terminal domain consisting of a cleavable signal sequence, a short periplasmic domain and a transmembrane segment, while translocation of the large C-terminal periplasmic domain requires the Sec-translocon and SecA [145].

Recently, it was shown in vitro that the Sec-dependent IMP Lactose Permease (LacY) does not require YidC for insertion into the inner membrane. However, YidC is absolutely required for proper folding of LacY [146]. This suggests that YidC may function both as an ‘insertase’ and a ‘foldase’ for IMPs. In paper II, we showed that YidC is also involved in the biogenesis of two lipoproteins, Lpp and BRP [147]. Thus, the role of YidC may not be restricted to the biogenesis of IMPs.

2.2.4 The TAT-pathway

Proteins are exported via the Sec-pathway in an unfolded state. In contrast, the Twin-Arginine protein Transport (TAT)-pathway is used for post- translational translocation of folded proteins across the inner membrane of E. coli [28]. The main components of the TAT-translocon in E. coli are the IMPs TatA, TatB, and TatC. Homologs of these proteins exist in Archaea, prokaryotes, chloroplasts and plant mitochondria [148]. Biochemical studies suggest that TatA, TatB and TatC participate in dynamic, multimeric membrane complexes that interact transiently with each other and the substrate during the translocation cycle [149]. The TatBC complex appears to be involved in the initial recognition and binding of the TAT signal peptide. Oligomeric TatA, which is thought to form the protein-conducting channel, is subsequently recruited in a process that is pmf dependent [149]. There is some evidence suggesting that the protein translocation step is also dependent on pmf [149].

In E. coli, substrates of the TAT-pathway include redox enzymes requiring cofactor insertion in the cytoplasm, oligomeric proteins that have to be assemble into a complex prior to export, and proteins whose folding is incompatible with Sec-export. Approximately 6% of all secretory proteins in E. coli are predicted to be TAT-dependent [28]. Substrates are directed to the TAT-translocon by a distinct twin-arginine motif present in the signal sequence (see section 2.1). A subset of TAT signal peptides exhibits a high degree of pathway selectivity, while others are promiscuous and can direct

27 proteins either to the Sec- or the TAT-translocon [28]. A number of TAT- dependent proteins that do not have a signal sequence of their own have been identified. These proteins form complexes with proteins that contain a TAT- signal and get secreted via the TAT-pathway by a “hitch-hiker mechanism” [150]. In addition, certain IMPs appear to depend on the TAT-translocon for insertion [149]. These IMPs have a single TM domain at their C-terminal end [151].

2.3 Maturation and sorting of lipoproteins

Lipoproteins are anchored to the periplasmic surface of the inner or outer membrane through a lipid moiety attached to a Cysteine (Cys) residue located at the N-terminus of the mature protein [18]. The E. coli genome is predicted to encode more then 90 different lipoproteins [18]. Most of these are located to the outer membrane, while only a few (<10) are localized to the inner membrane [18]. Lipoproteins are synthesized with an N-terminal signal sequence that targets them for translocation across the inner membrane by the Sec-translocon. It is generally assumed that lipoproteins are targeted to the Sec-translocon in a post-translational fashion. However, this has not been shown experimentally. In paper II, we studied the targeting and translocation of two lipoproteins, Lpp and BRP, and found that both proteins are co-translationally targeted to the Sec-translocon via the SRP- pathway [147]. Surprisingly, YidC was shown to play an important role in the biogenesis of these two lipoproteins [147]. These findings will be discussed further in section 4.2.

The signal sequence of lipoproteins ends with a so-called ‘lipobox’, a well conserved motif (-L-(A/S)-(G/A)+C-, where ‘+’ indicates the signal peptide cleavage site) required for the modification of lipoproteins in the periplasm [18]. Upon translocation across the inner membrane, the lipobox is recognized and a diacylglycerol moiety is attached to the Cys residue next to the cleavage site [18]. The signal peptide is removed by signal peptidase II (prolipoprotein signal peptidase) and the lipidated Cys residue thereby becomes the N-terminal residue. Subsequently, the Cys residue is further modified by acylation of the free α-amino group [18].

28 The sorting of lipoproteins to either the inner or the outer membrane is determined by the residue immediately after the N-terminal Cys. An Aspartic acid (Asp) in this position (position 2) retains the lipoprotein at the inner membrane while any other amino acid targets the protein for translocation to the outer membrane [18]. This transport is mediated by the ATP-binding cassette (ABC) transporter LolCDE along with the periplasmic carrier protein LolA [18]. Lipoproteins that do not have an Asp at position 2 are recognized by LolCDA, which catalyzes the release of the protein from the inner membrane [18]. The lipoprotein is subsequently delivered to LolA, which guides it across the periplasm to the lipoprotein specific receptor LolB at the outer membrane. The lipoprotein is transferred to LolB and subsequently to the outer membrane [18]. The LolCD catalyzed release of lipoproteins from the inner membrane is driven by ATP hydrolysis, while the LolA mediated transport of lipoproteins across the periplasm, their delivery to LolB and their insertion into the outer membrane does not require any input of energy [18].

Figure 9. Sorting of lipoproteins in the periplasm. After translocation across the inner membrane (IM) via the Sec-translocon (Sec), lipoproteins are modified (diacy- lation, processing by signal peptidase II and acylation) at the periplasmic side of the (IM). IM lipoproteins have an aspartic acid (D) at position 2 that functions as a Lol- avoidance signal and retains the proteins at the IM. Lipoproteins with any other amino acid (X) at position 2 are sorted to the outer membrane (OM) by the Lol system. The ABC transporter, LolCDE, use ATP to release OM lipoproteins from the inner membrane. The lipoprotein is bound by LolA, which carries the protein across the periplasmic space to the OM. At the OM, the lipoprotein is transferred to LolB, followed by incorporation into the outer membrane.

29 2.4 Insertion of β-barrel proteins into the outer membrane

β-barrel outer membrane proteins (OMPs) are synthesized with a signal peptide that directs them to the Sec-translocon. After translocation across the inner membrane, the signal peptide is cleaved off, and the OMP is released into the periplasm [152]. Several periplasmic chaperones involved in the biogenesis of OMPs have been identified, e.g., Skp, SurA, DegP and FkpA [152]. However, their precise functions and interplay are not yet known. Skp may function as a periplasmic chaperone that receives OMPs after translocation across the inner membrane since it selectively interacts with OMPs [153], and has been crosslinked to the newly translocated PhoE protein at the periplasmic side of the inner membrane [154]. Skp also binds to LPS and this binding is required for association of the Skp-OMP complex to the bilayers [14]. It has been suggested that Skp delivers OMPs to the outer membrane and that the cycling between the free form of Skp and Skp- OMP is modulated by LPS binding. However, cells lacking the skp gene are still viable and have only mild defects in OMP biogenesis [14]. Mutants in surA are viable but are affected in the folding of OMP monomers [152]. Double mutations in the skp and surA genes yield synthetic lethality, indicating that Skp and SurA either act in parallel pathways or sequentially in the same pathway [152]. DegP is a protease that degrades unfolded or misfolded proteins in the periplasm. However, DegP also has chaperone activity [155]. A combination of mutations in surA and degP also results in synthetic lethality [152]. This may be an effect of increased accumulation of misfolded OMPs in the periplasm in the absence of the proteolytic activity of DegP.

A complex consisting of the OMP YeaT and the lipoproteins YfgL, YfiO and NlpB was recently identified as an insertion-machinery for OMPs into the outer membrane of E. coli [156]. The gene encoding yfgL was identified in a genetic screen using ‘chemical conditionality’ to find suppressors of a leaky mutant (imp4213) [157]. YeaT, YfiO and NlpB were subsequently identified by co-immunoprecipitation experiments using YfgL as the bait [156]. Omp85, a homologue of YeaT in Neisseria meningitidis had previously been identified as an essential protein involved in the biogenesis of OMPs [158]. In E. coli, YeaT and YfiO are essential for viability, while YfgL and NlpB are not [14]. However, mutations in the yfgL and nlpB

30 genes cause outer membrane defects [14]. The precise function of the individual components in the OMP insertion machinery is not yet clear.

Figure 10. OMPs are translocated across the inner membrane (IM) by the Sec-translocon (Sec). In the periplasm, OMPs are transported to the outer membrane (OM) by an unknown mechanism, although the periplasmic chaperones Skp, DegP and SurA have been implicated. At the OM, the YaeT / YfgL / YfiO / NlpB complex assembles OMPs by an unknown mechanism (adapted from [14]).

31 3. Objectives

Approximately 10% of the ORFs in the E. coli genome encodes secretory proteins and 20% encodes IMPs [8]. Secretory proteins and IMPs depend on a protein sorting machineries to reach their correct cellular compartment. Using various genetic, biochemical and structural approaches much has been learned about the components involved in these processes. However, an extremely limited set of model proteins has been used to study substrate- pathway relationships. It is not clear how representative the biogenesis requirements for these model proteins are. In fact, previous work from our laboratory and others suggest that protein biogenesis is an unexpectedly versatile process and what is first viewed as an exception can very well turn out to be the general rule [135, 159]. Therefore, the main objective of all the papers presented in this thesis has been to extend our knowledge of substrate-pathway relationships by determining the biogenesis requirements of more proteins. In papers I and II, this was done using focused approaches. Selected model proteins (lipoproteins and putative OMPs) were expressed from plasmids and their targeting and translocation were analysed in vitro by crosslinking experiments and/or in vivo by pulse-chase analysis in different E. coli mutant strains. In papers III and IV, the aim was to study protein biogenesis using a ‘global’ approach. Cells depleted of SecB and SecE were compared to control cells by sub-cellular fractionation and comparative proteomics. In contrast to more focused approaches, a global approach does not depend on pre-selection of model substrates, or plasmid based ovexpression that might induce protein aggregation, saturate a pathway and/or force proteins to use a pathway that is not used under normal conditions.

32 4. Summary of papers

4.1 Paper I - New Escherichia coli outer membrane proteins identified through prediction and experimental verification

It is often difficult to predict OMPs from amino acid sequence. Compared to the transmembrane segments of α-helical IMPs, the transmembrane β- strands of OMPs are relatively short and much less hydrophobic [8]. It has been estimated that the E. coli genome encodes approximately 75 OMPs [160], but only 58 are currently listed in the EcoCyc database [8]. Recent proteomics studies (e.g., paper III) have experimentally verified the localization of several outer membrane proteins. However, not all proteins are expressed under the conditions used in these studies. Therefore, a complementary approach was used in paper I. Eight proteins predicted by the Hunter predictor as putative OMPs [160], were expressed from a plasmid and their localization was experimentally determined. Plasmid based expression can lead to formation of protein aggregates that co-sediment with outer membranes during density centrifugation. Therefore, a method based on urea extraction was developed to distinguish between proteins localized in the outer membrane and/or aggregates. Proteins that are properly integrated into the outer membrane are resistant to urea treatment, while proteins in aggregates are dissolved. Thus, upon urea treatment of the outer membrane fraction, properly integrated proteins can be pelleted by centrifugation, while aggregated proteins are found in the supernatant. Five proteins (YftM, YaiO, YfaZ, CsgF, and YliL) were shown to localize to the outer membrane, confirming the original prediction. Two proteins (YhjY and YagZ) were susceptible to urea extraction and their localization could therefore not be verified. Our data suggest that one of the proteins, YfaL, is an autotransporter.

The SecA and SecB dependencies of all proteins, except YfaL, were tested in vivo by pulse-chase analysis. All seven proteins were shown to depend on SecA for translocation. Translocation of YftM, YliL, YaiO, YhjY, YfaZ and CsgF was hampered in the absence of SecB. The degree of SecB dependence

33 varied between the different proteins. Notably, the translocation of some of the proteins tested was slow, also when SecB was expressed. This fits well with the observation that a fraction of the expressed proteins aggregate at 37ºC. Indeed, the proteins with the slowest translocation (YliL, YhjY, and YagZ) were the ones that were most sensitive to urea extraction after expression at 37ºC.

4.2 Paper II - Targeting and translocation of two lipoproteins in Escherichia coli via the SRP/Sec/YidC pathway

Lipoproteins are synthesized with an N-terminal signal sequence that targets them for translocation across the inner membrane. Although little experimental evidence exists, it is generally assumed that lipoproteins are targeted to the Sec-translocon in a post-translational fashion. In paper II, a combined in vitro crosslinking and in vivo depletion study was carried out to determine the targeting and translocation requirements of two lipoproteins - Lpp and BRP. Surprisingly, both proteins are targeted to the inner membrane via the SRP-dependent co-translational targeting pathway. Depletion of both the 4.5S RNA and Ffh impaired translocation. The signal sequences of Lpp and BRP were efficiently crosslinked to Ffh and L23/L29, ribosomal components located at the exit tunnel. In vivo depletion experiments indicated that Lpp and BRP do not require SecB for efficient translocation. However, it should be noted that the precursor form of endogenously expressed Lpp was identified in aggregates isolated from a secB null mutant strain (paper III). Thus, it is possible that a fraction of the newly synthesized Lpp can make use of both the SRP and the SecB targeting pathways.

As expected, translocation of both Lpp and BRP was dependent on the Sec- translocon and SecA. However, our data showed that YidC also plays an important role in the biogenesis of Lpp and BRP. The unprocessed form of both proteins accumulated when cells were depleted of YidC and furthermore, the signal sequences of Lpp and BRP could be crosslinked to YidC. Thus, YidC appears to be involved in the translocation of Lpp and BRP. One possibility is that YidC facilitate the lateral transfer of lipoproteins into the inner membrane via their signal sequences. To our knowledge, this is the first time that it has been shown that YidC can interact with secretory proteins and plays a role in their biogenesis.

34 4.3 Paper III - Defining the role of the Escherichia coli chaperone SecB using comparative proteomics

In E. coli, most secretory proteins are believed to cross the inner membrane via the Sec-translocon in a post-translational manner. A common assumption is that the targeting of these proteins is facilitated by the cytoplasmic chaperone SecB. However, SecB dependence has only been shown for a few selected model proteins, including the ones used in paper I. Several other secretory proteins, including the two lipoproteins studied in paper II, appear to be efficiently translocated in the absence of SecB. Although a SecB binding motif has been identified, it can not be used to predict whether SecB is required for efficient translocation since it is commonly found in both SecB dependent and independent secretory proteins, as well as in cytoplasmic proteins. In paper III, we chose a comparative proteomics approach to characterize a secB null mutant strain and identify novel substrates of the SecB-pathway. The secB null mutant strain was compared to a control strain that expresses normal levels of SecB. Whole cell lysates, protein aggregates and outer membrane fractions from these two strains were isolated and analysed using one-dimensional electrophoresis (1DE) and two- dimensional electrophoresis (2DE) in combination with mass spectrometry and immunoblotting.

Comparative 2DE analysis showed that levels of the processed forms of several secretory proteins were reduced in the secB null strain, suggesting that these proteins depend on SecB for efficient translocation. The analysis also showed that the levels of the σ32- regulated cytoplasmic chaperones DnaK, GroEL/ES, ClpB, IbpA/B, and HslU were increased in the secB null strain. This suggested that miss-targeting of secretory proteins in the absence of SecB may lead to protein aggregation in the cytoplasm. Indeed, protein aggregates containing mostly secretory proteins could be isolated from the secB null mutant, but were virtually absent in the control strain. It should be noted that the amount of protein found in aggregates corresponded to only 0.5% of the total protein content of the cell. Interestingly, pulse-chase analysis showed that cytoplasmic OmpA was removed from the aggregate fraction over time, indicating that miss-targeted OmpA is either degraded or reactivated for translocation.

The outer membrane proteome consists of OMPs and lipoproteins, which are all potential substrates of the SecB targeting pathway. To study the SecB dependence of the outer membrane proteome, we developed a protocol for

35 2DE analysis of membranes isolated from [35S]methionine-labeled cells. Proteins were visualized both by protein staining and by phosphorimaging. This allowed us to study protein steady state levels and outer membrane protein insertion kinetics on the same set of gels. The steady state levels of the iron-siderophore transporter FhuA, the ferrichrome-iron receptor FhuE, and peptidoglycan-associated lipoprotein (Pal) were decreased 2-fold in the absence of SecB. In contrast, the level of the ferrienterobactin receptor FepA was doubled, possibly compensating for the decrease of the FhuA and FhuE levels. Steady state levels of all the other outer membrane proteins identified were unchanged, suggesting that SecB is not required for their targeting to the Sec-translocon. Importantly, the analysis of [35S]methionine-labeled proteins revealed that many OMPs, such as BtuB, FhuA, FhuE, FadL, OmpT, OmpX, and TolC need more time to reach the outer membrane in the secB null mutant compared to the control. This demonstrates that SecB is not strictly required but improves the secretion efficiency of these proteins.

Collectively, the analysis of whole cell lysates, aggregates and outer membranes from the secB null mutant and the control strain lead to the identification of 19 secretory proteins that were affected in the absence of SecB. We hypothesized that these proteins are substrates of the SecB- targeting pathway. To test if this is indeed the case, the SecB dependence of 12 of these proteins was tested by classical pulse-chase analysis. Strikingly, the translocation of all 12 proteins was hampered in the secB null mutant strain.

4.4 Paper IV - Effects of SecE depletion on the inner and outer membrane proteomes of E. coli

The Sec-translocon is a protein-conducting channel involved in the translocation of secretory proteins across, and the insertion of IMPs into, the inner membrane of E coli. Sec-translocon requirements are usually studied using focused approaches and a very limited set of model proteins. To characterize the Sec-translocon dependence of secretory and IMPs in a global way, we performed a comparative sub-proteome analysis of SecE depleted and non-depleted cells. Both the steady-state proteomes and the proteome dynamics were evaluated using comparative 1DE and 2DE, followed by mass spectrometry based protein identification and extensive immunoblotting. This analysis resulted in several testable hypotheses and new substrates to further discover guiding principles for protein translocation

36 and insertion in E. coli.

Depletion of SecE resulted in a 90% (10-fold) reduction of SecE and a 50%- 70% (2-fold) reduction of the SecY,G translocon components. One- dimensional (1D) Blue-Native (BN) polyacrylamide gel-electrophoresis (PAGE) combined with immunoblotting showed that the level of the SecYEG-complex was strongly reduced upon SecE-depletion. Furthermore, the total level of SecA was increased 1.7 fold, consistent with insufficient Sec-translocon capacity. SecE depletion did not affect the SecB and SRP- targeting pathway capacity.

Analysis of whole cell lysates by 2DE and immunoblotting showed that the levels of the mature forms of several secretory proteins were reduced, while precursors accumulated in the cytoplasm upon SecE depletion. Protein aggregates containing secretory proteins were isolated from SecE depleted cells. The cytoplasmic σ32-stress response was activated in these cells, leading to increased levels of the chaperones DnaK, GroEL, GroES, ClpB and IbpA/B.

To study the effect of SecE depletion on the insertion and composition of the membrane proteomes, inner and outer membrane fractions were isolated from cells labeled with [35S]methionine. The outer membrane proteome was analysed by 2DE and the inner membrane proteome was analysed by two- dimensional BN-PAGE. The insertion and steady state levels of most outer membrane proteins were reduced upon SecE depletion. However, substantial translocation activity was still observed, indicating that translocation across the inner membrane is hampered but not abolished. The components of the outer membrane proteome were differentially affected by the depletion of SecE. OmpA, FadL, and Pal were unaffected or increased in the outer membrane upon SecE depletion, while most other OMPs and outer membrane lipoproteins were reduced to different extents.

Interestingly, the depletion has a differential effect on components of the inner membrane proteome. Steady state levels and insertion of approximately 30 IMPs were reduced, while a similar number of proteins were not affected, or even increased, in the membrane upon SecE depletion. The IMPs that were unaffected or increased lacked large translocated domains, and/or consisted of only one or two transmembrane segments. Interestingly, all established substrates of the Sec-translocon independent/YidC dependent pathway share similar features. Thus, it is

37 tempting to speculate that the proteins that are unaffected or increased in the membrane of SecE depleted cells are substrates of the YidC only pathway.

38 5. Concluding remarks

The work presented in this thesis has brought several unexpected findings and new questions that are burning to be answered. In paper II, both the SRP and YidC were found to be involved in the biogenesis of the lipoproteins Lpp and BRP. Although it is commonly assumed that secretory proteins - including lipoproteins - follow the SecB-pathway to the Sec-translocon, translocation of these two lipoproteins was not detectably affected by the absence of SecB. To date, it is not clear whether SRP and YidC have a general role in the biogenesis of lipoproteins. Therefore, targeting and translocation of more lipoproteins have to be studied. In paper III, we found that the steady state composition of the outer membrane proteome was hardly affected in the absence of SecB. However, using a combination of global and focused methods, several novel SecB-substrates were identified (papers I and III). Although not absolutely required, SecB was shown to facilitate translocation of these proteins. In paper IV, we used a proteomics approach to study the effects of SecE depletion. Given the central role of the Sec-translocon in the biogenesis of OMPs, IMPs and lipoproteins, SecE depletion was expected to have large effects on the outer and inner membrane proteomes. The outer membrane proteome was indeed strongly affected upon SecE depletion. However, the effect was differential, suggesting that some outer membrane proteins may have superior access to the Sec-translocon. The mechanism behind this observation needs to be studied further. In addition, the levels of a many IMPs were unaffected or even increased in the inner membrane of SecE depleted cells. It is tempting to speculate that these proteins are substrates of a Sec-translocon independent pathway, e.g., the YidC pathway. If this is indeed the case, Sec- translocon independent insertion of IMPs is far more common in E. coli than thought previously. Notably, papers III and IV show that proteomics approaches have great potential helping us to further our knowledge of protein targeting, translocation and insertion. Luckily for newcomers in the field, there is still plenty of work to be done!

39 6. Sammanfattning på svenska – Hur hittar proteiner rätt i en cell?

Inuti alla celler tillverkas en mängd olika proteiner med varierande funktioner. Exempel på olika typer av proteiner är enzymer, hormoner, antikroppar och muskelfibrer. När ett protein har bildats måste det sorteras och transporteras till rätt ställe i eller utanför cellen för att fungera. Det kan till och med vara skadligt för cellen om denna sortering inte fungerar eftersom proteiner till exempel kan aggregera (klumpa ihop sig) om de hamnar på fel plats. Den här avhandlingen syftar till att klargöra hur proteiner sorteras och transporteras i en cell. Bakterien E. coli har använts som modellorganism, men många av de principer som styr proteinsorteringen har bevarats under evolutionens gång. Detta innebär att upptäckter som görs i E. coli bakterier ofta kan överföras till specialiserade celler i högre organismer och vice versa.

Proteiner kan delas in i två grupper beroende på vilka lösningsegenskaper de har. Globulära proteiner är vattenlösliga medan membranproteiner inte är det. Membranproteiner är istället lösliga i den tunna film av lipider (fett) som bildar stommen i alla cellmembran. E. coli bakterien kan delas upp i fyra avdelningar som alla innehåller proteiner: cytoplasman och periplasman som innehåller globulära proteiner, och det inre och det yttre membranet som innehåller membranproteiner omgivna av lipider. Cytoplasman kapslas in av innermembranet, som i sin tur omges av yttermembranet. Periplasman finns i hålrummet mellan de båda membranen (för bild av E. coli bakteriens struktur, se Figur 3, sida 17). I cytoplasman finns bakteriens DNA som innehåller koden (ritningen) för hur alla proteiner ska se ut. När ett visst protein ska bildas så tillverkas först en sorts arbetskopia, ett så kallat mRNA, av den lilla del av DNA:t som innehåller koden för just det proteinet. Därefter tillverkas proteinet av så kallade ribosomer, som kan läsa av koden i mRNA:t. I E. coli bakterien sker allt detta i cytoplasman. Proteiner som hör hemma i periplasman eller i yttermembranet måste därför transporteras till och över innermembranet för att nå sina respektive avdelningar i cellen. Proteiner som hör hemma i innermembranet måste transporteras till

40 membranet för att sedan sättas in i det. Enligt rådande modeller sker i de flesta fall både insättning i och transport över innermembranet med hjälp av det så kallade Sec-translokonet som bildar en kanal inuti membranet. För att nybildade proteiner ska nå fram till Sec-translokonet innehåller cytoplasman olika sorteringsfaktorer. Dessa arbetar enligt samma princip som en brevbärare: de känner igen specifika egenskaper (adresslappar) hos nybildade proteiner och levererar dem till Sec-translokonet.

I den här avhandlingen har vi undersökt en av E. coli bakteriens sorterings faktorer, proteinet SecB. I artiklarna I, II och III har vi använt olika metoder för att testa vilka proteiner som behöver SecB för att nå Sec-translokonet och transporteras över membranet. Vi identifierade en rad olika proteiner i periplasman och yttermembranet som använder SecB. Resultaten i artikel III visar dock att även om proteintransporten är fördröjd i muterade E. coli celler som helt saknar SecB, så når de flesta proteiner som hör hemma i yttermembranet faktiskt fram ändå.

I artikel IV undersökte vi vilka proteiner som behöver, respektive inte behöver, Sec-translonet genom att testa vilka proteiner som påverkas i muterade E. coli celler som har mindre än en tiondel så många Sec- translokon som vanliga E. coli celler. Vi upptäckte att nivån av många proteiner i innermembranet inte påverkades, eller till och med ökade, när mängden Sec-translokon minskades. En möjlig förklaring till detta är att vissa proteiner sätts in i membranet med hjälp av något annat insättningsystem än Sec-translokonet. En annan möjlighet är att vissa proteiner har egenskaper som ger dem förtur till Sec-translokonet, och därför inte påverkas när mängden Sec-translokon i cellen minskas. Sammantaget visar resultaten att insättningen av proteiner är mer komplicerad än man tidigare trott, och att de rådande modellerna därför måste omprövas. Mycket arbete återstår för att nå full förståelse för hur proteiner sorteras och transporteras.

41 7. Acknowledgements

Jan-Willem, it all started with an email from Katmandu… Since then, we have discovered the “do’s and don’ts” of proteomics together. I think we can both agree it has been a bumpy, but rewarding journey! Thank you for your patience, spelling corrections and dedication to science. A special thanks for the colorful metaphors - those poor ants!!!

Klaas, thank you for welcoming me into your lab, for introducing me to proteomics and for stimulating and rewarding discussions. I had a great time playing with your nice machines! To all members of the KvW lab, thanks for making me feel at home, for being so helpful and nice to work with. Speciellt tack till Jimmy, det skulle inte gått utan dig!

Gunnar, thanks for introducing me to the wonderful world of membrane proteins! It has been great working in the “GvH cluster”.

Stefan N, thank you for being such an excellent boss of the DBB PhD program. To everyone at DBB, tank you for making the department such a happy place!

Linda, it was excellent working with you and I still miss you in the lab. Thanks for setting up the lab, for “breaking in” JW and most of all - for being such a friend. David D, thank you for the good times, and for helping me with the cloning for the SecB project. Samuel, I still wonder if I can’t take you with me to my next position! How will I ever manage without you?! Thank you for all the help (including life saving chocolate) and for breaking more glass-ware than me. David W, thank you for all the laughs and weird discussions in the lab. It has been a pure pleasure working with you! However, I will never do your laundry again. Mirjam, soon it will be your job to “hålla killarna på mattan”. I trained you well, don’t let me down! Thanks for all the nice chats, and for eating more chocolate then me ;-).

Joy and Dan, thank you both for organizing journal clubs etc! It has been great working with you. I especially appreciate both of you sharing your ‘senior’ experiences and opinions on science and career with me. Carolina and Filippa, thank you for keeping track of birthdays etc! Ing-Marie, thanks for taking care of so many things…freezers, gel dryers etc. Kalle S, thanks

42 for help with the SecE westerns! Kalle E, for organizing cake-clubs, thanks! Marie, coffee breaks are more fun when you are around! Mikaela, thank you for calling me organized! It has been great fun working with you. Hope to see you on a rock or a wall in the future. Mirjam L, I truly enjoy the heated discussions during coffee breaks, it is great to have you here! Susanna, for your contagious enthusiasm, thanks! Tara, we shared misery at KÖL and good times at Corsica. Thank you for being a good friend. Good luck in the new lab.

Eva Severinsson och alla i Biomedicinska Forskarskolan 99/00, tack för ett fantastiskt år! Speciellt tack till Hanna, Julian, Maria K, Maria S och Martin G, Martin L och Camilla, och Sofia för trevliga middagar och fester. Maria E, Henke och Holger (coolaste killen!), bonus-tack för alla trevliga midsommarfester på Gräskö. Maria, jag saknar våra vandrings strapatser, hoppas de blir fler!

Lisa and Andrea, my dear Italians who saved my life and sanity in not so gorgeous Ithaca! You are both (yes, you too Andrea) fantastic people and I feel blessed to have you as friends.

Jenny, Linnea och Tesan (alias brudarna babes), som tur är har vi inte bara ett glödande intresse för böcker gemensamt... Tack för en underbar vänskap och för alla roliga dagar, kvällar och nätter.

Frida och Kristina, tack för er vänskap som betyder så mycket för mig. Tack för att ni finns och för att ni har så stort tålamod. Vi måste ses oftare, puss!

Mattias, för att du säkrar mig när jag faller, och för alla berg vi kommer att bestiga tillsammans - tack. Du gör mig lycklig.

Pappa, Marie, Vincent, Leo och Adde, det blir alldeles för sällan men det är alltid kul när vi väl ses! Tack för allt stöd och för roliga studer på Öland, i skidbacken och vid middagsbordet.

Carl-Fredrik, du är min idol, det du inte kan fixa är inte värt att fixa! Tack för att du är - och alltid har varit - en sådan fantastisk storebror. Liza, tack för alla gourmé middagar och trevliga stunder. Emma, Julia och Marcus, ni är världens sötaste och jag blir så glad av att vara med er!

Mamma, du stöttar mig i vått och torrt. Stort tack för all din kärlek, uppmuntran och omsorg.

43 8. References

1. Engelman, D.M., (2005) Membranes are more mosaic than fluid. Nature 438: 578-580. 2. White, S.H., (2003) Translocons, thermodynamics, and the folding of membrane proteins. FEBS Lett 555: 116-121. 3. Mouritsen, O., Life- As a matter of fat, the emerging science of lipidomics. The frontier collection, ed. D. Dragoman, Dragoman, M., Elitzur, A.C., Silverman, M.P., Tuszymski, J., Zeh, H.D. . 2005, Heidelberg: Springer. 4. Singer, S.J. and G.L. Nicolson, (1972) The fluid mosaic model of the structure of cell membranes. Science 175: 720-731. 5. Lee, A.G., (2004) How lipids affect the activities of integral membrane proteins. Biochim Biophys Acta 1666: 62-87. 6. Killian, J.A. and G. von Heijne, (2000) How proteins adapt to a membrane-water interface. Trends Biochem Sci 25: 429-434. 7. Ulmschneider, M.B., M.S. Sansom, and A. Di Nola, (2005) Properties of integral membrane protein structures: derivation of an implicit membrane potential. Proteins 59: 252-265. 8. Elofsson, A. and G. Heijne, (2007) Membrane Protein Structure: Prediction versus Reality. Annu Rev Biochem 76: 125-140. 9. von Heijne, G., (1986) The distribution of positively charged residues in bacterial inner membrane proteins correlates with the trans-membrane topology. EMBO J 5: 3021-3027. 10. Nilsson, I. and G. von Heijne, (1990) Fine-tuning the topology of a polytopic membrane protein: role of positively and negatively charged amino acids. Cell 62: 1135-1141. 11. Nilsson, J., B. Persson, and G. von Heijne, (2005) Comparative analysis of amino acid distributions in integral membrane proteins from 107 genomes. Proteins 60: 606-616. 12. Blattner, F.R., G. Plunkett, C.A. Bloch, N.T. Perna, V. Burland, M. Riley, J. Collado-Vides, J.D. Glasner, C.K. Rode, G.F. Mayhew, J. Gregor, N.W. Davis, H.A. Kirkpatrick, M.A. Goeden, D.J. Rose, B. Mau, and Y. Shao, (1997) The complete genome sequence of Escherichia coli K-12. Science 277: 1453-1462. 13. Lugtenberg, E.J. and R. Peters, (1976) Distribution of lipids in cytoplasmic and outer membranes of Escherichia coli K12. Biochim

44 Biophys Acta 441: 38-47. 14. Ruiz, N., D. Kahne, and T.J. Silhavy, (2006) Advances in understanding bacterial outer-membrane biogenesis. Nat Rev Microbiol 4: 57-66. 15. Zimmerman, S.B. and S.O. Trach, (1991) Estimation of macromolecule concentrations and excluded volume effects for the cytoplasm of Escherichia coli. J Mol Biol 222: 599-620. 16. Krogh, A., B. Larsson, G. von Heijne, and E. Sonnhammer, (2001) Predicting transmembrane protein topology with a hidden Markov model. Application to complete genomes. J Mol Biol 305: 567-580. 17. Messens, J. and J.F. Collet, (2006) Pathways of disulfide bond formation in Escherichia coli. Int J Biochem Cell Biol 38: 1050- 1062. 18. Tokuda, H. and S. Matsuyama, (2004) Sorting of lipoproteins to the outer membrane in E. coli. Biochim Biophys Acta 1694: IN1-9. 19. Sugawara, E. and H. Nikaido, (1992) Pore-forming activity of OmpA protein of Escherichia coli. J Biol Chem 267: 2507-2511. 20. Dong, C., K. Beis, J. Nesper, A.L. Brunkan-Lamontagne, B.R. Clarke, C. Whitfield, and J.H. Naismith, (2006) Wza the translocon for E. coli capsular polysaccharides defines a new class of membrane protein. Nature 444: 226-229. 21. Nandakumar, M.P., A. Cheung, and M.R. Marten, (2006) Proteomic analysis of extracellular proteins from Escherichia coli W3110. J Proteome Res 5: 1155-1161. 22. McBroom, A.J. and M.J. Kuehn, (2007) Release of outer membrane vesicles by Gram-negative bacteria is a novel envelope stress response. Mol Microbiol 63: 545-558. 23. Blobel, G. and B. Dobberstein, (1975) Transfer of proteins across membranes. I. Presence of proteolytically processed and unprocessed nascent immunoglobulin light chains on membrane- bound ribosomes of murine myeloma. J Cell Biol 67: 835-851. 24. Luirink, J., G. von Heijne, E. Houben, and J.W. de Gier, (2005) Biogenesis of inner membrane proteins in Escherichia coli. Annu Rev Microbiol 59: 329-355. 25. Van den Berg, B., W.M. Clemons, Jr., I. Collinson, Y. Modis, E. Hartmann, S.C. Harrison, and T.A. Rapoport, (2004) X-ray structure of a protein-conducting channel. Nature 427: 36-44. 26. Kiefer, D. and A. Kuhn, (2007) YidC as an essential and multifunctional component in membrane protein assembly. Int Rev Cytol 259: 113-138. 27. Facey, S.J. and A. Kuhn, (2003) The sensor protein KdpD inserts into the Escherichia coli membrane independent of the Sec translocase and YidC. Eur J Biochem 270: 1724-1734. 28. Lee, P.A., D. Tullman-Ercek, and G. Georgiou, (2006) The bacterial twin-arginine translocation pathway. Annu Rev Microbiol 60: 373- 395. 29. Dalbey, von Heijne., Protein Targeting Transport and

45 Translocation, 2002, Elsevier Science. 30. von Heijne, G., (1983) Patterns of amino acids near signal-sequence cleavage sites. Eur J Biochem 133: 17-21. 31. von Heijne, G., (1985) Signal sequences. The limits of variation. J Mol Biol 184: 99-105. 32. Berks, B.C., (1996) A common export pathway for proteins binding complex redox cofactors? Mol Microbiol 22: 393-404. 33. Cristóbal, S., J.W. de Gier, H. Nielsen, and G. von Heijne, (1999) Competition between Sec- and TAT-dependent protein translocation in Escherichia coli. EMBO J 18: 2982-2990. 34. Bogsch, E., S. Brink, and C. Robinson, (1997) Pathway specificity for a delta pH-dependent precursor thylakoid lumen protein is governed by a 'Sec-avoidance' motif in the transfer peptide and a 'Sec-incompatible' mature protein. Embo J 16: 3851-3859. 35. von Heijne, G., (1997) Getting greasy: How transmembrane polypeptide segments integrate into the lipid bilayer. Mol Microbiol 24: 249-253. 36. Luirink, J. and I. Sinning, (2004) SRP-mediated protein targeting: structure and function revisited. Biochim Biophys Acta 1694: 17-35. 37. Lee, H.C. and H.D. Bernstein, (2002) Trigger factor retards protein export in Escherichia coli. J Biol Chem 277: 43527-43535. 38. Randall, L.L. and S.J. Hardy, (2002) SecB, one small chaperone in the complex milieu of the cell. Cell Mol Life Sci 59: 1617-1623. 39. Dekker, C., B. de Kruijff, and P. Gros, (2003) Crystal structure of SecB from Escherichia coli. J Struct Biol 144: 313-319. 40. Xu, Z., J.D. Knafels, and K. Yoshino, (2000) Crystal structure of the bacterial protein export chaperone secB. Nat Struct Biol 7: 1172- 1177. 41. Knoblauch, N.T., S. Rudiger, H.J. Schonfeld, A.J. Driessen, J. Schneider-Mergener, and B. Bukau, (1999) Substrate specificity of the SecB chaperone. J Biol Chem 274: 34219-34225. 42. Powers, E.L. and L.L. Randall, (1995) Export of periplasmic galactose-binding protein in Escherichia coli depends on the chaperone SecB. J Bacteriol 177: 1906-1907. 43. Gannon, P.M., P. Li, and C.A. Kumamoto, (1989) The mature portion of Escherichia coli maltose-binding protein (MBP) determines the dependence of MBP on SecB for export. J Bacteriol 171: 813-818. 44. Randall, L.L., T.B. Topping, and S. Hardy, (1990) No Specific Recognition of Leader Peptide by SecB, a Chaperone Involved in Protein Export. Science 248: 860-863. 45. Topping, T.B. and L.L. Randall, (1994) Determination of the binding frame within a physiological ligand for the chaperone SecB. Prot Sci 3: 730-736. 46. Khisty, V.J., G.R. Munske, and L.L. Randall, (1995) Mapping of the binding frame for the chaperone SecB within a natural ligand, galactose-binding protein. J Biol Chem 270: 25920-25927.

46 47. Smith, V.F., S.J.S. Hardy, and L.L. Randall, (1997) Determination of the binding frame of the chaperone SecB within the physiological ligand oligopeptide-binding protein. Protein Sci 6: 1746-1755. 48. Zhou, J. and Z. Xu, (2005) The structural view of bacterial translocation-specific chaperone SecB: implications for function. Mol Microbiol 58: 349-357. 49. Watanabe, M. and G. Blobel, (1989) SecB functions as a cytosolic signal recognition factor for protein export in E. coli. Cell 58: 695- 705. 50. Lecker, S., R. Lill, T. Ziegelhoffer, C. Georgopoulos, P.J. Bassford, Jr., C.A. Kumamoto, and W. Wickner, (1989) Three pure chaperone proteins of Escherichia coli -SecB, trigger factor and GroEL-form soluble complexes with precursor proteins in vitro. EMBO J 8: 2703-2709. 51. den Blaauwen, T., E. Terpetschnig, J.R. Lakowicz, and A.J. Driessen, (1997) Interaction of SecB with soluble SecA. FEBS Lett 416: 35-38. 52. Fekkes, P., C. van der Does, and A.J. Driessen, (1997) The molecular chaperone SecB is released from the carboxy-terminus of SecA during initiation of precursor protein translocation. EMBO J 16: 6105-6113. 53. Hartl, F.U., S. Lecker, E. Schiebel, J.P. Hendrick, and W. Wickner, (1990) The Binding Cascade of SecB to SecA to SecY/E Mediates Preprotein Targeting to the E. Coli Plasma Membrane. Cell 63: 269- 279. 54. Woodbury, R.L., T.B. Topping, D.L. Diamond, D. Suciu, C.A. Kumamoto, S.J. Hardy, and L.L. Randall, (2000) Complexes between protein export chaperone SecB and SecA. Evidence for separate sites on SecA providing binding energy and regulatory interactions. J Biol Chem 275: 24191-24198. 55. Randall, L.L., J.M. Crane, G. Liu, and S.J. Hardy, (2004) Sites of interaction between SecA and the chaperone SecB, two proteins involved in export. Protein Sci 13: 1124-1133. 56. Randall, L.L., J.M. Crane, A.A. Lilly, G. Liu, C. Mao, C.N. Patel, and S.J. Hardy, (2005) Asymmetric binding between SecA and SecB two symmetric proteins: implications for function in export. J Mol Biol 348: 479-489. 57. Schiebel, E., A.J. Driessen, F.U. Hartl, and W. Wickner, (1991) Delta mu H+ and ATP function at different steps of the catalytic cycle of preprotein translocase. Cell 64: 927-939. 58. Economou, A. and W. Wickner, (1994) SecA promotes preprotein translocation by undergoing ATP-driven cycles of membrane insertion and deinsertion. Cell 78: 835-843. 59. Economou, A., J.A. Pogliano, J. Beckwith, D.B. Oliver, and W. Wickner, (1995) SecA membrane cycling at SecYEG is driven by distinct ATP binding and hydrolysis events and is regulated by SecD and SecF. Cell 83: 1171-1181.

47 60. Miller, A., L. Wang, and D.A. Kendall, (2002) SecB modulates the nucleotide-bound state of SecA and stimulates ATPase activity. Biochemistry 41: 5325-5332. 61. Driessen, A.J.M., E.H. Manting, and C. van der Does, (2001) The structural basis of protein targeting and translocation in bacteria. Nat Struct Biol 8: 492-498. 62. Park, S., G. Liu, T.B. Topping, W.H. Cover, and L.L. Randall, (1988) Modulation of folding pathways of exported proteins by the leader sequence. Science 239: 1033-1035. 63. Liu, G.P., T.B. Topping, and L.L. Randall, (1989) Physiological Role During Export for the Retardation of Folding by the Leader Peptide of Maltose-Binding Protein. Proc Natl Acad Sci U S A 86: 9213-9217. 64. Wild, J., W.A. Walter, C.A. Gross, and E. Altman, (1993) Accumulation of secretory protein precursors in Escherichia coli induces the heat shock response. J Bacteriol 175: 3992-3997. 65. Ito, K., Y. Akiyama, T. Yura, and K. Shiba, (1986) Diverse effects of the MalE-LacZ hybrid protein on Escherichia coli cell physiology. J Bacteriol 167: 201-204. 66. Ullers, R.S., J. Luirink, N. Harms, F. Schwager, C. Georgopoulos, and P. Genevaux, (2004) SecB is a bona fide generalized chaperone in Escherichia coli. Proc Natl Acad Sci U S A 101: 7583-7588. 67. Muller, J.P., (1996) Influence of impaired chaperone or secretion function on SecB production in Escherichia coli. J Bacteriol 178: 6097-6104. 68. Valent, Q.A., J.W. de Gier, G. von Heijne, D.A. Kendall, C.M. ten Hagen-Jongman, B. Oudega, and J. Luirink, (1997) Nascent membrane and presecretory proteins synthesized in Escherichia coli associate with signal recognition particle and trigger factor. Mol Microbiol 25: 53-64. 69. Schaffitzel, C., M. Oswald, I. Berger, T. Ishikawa, J.P. Abrahams, H.K. Koerten, R.I. Koning, and N. Ban, (2006) Structure of the Escherichia coli signal recognition particle bound to a translating ribosome. Nature 444: 503-506. 70. (Gill and Salmond, B., R., E. Crooke, K. Shiba, W. Wickner, and K. Ito, (1986) The secY protein can act post-translationally to promote bacterial protein export. J Biol Chem 261: 12907-12910. 71. Luirink, J., C.M. ten Hagen-Jongman, C.C. van der Weijden, B. Oudega, S. High, B. Dobberstein, and R. Kusters, (1994) An alternative protein targeting pathway in Escherichia coli: studies on the role of FtsY. EMBO J 13: 2289-2296. 72. Angelini, S., S. Deitermann, and H.G. Koch, (2005) FtsY, the bacterial signal-recognition particle receptor, interacts functionally and physically with the SecYEG translocon. EMBO Rep 6: 476-481. 73. Valent, Q.A., P.A. Scotti, S. High, J.W. de Gier, G. von Heijne, G. Lentzen, W. Wintermeyer, B. Oudega, and J. Luirink, (1998) The Escherichia coli SRP and SecB targeting pathways converge at the

48 translocon. EMBO J 17: 2504-2512. 74. Herskovits, A.A. and E. Bibi, (2000) Association of Escherichia coli ribosomes with the inner membrane requires the signal recognition particle receptor but is independent of the signal recognition particle. Proc Natl Acad Sci U S A 97: 4621-4626. 75. Shan, S.O. and P. Walter, (2005) Co-translational protein targeting by the signal recognition particle. FEBS Lett 579: 921-926. 76. Mitra, K., C. Schaffitzel, T. Shaikh, F. Tama, S. Jenni, C.L. Brooks, 3rd, N. Ban, and J. Frank, (2005) Structure of the E. coli protein- conducting channel bound to a translating ribosome. Nature 438: 318-324. 77. Scotti, P.A., Q.A. Valent, E.H. Manting, M.L. Urbanus, A.J. Driessen, B. Oudega, and J. Luirink, (1999) SecA is not required for signal recognition particle-mediated targeting and initial membrane insertion of a nascent inner membrane protein. J Biol Chem 274: 29883-29888. 78. Bowers, C.W., F. Lau, and T.J. Silhavy, (2003) Secretion of LamB- LacZ by the signal recognition particle pathway of Escherichia coli. J Bacteriol 185: 5697-5705. 79. Kim, J., S. Rusch, J. Luirink, and D.A. Kendall, (2001) Is Ffh required for export of secretory proteins? FEBS Lett 505: 245-248. 80. Lee, H.C. and H.D. Bernstein, (2001) The targeting pathway of Escherichia coli presecretory and integral membrane proteins is specified by the hydrophobicity of the targeting signal. Proc Natl Acad Sci U S A 98: 3471-3476. 81. Schierle, C.F., M. Berkmen, D. Huber, C. Kumamoto, D. Boyd, and J. Beckwith, (2003) The DsbA signal sequence directs efficient, cotranslational export of passenger proteins to the Escherichia coli periplasm via the signal recognition particle pathway. J Bacteriol 185: 5706-5713. 82. Huber, D., D. Boyd, Y. Xia, M.H. Olma, M. Gerstein, and J. Beckwith, (2005) Use of thioredoxin as a reporter to identify a subset of Escherichia coli signal sequences that promote signal recognition particle-dependent translocation. J Bacteriol 187: 2983- 2991. 83. Nakatogawa, H. and K. Ito, (2001) Secretion monitor, SecM, undergoes self-translation arrest in the cytosol. Mol Cell 7: 185-192. 84. Sijbrandi, R., M.L. Urbanus, C.M. ten Hagen-Jongman, H.D. Bernstein, B. Oudega, B.R. Otto, and J. Luirink, (2003) Signal recognition particle (SRP)-mediated targeting and sec- dependent translocation of an extracellular Escherichia coli protein. J Biol Chem 278: 4654-4659. 85. Drew, D., L. Fröderberg, L. Baars, and J.W. de Gier, (2003) Assembly and overexpression of membrane proteins in Escherichia coli. Biochim Biophys Acta 1610: 3-10. 86. Newitt, J.A., N.D. Ulbrandt, and H.D. Bernstein, (1999) The structure of multiple polypeptide domains determines the signal

49 recognition particle targeting requirement of Escherichia coli inner membrane proteins. J Bacteriol 181: 4561-4567. 87. Valent, Q.A., P.A. Scotti, S. High, J.W. de Gier, G. von Heijne, G. Lentzen, W. Wintermeyer, B. Oudega, and J. Luirink, (1998) The Escherichia coli SRP and SecB targeting pathways converge at the translocon. EMBO J 17: 2504-2512. 88. Osborne, A.R., T.A. Rapoport, and B. van den Berg, (2005) Protein translocation by the Sec61/SecY channel. Annu Rev Cell Dev Biol 21: 529-550. 89. Rusch, S.L. and D.A. Kendall, (2007) Oligomeric states of the SecA and SecYEG core components of the bacterial Sec translocon. Biochim Biophys Acta 1768: 5-12. 90. Deitermann, S., G.S. Sprie, and H.G. Koch, (2005) A dual function for SecA in the assembly of single spanning membrane proteins in Escherichia coli. J Biol Chem 280: 39077-39085. 91. Manting, E.H. and A.J. Driessen, (2000) Escherichia coli translocase: the unravelling of a molecular machine. Mol Microbiol 37: 226-238. 92. Brundage, L., J.P. Hendrick, E. Schiebel, A.J. Driessen, and W. Wickner, (1990) The purified E. coli integral membrane protein SecY/E is sufficient for reconstitution of SecA-dependent precursor protein translocation. Cell 62: 649-657. 93. Akimaru, J., S.I. Matsuyama, H. Tokuda, and S. Mizushima, (1991) Reconstitution of a Protein Translocation System Containing Purified SecY, SecE, and SecA from Escherichia coli. Proc Natl Acad Sci USA 88: 6545-6549. 94. Nakatogawa, H. and K. Ito, (2001) Secretion monitor, SecM, undergoes self-translation arrest in the cytosol. Mol Cell 7: 185-192. 95. Osborne, A.R. and T.A. Rapoport, (2007) Protein translocation is mediated by oligomers of the SecY complex with one SecY copy forming the channel. Cell 129: 97-110. 96. Hessa, T., H. Kim, K. Bihlmaier, C. Lundin, J. Boekel, H. Andersson, I. Nilsson, S.H. White, and G. von Heijne, (2005) Recognition of transmembrane helices by the endoplasmic reticulum translocon. Nature 433: 377-381. 97. Kihara, A., Y. Akiyama, and K. Ito, (1995) FtsH is required for proteolytic elimination of uncomplexed forms of SecY, an essential protein translocase subunit. Proc Natl Acad Sci USA 92: 4532-4536. 98. Schatz, P.J., K.L. Bieker, K.M. Ottemann, T.J. Silhavy, and J. Beckwith, (1991) One of three transmembrane stretches is sufficient for the functioning of the SecE protein, a membrane component of the E. coli secretion machinery. EMBO J 10: 1749-1757. 99. Saparov, S.M., K. Erlandson, K. Cannon, J. Schaletzky, S. Schulman, T.A. Rapoport, and P. Pohl, (2007) Determining the conductance of the SecY protein translocation channel for small molecules. Mol Cell 26: 501-509. 100. Li, W., S. Schulman, D. Boyd, K. Erlandson, J. Beckwith, and T.A.

50 Rapoport, (2007) The plug domain of the SecY protein stabilizes the closed state of the translocation channel and maintains a membrane seal. Mol Cell 26: 511-521. 101. Harris, C.R. and T.J. Silhavy, (1999) Mapping an interface of SecY (PrlA) and SecE (PrlG) by using synthetic phenotypes and in vivo cross-linking. J Bacteriol 181: 3438-3444. 102. Flower, A.M., (2007) The SecY translocation complex: convergence of genetics and structure. Trends Microbiol 15: 203-210. 103. White, S.H. and G. von Heijne, (2004) The machinery of membrane protein assembly. Curr Opin Struct Biol 14: 397-404. 104. Hamman, B., J.-C. Chen, E. Johnson, and A. Johnson, (1997) The aqueous pore through the translocon has a diameter of 40-60 Å during cotrsnaltional protein translocation at the ER membrane. Cell 89: 535-544. 105. Sadlish, H., D. Pitonzo, A.E. Johnson, and W.R. Skach, (2005) Sequential triage of transmembrane segments by Sec61alpha during biogenesis of a native multispanning membrane protein. Nat Struct Mol Biol 12: 870-878. 106. Breyton, C., W. Haase, T.A. Rapoport, W. Kühlbrandt, and I. Collinson, (2002) Three-dimensional structure of the bacterial protein-translocation complex SecYEG. Nature 418: 662-665. 107. Bostina, M., B. Mohsin, W. Kuhlbrandt, and I. Collinson, (2005) Atomic model of the E. coli membrane-bound protein translocation complex SecYEG. J Mol Biol 352: 1035-1043. 108. Woodbury, R.L., S.J. Hardy, and L.L. Randall, (2002) Complex behavior in solution of homodimeric SecA. Protein Sci 11: 875-882. 109. Or, E., A. Navon, and T. Rapoport, (2002) Dissociation of the dimeric SecA ATPase during protein translocation across the bacterial membrane. EMBO J 21: 4470-4479. 110. Benach, J., Y.T. Chou, J.J. Fak, A. Itkin, D.D. Nicolae, P.C. Smith, G. Wittrock, D.L. Floyd, C.M. Golsaz, L.M. Gierasch, and J.F. Hunt, (2003) Phospholipid-induced monomerization and signal- peptide-induced oligomerization of SecA. J Biol Chem 278: 3628- 3638. 111. Hunt, J.F., S. Weinkauf, L. Henry, J.J. Fak, P. McNicholas, D.B. Oliver, and J. Deisenhofer, (2002) Nucleotide control of interdomain interactions in the conformational reaction cycle of SecA. Science 297: 2018-2026. 112. Vassylyev, D.G., H. Mori, M.N. Vassylyeva, T. Tsukazaki, Y. Kimura, T.H. Tahirov, and K. Ito, (2006) Crystal structure of the translocation ATPase SecA from Thermus thermophilus reveals a parallel, head-to-head dimer. J Mol Biol 364: 248-258. 113. Zimmer, J., W. Li, and T.A. Rapoport, (2006) A novel dimer interface and conformational changes revealed by an X-ray structure of B. subtilis SecA. J Mol Biol 364: 259-265. 114. Alami, M., K. Dalal, B. Lelj-Garolla, S.G. Sligar, and F. Duong, (2007) Nanodiscs unravel the interaction between the SecYEG

51 channel and its cytosolic partner SecA. EMBO J 26: 1995-2004. 115. Driessen, A.J., (1993) SecA, the peripheral subunit of the Escherichia coli precursor protein translocase, is functional as a dimer. Biochemistry 32: 13190-13197. 116. de Keyzer, J., E.O. van der Sluis, R.E. Spelbrink, N. Nijstad, B. de Kruijff, N. Nouwen, C. van der Does, and A.J. Driessen, (2005) Covalently dimerized SecA is functional in protein translocation. J Biol Chem 280: 35255-35260. 117. Jilaveanu, L.B. and D. Oliver, (2006) SecA dimer cross-linked at its subunit interface is functional for protein translocation. J Bacteriol 188: 335-338. 118. McFarland, L., O. Francetic, and C.A. Kumamoto, (1993) A mutation of Escherichia coli SecA protein that partially compensates for the absence of SecB. J Bacteriol 175: 2255-2262. 119. Or, E. and T. Rapoport, (2007) Cross-linked SecA dimers are not functional in protein translocation. FEBS Lett 581: 2616-2620. 120. Chen, Y., P.C. Tai, and S.F. Sui, (2007) The active ring-like structure of SecA revealed by electron crystallography: Conformational change upon interaction with SecB. J Struct Biol 159: 149-153. 121. Duong, F. and W. Wickner, (1997) Distinct catalytic roles of the SecYE, SecG and SecDFyajC subunits of preprotein translocase holoenzyme. EMBO J 16: 2756-2768. 122. Scotti, P.A., M.L. Urbanus, J. Brunner, J.W. de Gier, G. von Heijne, C. van der Does, A.J. Driessen, B. Oudega, and J. Luirink, (2000) YidC, the Escherichia coli homologue of mitochondrial Oxa1p, is a component of the Sec translocase. Embo J 19: 542-549. 123. Nouwen, N. and A.J. Driessen, (2002) SecDFyajC forms a heterotetrameric complex with YidC. Mol Microbiol 44: 1397-1405. 124. Matsuyama, S., Y. Fujita, K. Sagara, and S. Mizushima, (1992) Overproduction, purification and characterization of SecD and SecF, integral membrane components of the protein translocation machinery of Escherichia coli. Biochim Biophys Acta 1222: 77-84. 125. Pogliano, J.A. and J. Beckwith, (1994) SecD and SecF facilitate protein export in Escherichai coli. EMBO J 13: 554-561. 126. Yi, L., N. Celebi, M. Chen, and R.E. Dalbey, (2004) Sec/SRP requirements and energetics of membrane insertion of subunits a, b, and c of the Escherichia coli F1F0 ATP synthase. J Biol Chem 279: 39260-39267. 127. Matsuyama, S., Y. Fujita, and S. Mizushima, (1993) SecD Is Involved in the Release of Translocated Secretory Proteins from the Cytoplasmic Membrane of Escherichia coli. EMBO J 12: 265-270. 128. Duong, F. and W. Wickner, (1997) The SecDFyajC domain of preprotein translocase controls preprotein movement by regulating SecA membrane cycling. EMBO J 16: 4871-4879. 129. Albers, S.V., Z. Szabo, and A.J. Driessen, (2006) Protein secretion in the Archaea: multiple paths towards a unique cell surface. Nat

52 Rev Microbiol 4: 537-547. 130. van de Vossenberg, J.L., S.V. Albers, C. van der Does, A.J. Driessen, and W. van Klompenburg, (1998) The positive inside rule is not determined by the polarity of the delta psi (transmembrane electrical potential) [letter]. Mol Microbiol 29: 1125-1127. 131. Urbanus, M.L., L. Froderberg, D. Drew, P. Bjork, J.W. de Gier, J. Brunner, B. Oudega, and J. Luirink, (2002) Targeting, insertion, and localization of Escherichia coli YidC. J Biol Chem 277: 12718- 12723. 132. Pohlschroder, M., E. Hartmann, N.J. Hand, K. Dilks, and A. Haddad, (2005) Diversity and evolution of protein translocation. Annu Rev Microbiol 59: 91-111. 133. Samuelson, J.C., M. Chen, F. Jiang, I. Moller, M. Wiedmann, A. Kuhn, G.J. Phillips, and R.E. Dalbey, (2000) YidC mediates membrane protein insertion in bacteria. Nature 406: 637-641. 134. Chen, M., J.C. Samuelson, F. Jiang, M. Müller, A. Kuhn, and R.E. Dalbey, (2002) Direct interaction of YidC with the Sec-independent Pf3 coat protein during its membrane protein insertion. J Biol Chem 277: 7670-7675. 135. Froderberg, L., E. Houben, J.C. Samuelson, M.Y. Chen, S.K. Park, G.J. Phillips, R. Dalbey, J. Luirink, and J.W.L. de Gier, (2003) Versatility of inner membrane protein biogenesis in Escherichia coli. Mol Microbiol 47: 1015-1027. 136. van der Laan, M., M.L. Urbanus, C.M. Ten Hagen-Jongman, N. Nouwen, B. Oudega, N. Harms, A.J. Driessen, and J. Luirink, (2003) A conserved function of YidC in the biogenesis of respiratory chain complexes. Proc Natl Acad Sci U S A 100: 5801- 5806. 137. Van Der Laan, M., P. Bechtluft, S. Kol, N. Nouwen, and A.J. Driessen, (2004) F1F0 ATP synthase subunit c is a substrate of the novel YidC pathway for membrane protein biogenesis. J Cell Biol 165: 213-222. 138. van Bloois, E., G. Jan Haan, J.W. de Gier, B. Oudega, and J. Luirink, (2004) F(1)F(0) ATP synthase subunit c is targeted by the SRP to YidC in the E. coli inner membrane. FEBS Lett 576: 97-100. 139. Facey, S.J., S.A. Neugebauer, S. Krauss, and A. Kuhn, (2007) The mechanosensitive channel protein MscL is targeted by the SRP to the novel YidC membrane insertion pathway of Escherichia coli. J Mol Biol 365: 995-1004. 140. Scotti, P.A., M.L. Urbanus, J. Brunner, J.W.L. de Gier, G. von Heijne, C. van der Does, A.J.M. Driessen, B. Oudega, and J. Luirink, (2000) YidC, the Escherichia coli homologue of mitochondrial Oxa1p, is a component of the Sec translocase. EMBO J 19: 542-549. 141. Facey, S.J. and A. Kuhn, (2004) Membrane integration of E. coli model membrane proteins. Biochim Biophys Acta 1694: 55-66. 142. Urbanus, M.L., P.A. Scotti, L. Froderberg, A. Saaf, J.W. de Gier, J.

53 Brunner, J.C. Samuelson, R.E. Dalbey, B. Oudega, and J. Luirink, (2001) Sec-dependent membrane protein insertion: sequential interaction of nascent FtsQ with SecY and YidC. EMBO Rep 2: 524- 529. 143. van der Laan, M., E.N. Houben, N. Nouwen, J. Luirink, and A.J. Driessen, (2001) Reconstitution of Sec-dependent membrane protein insertion: nascent FtsQ interacts with YidC in a SecYEG-dependent manner. EMBO Rep 2: 519-523. 144. Houben, E.N., P.A. Scotti, Q.A. Valent, J. Brunner, J.L. de Gier, B. Oudega, and J. Luirink, (2000) Nascent Lep inserts into the Escherichia coli inner membrane in the vicinity of YidC, SecY and SecA. FEBS Lett 476: 229-233. 145. van Bloois, E., G.J. Haan, J.W. de Gier, B. Oudega, and J. Luirink, (2006) Distinct requirements for translocation of the N-tail and C- tail of the Escherichia coli inner membrane protein CyoA. J Biol Chem 281: 10002-10009. 146. Nagamori, S., I.N. Smirnova, and H.R. Kaback, (2004) Role of YidC in folding of polytopic membrane proteins. J Cell Biol 165: 53-62. 147. Froderberg, L., E.N. Houben, L. Baars, J. Luirink, and J.W. de Gier, (2004) Targeting and translocation of two lipoproteins in Escherichia coli via the SRP/Sec/YidC pathway. J Biol Chem 279: 31026-31032. 148. Yen, M.R., Y.H. Tseng, E.H. Nguyen, L.F. Wu, and M.H. Saier, Jr., (2002) Sequence and phylogenetic analyses of the twin-arginine targeting (Tat) protein export system. Arch Microbiol 177: 441-450. 149. Sargent, F., B.C. Berks, and T. Palmer, (2006) Pathfinders and trailblazers: a prokaryotic targeting system for transport of folded proteins. FEMS Microbiol Lett 254: 198-207. 150. Rodrigue, A., A. Chanal, K. Beck, M. Muller, and L.F. Wu, (1999) Co-translocation of a periplasmic enzyme complex by a hitchhiker mechanism through the bacterial tat pathway. J Biol Chem 274: 13223-13228. 151. Hatzixanthis, K., T. Palmer, and F. Sargent, (2003) A subset of bacterial inner membrane proteins integrated by the twin-arginine translocase. Mol Microbiol 49: 1377-1390. 152. Bos, M.P., V. Robert, and J. Tommassen, (2006) Biogenesis of the Gram-Negative Bacterial Outer Membrane. Annu Rev Microbiol. 153. Chen, R. and U. Henning, (1996) A periplasmic protein (Skp) of Escherichia coli selectively binds a class of outer membrane proteins. Mol Microbiol 19: 1287-1294. 154. Harms, N., G. Koningstein, W. Dontje, M. Muller, B. Oudega, J. Luirink, and H. de Cock, (2001) The early interaction of the outer membrane protein PhoE with the periplasmic chaperone Skp occurs at the cytoplasmic membrane. J Biol Chem 276: 18804-18811. 155. Spiess, C., A. Beil, and M. Ehrmann, (1999) A temperature- dependent switch from chaperone to protease in a widely conserved

54 heat shock protein. Cell 97: 339-347. 156. Wu, T., J. Malinverni, N. Ruiz, S. Kim, T.J. Silhavy, and D. Kahne, (2005) Identification of a multicomponent complex required for outer membrane biogenesis in Escherichia coli. Cell 121: 235-245. 157. Ruiz, N., B. Falcone, D. Kahne, and T.J. Silhavy, (2005) Chemical conditionality: a genetic strategy to probe organelle assembly. Cell 121: 307-317. 158. Voulhoux, R. and J. Tommassen, (2004) Omp85, an evolutionarily conserved bacterial protein involved in outer-membrane-protein assembly. Res Microbiol 155: 129-135. 159. Hegde, R.S. and H.D. Bernstein, (2006) The surprising complexity of signal sequences. Trends Biochem Sci 31: 563-571. 160. Casadio, R., P. Fariselli, G. Finocchiaro, and P.L. Martelli, (2003) Fishing new proteins in the twilight zone of genomes: the test case of outer membrane proteins in Escherichia coli K12, Escherichia coli O157:H7, and other Gram-negative bacteria. Protein Sci 12: 1158-1168.

55 Downloaded from www.proteinscience.org on August 6, 2007

New Escherichia coli outer membrane proteins identified through prediction and experimental verification

PAOLA MARANI,1,2,4 SAMUEL WAGNER,1,4 LOUISE BAARS,1 PIERRE GENEVAUX,3 JAN-WILLEM DE GIER,1 INGMARIE NILSSON,1 2 1 RITA CASADIO, AND GUNNAR VON HEIJNE 1Department of Biochemistry and Biophysics, Stockholm University, SE-106 91 Stockholm, Sweden 2Laboratory of Biocomputing, CIRB/Department of Biology, University of Bologna, Bologna, Italy 3Department of Microbiology and Molecular Medicine, Centre Me´dical Universitaire, CH-1211 Geneva, Switzerland

(RECEIVED October 5, 2005; FINAL REVISION December 23, 2005; ACCEPTED December 23, 2005)

Abstract Many new Escherichia coli outer membrane proteins have recently been identified by proteomics techniques. However, poorly expressed proteins and proteins expressed only under certain conditions may escape detection when wild-type cells are grown under standard conditions. Here, we have taken a complementary approach where candidate outer membrane proteins have been identified by bioinformatics prediction, cloned and overexpressed, and finally localized by cell fractionation experiments. Out of eight predicted outer membrane proteins, we have confirmed the outer membrane localization for five—YftM, YaiO, YfaZ, CsgF, and YliI—and also provide preliminary data indicating that a sixth—YfaL—may be an outer membrane autotransporter. Keywords: outer membrane protein; bioinformatics; SecB; autotransporter

From the known high-resolution structures of trans- An important use of bioinformatics prediction schemes membrane proteins, only two basic architectures have is to guide the experimentalist toward targets that been identified so far: the helix bundle and the b-barrel are highly likely to correspond to true instances of (von Heijne 1999). Helix bundle proteins have been the particular kind of gene or protein of interest. Here, extensively studied, both from an experimental and we have used the recently developed Hunter predictor from a bioinformatics perspective, and rather reliable (Casadio et al. 2003) to select likely outer membrane prediction methods exist for their identification from proteins among the nonannotated part of the Escheri- sequence data alone (Chen et al. 2002; Mele´n et al. chia coli proteome, and have experimentally verified the 2003). b-Barrel proteins have received comparatively predicted outer membrane localization of five hitherto less attention, and only a few methods have been pro- uncharacterized proteins: YftM, YaiO, YfaZ, CsgF, and posed for identification and topology prediction of such YliI. We further provide data indicating that a sixth pro- proteins (Casadio et al. 2003; Bagos et al. 2004; Berven tein, YfaL, is an outer membrane autotransporter. et al. 2004; Bigelow et al. 2004).

Results

4These authors contributed equally to this work. Selection of target proteins Reprint requests to: Gunnar Von Heijne, Department of Biochem- istry and Biophysics, Stockholm University, SE-106 91 Stockholm, From the list of 18 new outer membrane proteins pre- Sweden; e-mail: [email protected]; fax: +46-8-15-36-79. Article published online ahead of print. Article and publication date are dicted by the Hunter predictor in the E. coli proteome at http://www.proteinscience.org/cgi/doi/10.1110/ps.051889506. (see Table 3 in Casadio et al. 2003), we initially chose 11

New outer membrane proteins proteins, characterized by different lengths and different numbers of predicted b-strands, for further analysis. Despite repeated attempts, only eight of these genes could be cloned in our vector system. Therefore, we focused our experimental analysis on this set of putative outer membrane proteins (Table 1).

Cloning and expression of target proteins The eight target genes were cloned into the pING vector (Johnston et al. 1985), and a hemagglutinin (HA) tag was added to the C terminus of the gene products for immu- nodetection. Induction with arabinose and labeling with [35S]-Met in all cases gave rise to a protein product that could be immunoprecipitated by an HA antibody and was Figure 1. Translocation of the target proteins across the inner mem- of the expected molecular weight (data not shown). brane is SecA dependent. Protein expression was induced for 5 min 35 In initial [ S]-Met labeling experiments, we noted that with arabinose, followed by labeling for 1 min with [35S]-Met (YliI was seven of the eight proteins appeared as doublets (Fig. 1), labeled for 3 min). Sodium azide was added to a final concentration of possibly reflecting inefficient removal of the signal peptide 2 mM (+ lanes) 30 sec prior to radio-labeling. Proteins were immunoprecipitated with antisera against the HA-tag. Precursor (p) and (because of its large size, small molecular weight differences mature (m) forms of the proteins are indicated. could not be detected in pulse-chase experiments with YfaL). To study this possibility further, we blocked SecA- dependent translocation through the inner membrane centrifugation. The purity of the inner and outer mem- SecYEG translocon by adding sodium azide 30 sec prior brane fractions was determined by Western blotting to the addition of [35S]-Met (Oliver et al. 1990). As seen in against the inner and outer membrane marker proteins Figure 1, after a 1-min pulse, significantly more of the higher Lep and OmpA, respectively. molecular-weight form was seen in the presence than in the All target proteins were found in the outer membrane absence of azide for all seven proteins, strongly suggesting fraction, together with the control outer membrane pro- that these proteins are translocated across the inner mem- tein OmpA (Fig. 2). A higher molecular-weight form brane in a SecA- and translocon-dependent process. migrating slightly slower than the 150-kDa standard was seen for YtfM, possibly representing an SDS-resistant dimeric form of the protein. For the relatively Outer membrane localization of target proteins strongly expressed YaiO and YliI proteins, trace amounts To assay the possible outer membrane localization of the were also present in the inner membrane fraction, most target proteins, we separated outer and inner mem- likely due to cross-contamination. branes by successive two- and six-step sucrose gradient It was recently shown that cytosolic aggregates of misfolded proteins cosediment with the outer membrane fraction upon sucrose density gradient centrifugation, and Table 1. Proteins included in the study that the inclusion-body binding proteins IbpA and IbpB can be used as a marker for these aggregates (Laskowska Molecular weight et al. 2004). To evaluate if our overexpressed target pro- (processed, Urea SecB UniProt teins are inserted into the outer membrane and do not UniProt code HA tagged) pellet dependence annotation simply copurify in aggregates, we developed a protocol in CSGF_ECOLI 14.2 + + Biogenesis of which the purified outer membrane fraction is washed curli organelles with 5 M urea to dissolve potential aggregates but leave YHJY_ECOLI 25.1 – + Lipase 1 membranes intact. Similar procedures are often used to YTFM_ECOLI 63.8 ++ + Hypothetical demonstrate the correct insertion of helix bundle proteins protein into membranes (Chen et al. 2003). Western blots against YFAL_ECOLI 129.7 N.D. N.D. Putative autotransporter IbpA,B showed that aggregates are solubilized by urea YAIO_ECOLI 28.3 ++ + – treatment, whereas the major outer membrane protein YLII_ECOLI 40.0 + + Putative glucose OmpA remains in the membrane pellet (Fig. 3A). dehydrogenase As shown in Figure 3B, only YaiO and YftM remained YFAZ_ECOLI 17.9 + + – totally in the urea-resistant outer membrane fraction (the YAGZ_ECOLI 19.1 – + Fimbrillin latter gave rise to two additional lower molecular weight

www.proteinscience.org 885 Downloaded from www.proteinscience.org on August 6, 2007

Marani et al.

We considered the possibility that the 55-kDa fragment is the cleaved translocator domain of the autotransporter, which would be similar in size to the AIDA-I translocator domain (47.5 kDa). As many autotransporters are serine/ threonine proteases, we tested if the cleavage of the putative translocator domain could be inhibited by the serine/ threonine protease inhibitor Pefabloc SC. Indeed, when YfaL was expressed in the presence of Pefabloc SC, the 55-kDa band disappeared (Fig. 4). The putative cleaved 75-kDa N-terminal passenger domain does not contain a HA-tag and thus cannot be detected by Western blotting. Figure 2. Membrane fractionation. Cells were grown at 37C, and expression of HA-tagged target proteins was induced with arabinose SecB dependence at an OD600 of 0.4–0.6. Cells were harvested 45 min after induction and lysed by French pressing. Inner and outer membrane fractions were It is generally assumed that the cytoplasmic chaperone prepared by sucrose density gradient centrifugation, separated by SDS- SecB facilitates the export of precursor polypeptides by PAGE, and probed by immunodecoration of the HA-tagged proteins. Western blots of the outer membrane marker OmpA and the inner membrane marker Lep are also shown. bands upon urea treatment; the identity of these bands is unknown). CsgF and YfaZ were also largely urea-resistant, while YfaL, YliI, YhjY, and YagZ were to a greater or lesser extent removed from the outer membrane fraction. Since the formation of inclusion bodies is often reduced at lower temperature, YfaL, YliL, YhjY, and CsgF were expressed also at 30C (Fig. 3C). The amounts of CsgF and YliI in the outer membrane fraction increased, whereas YhjY was still mainly extracted. Expression of YfaL was too low for detection under these conditions. Our results strongly suggest that at least five of the eight proteins (YftM, YaiO, YfaZ, CsgF, YliI) are localized to the outer membrane. The 130-kDa protein YfaL has been predicted to be an autotransporter (Yen et al. 2002). Supporting this, the C-terminal part shows considerable homology with the AIDA-I autotransporter of E. coli O126:H27, which mediates binding to an integral membrane glycoprotein on HeLa cells (Laarmann and Schmidt 2003). In general, autotransporters consist of an outer membrane C-terminal translocator domain and a globular N-terminal passenger domain that mediates the ultimate function of the protein. After translocation through the outer membrane, the passenger domain is (often autolytically) cleaved off (Henderson et al. 1998). As noted above, we could not detect YfaL in the outer membrane after cell fractionation and urea extraction. A Figure 3. Urea wash of outer membrane fractions. Outer membrane fractions from MC1061 cells overexpressing the different target proteins substantial amount of the expressed protein seems to were prepared as described in Figure 2. After fractionation, membranes end up in aggregates, which is in concert with the fact were washed with PBS plus 5 M urea for 1 h in the cold. Urea-treated that the inclusion body binding protein IbpB is up-regu- membranes (+ wash) and untreated controls (– wash) were analyzed by lated upon overexpression of YfaL (data not shown). Western blotting. (A) Western blots of the inclusion body binding protein Interestingly, when whole cells were subjected to SDS- B (IbpB; the cells in this experiment were induced for expression of YfaL) and the major outer membrane protein OmpA. (B) Western blots of the PAGE and Western blotting against the C-terminal HA- indicated HA-tagged target proteins expressed at 37C. (C) Western tag, an additional band at 55 kDa appeared (Fig. 4). blots of the indicated HA-tagged target proteins expressed at 30C.

886 Protein Science, vol. 15 Downloaded from www.proteinscience.org on August 6, 2007

New outer membrane proteins

identification of bacterial outer membrane proteins, and may be particularly effective for low-abundance proteins or proteins that are expressed only under certain conditions. The outer membrane localization of the five proteins was experimentally verified by sucrose density gradient centrifugation and urea treatment of the outer membranes. Only YhjY gave ambiguous results: We could not determine if it is located in the outer membrane as it is difficult to overex- press and ends up mostly in inclusion bodies. YagZ could be Figure 4. Analysis of the putative outer membrane autotransporter extracted completely from the outer membrane fraction YfaL. Cells were grown in LB medium at 30C, and expression of with 5 M urea and is thus not embedded in the outer mem- HA-tagged YfaL was induced with arabinose at an OD600 of 0.4–0.6 for 2 h. Induced cells were grown in the presence or the absence of brane. YagZ is identical with the protein MatB (meningitis- Pefabloc SC. Cells were harvested and subsequently analyzed by Wes- associated and temperature-regulated), which has been tern blotting against the HA-tag. Full-length YfaL (I) and the putative shown recently to be the major fimbrillin of the Mat fimbria 55-kDa translocator domain (II) are indicated. (and thus not directly inserted in the outer membrane) (Pouttu et al. 2001). MatB is expressed in some pathogenic maintaining them in a translocation competent conforma- strains (MENEC) but not in the laboratory strain K12. tion and by delivering them to SecA (Randall and Hardy Finally, our results indicate that YfaL is an autotransporter 2002). However, up until recently, it has been shown for with a cleavable 55-kDa translocator domain. only six proteins (PhoE, LamB, MBP, GBP, OmpF, and In common with previously studied outer membrane OmpA) that their export is facilitated by SecB (Kuma- proteins, we find that targeting of the outer membrane moto and Beckwith 1985; de Cock et al. 1992; Powers and proteins identified here is facilitated by the SecB chaper- Randall 1995), whereas four proteins (PhoA, Lpp, RbsB, one, suggesting that SecB dependence may be a common and AmpC) do not seem to require SecB (Knoblauch et al. characteristic of outer membrane proteins. 1999; Xu et al. 2000; Randall and Hardy 2002; Dekker et al. 2003). Twelve additional proteins were recently identi- Materials and methods fied as SecB substrates in a proteomics screen (Baars et al. 2006). Enzymes and chemicals In an attempt to expand the list even further, we expressed HA-tagged CsgF, YfaZ, YagZ, YhjY, YaiO, Unless otherwise stated, all enzymes were from Promega or YliI, and YftM in the secB null strain MC4100DsecB (no New England Biolabs. [35S]-Met and [14C]-methylated marker expression was seen for YfaL) and the control strain MC4100 (Fig. 5). With the possible exception of YhjY, the relative amount of precursor was clearly increased in the secB null strain compared with wild type for all seven proteins, indicating that SecB facilitates their targeting. As expected, the uncleaved precursor form pro-OmpA accumulated in the transformed secB null strains but not in the transformed control strains (data not shown).

Discussion Until recently, identification of bacterial outer membrane proteins by computational approaches (other than standard sequence similarity searches) has been a neglected field in bioinformatics. Here, we have experimentally verified that at least five of the top candidate outer membrane proteins identified by the Hunter predictor among the Figure 5. Analysis of SecB dependence. The indicated HA-tagged unannotated portion of the E. coli proteome (Casadio et target proteins were expressed in the secB null mutant MC4100DsecB al. 2003)—YftM, YaiO, YfaZ, CsgF, and YliI—are in fact (– SecB) and the wild-type strain MC4100 (+ SecB). Protein expres- localized in the outer membrane. Target protein selection sion was induced with arabinose for 5 min, followed by labeling for 1 min with [35S]-Met. Proteins were immunoprecipitated with antisera based on bioinformatics predictions followed by experi- against the HA-tag. Precursor (p) and mature (m) forms of the proteins mental verification is thus a viable alternative to large- are indicated. Note the slow cleavage of YliI, which is complete only scale proteomics approaches (Molloy et al. 2000) for the after a 3-min pulse in wild-type cells (Fig. 1).

www.proteinscience.org 887 Downloaded from www.proteinscience.org on August 6, 2007

Marani et al.

proteins were from Amersham-Pharmacia Biotech. Protein A– grown at 37C and 30C, respectively. Expression of the outer Sepharose and sodium azide were from Sigma Chemical. Pan- membrane proteins was induced by the addition of 0.1% arab- sorbin was from Calbiochem Biochemicals &Immunochemicals. inose at an OD600 of 0.4–0.6. Cells were harvested 45 min after BigDye Terminator v1.1 Cycle Sequencing Kit was from AB induction at 6000g using a Beckman 8.1000 rotor. Applied Biosystems,, and oligonucleotides were from Cyber- The 1000 OD600 units of cells were resuspended in 6 mL Gene AB. The QuikChange site-directed mutagenesis kit was buffer K (50 mM triethanolamine [TEA], 250 mM sucrose, 1 from Stratagene. The Expand Long Template PCR System was mM EDTA, 1 mM dithiothreitol [DTT], 0.1 mg/mL Pefabloc from Roche Diagnostics GmbH, and the QIAquick PCR puri- at pH 7.5) and lysed by two cycles of French pressing (18,000 fication kit was from Qiagen. All mutants were confirmed by psi). The lysate was clarified of unbroken cells by 20-min sequencing of plasmid DNA at BM Labbet AB. Rabbit poly- centrifugation at 8,000g. The supernatant was transferred on clonal anti-HA-tag (influenza HA-epitope) antibody was from top of a two-step sucrose gradient: bottom to top, 1 mL 55% Abcam Limited. BCA protein concentration assay was from (w/w), 5.5 mL 9% (w/w). All sucrose gradients were prepared Pierce, and Pefabloc was from Biomol. in buffer M: 50 mM TEA, 1 mM EDTA, and 1 mM DTT (pH 7.5). The gradients were spun for 2.5 h at 210,000g in a Beck- man SW 40 rotor, and the membrane fraction was collected DNA techniques from the top of the 55% sucrose step. This fraction, which contains the entire membranes, was diluted 1:1 with buffer M The genes encoding the eight E. coli target proteins were ampli- and subjected to a six-step sucrose gradient to obtain pure fied from E. coli strains MG1655 (Blattner et al. 1997) or inner and outer membrane fractions. The assembly of this MC1061 using Expand Long Polymerase. For cloning into and second gradient was as follows (from bottom to top): 0.8 mL in vivo expression from the pING1 plasmid (see Whitley et al. at 55%; 2.0 mL at 50%,45%,40%, and 35%; 0.8 mL at 30% (all 1994), both ends of the gene were modified during PCR ampli- w/w) and 3.3 mL of the sample. The gradients were spun for 15 fication by introducing a XhoI site and an initiator ATG codon h at 210,000g in a Beckman SW 40 rotor, and the inner and encoded by the 5¢ primer, and by changing the 3¢ end of the gene outer membrane fractions were collected from the top of the by a reverse primer encoding a HA-tag, YPYDVPDYA, two 40% and 50% sucrose steps, respectively. stop codons (TAA TAG), and a SmaI site. Thus, the 5¢ region of The purity of the fractions was confirmed by Western blot- the gene was modified to …CTCGAGTATG… (XhoI site and ting against Lep and OmpA as inner and outer membrane initiator codon underlined). The resulting fragment was cloned markers, respectively. The protein concentration of the frac- into the pING vector behind the ara promoter using an XhoI site tions was determined by a BCA assay according to the instruc- and a SmaI site introduced by site-specific mutagenesis. tions of the manufacturer (Pierce).

Strains, plasmids, culture conditions, and pulse Aggregate removal experiments Outer membranes containing 50 mg of protein were resus- Experiments were performed in E. coli strain MC1061 (Dalbey pended in PBS/5 M urea and washed by rotating for 1 h in and Wickner 1986), MC4100 (Casadaban and Cohen 1979), the cold room. Membranes were collected in Beckman TLA and MC4100DsecB (R.S. Ullers., F. Schwager, D. Ang, C. Geor- 100.3 at 194,000g for 20 min. Urea-washed membranes and gopoulos, and P. Genevaux, in prep.). Constructs were expressed unwashed control membranes were analyzed by Western blot- from the pING plasmid (Johnston et al. 1985) by induction with ting against the HA-tag. Blotting against OmpA was used as a L-arabinose. control for a protein that is properly inserted into the outer E. coli strains were transformed with the pING vector carry- membrane and that cannot be washed away. Blotting against ing the relevant constructs under control of the arabinose IbpB was used to show that the aggregates could be washed promoter were grown at 37C in M9 minimal medium supple- away by the 5 M urea treatment. mented with 100 mg/mL ampicillin, 0.5% (w/v) fructose, 100 mg/mL thiamine, and all amino acids (50 mg/mL each) except methionine. An overnight culture was diluted 1:25 in fresh Immunoblot analysis medium, shaken for 3.5 h at 37C, induced with arabinose (0.2% [w/v]) for 5 min, labeled with [35S]-Met (75 mCi/mL) The expression of the target proteins (with HA-tag fused to the for 1 min, and put on ice. Sodium azide (final concentration C terminus), Lep, OmpA, and IbpA,B (the IbpB antiserum 2 mM) was added 30 sec before radiolabeling. Samples were cross-reacts with IbpA) in the inner/outer membranes and acid-precipitated with trichloroacetic acid (TCA) (10% [v/v] aggregates was monitored by immunoblot analysis. Cells final concentration), resuspended, and then analyzed by immu- were cultured as described above. Purified inner/outer mem- noprecipitation with HA-antiserum combined with SDS- branes or aggregates (5 mg protein) were solubilized in Lae- PAGE as described previously (Fro¨derberg et al. 2004). Pro- mmli solubilization buffer and were separated by SDS-PAGE. teins were visualized in a Fuji FLA-3000 PhosphorImager Proteins were transferred from the polyacrylamide gel to a using the Image Reader V1.8J/Image Gauge V 3.45 software. polyvinylidene fluoride (PVDF) membrane (Millipore). Subse- quently, membranes were blocked with 5% milk and decorated with antisera to the components listed above. Proteins were Separation of outer and inner membrane fractions visualized with secondary HRP-conjugated antibodies (Bio- Rad) using the ECL system according to the instructions Cell fractionation was carried out essentially as described in of the manufacturer (Amersham Pharmacia) and a Fuji LAS Laskowska et al. (2004) using two subsequent sets of sucrose 1000-Plus CCD camera. Blots were quantified using the Image density gradients. Samples (1000 mL) of strain MC1061 trans- Gauge software (version 3.4). Experiments were repeated at formed with a pING vector harboring the different OMPs were least twice. If the membrane had to be tested with more than

888 Protein Science, vol. 15 Downloaded from www.proteinscience.org on August 6, 2007

New outer membrane proteins

one antibody, it was washed with 5 M urea and 10 mM DTT Dalbey, R.E. and Wickner, W. 1986. The role of the polar, carboxyl- overnight at 37C, blocked, and reused as before (Terzi et al. terminal domain of Escherichia coli leader peptidase in its translocation 2004). across the plasma membrane. J. Biol. Chem. 261: 13844–13849. de Cock, H., Overeem, W., and Tommassen, J. 1992. Biogenesis of outer membrane protein PhoE of Escherichia coli: Evidence for multiple SecB-binding sites in the mature portion of the PhoE protein. J. Mol. Protease inhibition assay for YfaL Biol. 224: 369–379. Dekker, C., de Kruijff, B., and Gros, P. 2003. Crystal structure of SecB MC1061 transformed with a pING vector harboring the yfaL from Escherichia coli. J. Struct. Biol. 144: 313–319. gene fused to a C-terminal HA-tag was grown at 30CinLB Fro¨derberg, L., Houben, E.N., Baars, L., Luirink, J., and de Gier, J.W. m 2004. Targeting and translocation of two lipoproteins in Escherichia medium supplemented with 100 g/mL ampicillin. Expression coli via the SRP/Sec/YidC pathway. J. Biol. Chem. 279: 31026– was induced by the addition of 0.1% arabinose at an OD600 of 31032. 0.4–0.6 in the presence or absence of 1 mg/mL Pefabloc serine Henderson, I.R., Navarro-Garcia, F., and Nataro, J.P. 1998. The great protease inhibitor. Cells were harvested 2 h after induction, escape: Structure and function of the autotransporter proteins. Trends and 0.15 OD units of whole cells/well was run on an SDS- Microbiol. 6: 370–378. 600 Johnston, S., Lee, J.H., and Ray, D.S. 1985. High-level expression of M13 PAGE and analyzed by Western blotting as described above. gene II protein from an inducible polycistronic messenger RNA. Gene 34: 137–145. Knoblauch, N., Ru¨diger, S., Scho¨nfeld, H.-J., Driessen, A., Schneider- Acknowledgments Mergener, J., and Bukau, B. 1999. Substrate specificity of the SecB chaperone. J. Biol. Chem. 274: 34219–34225. We thank B. Bukau for gift of IbpB antiserum, and C. Geor- Kumamoto, C.A. and Beckwith, J. 1985. Evidence for specificity at an gopoulos, in whose laboratory part of the work was per- early step in protein export in Escherichia coli. J. Bacteriol. 163: 267– 274. formed. This work was supported by grants from the Swedish Laarmann, S. and Schmidt, M.A. 2003. The Escherichia coli AIDA auto- Research Council and the Marianne and Marcus Wallenberg transporter adhesin recognizes an integral membrane glycoprotein as Foundation to G.v.H.; from FIRB and the European Commu- receptor. Microbiology 149: 1871–1882. nity BioSapiens and Functional Genomics programs to R.C.; Laskowska, E., Bohdanowicz, J., Kuczynska-Wisnik, D., Matuszewska, from Bologna University, CNR, and AIRBBC to P.M.; and E., Kedzierska, S., and Taylor, A. 2004. Aggregation of heat-shock – from the Swiss National Science Foundation (FN-31-65403) to denatured, endogenous proteins and distribution of the IbpA/B and Fda marker-proteins in Escherichia coli WT and grpE280 cells. Micro- P.G. biology 150: 247–259. Mele´n, K., Krogh, A., and von Heijne, G. 2003. Reliability measures for membrane protein topology prediction algorithms. J. Mol. Biol. 327: References 735–744. Molloy, M.P., Herbert, B.R., Slade, M.B., Rabilloud, T., Nouwens, Baars, L., Ytterberg, J., Drew, D., Wagner, S., Thilo, C., van Wijk, K.-J., A.S., Williams, K.L., and Gooley, A.A. 2000. Proteomic analysis and de Gier, J.-W. 2006. Defining the role of the E. coli chaperone SecB of the Escherichia coli outer membrane. Eur. J. Biochem. 267: 2871– using comparative proteomics. J. Biol. Chem. (in press). 2881. Bagos, P.G., Liakopoulos, T.D., Spyropoulos, I.C., and Hamodrakas, S.J. Oliver, D.B., Cabelli, R.J., Dolan, K.M., and Jarosik, G.P. 1990. Azide- 2004. PRED-TMBB: A web server for predicting the topology of b- resistant mutants of Escherichia coli alter the SecA protein, an azide- barrel outer membrane proteins. Nucleic Acids Res. 32: W400–W404. sensitive component of the protein export machinery. Proc. Natl. Acad. Berven, F.S., Flikka, K., Jensen, H.B., and Eidhammer, I. 2004. BOMP: A Sci. 87: 8227–8231. program to predict integral b-barrel outer membrane proteins encoded Pouttu, R., Westerlund-Wikstrom, B., Lang, H., Alsti, K., Virkola, R., within genomes of Gram-negative bacteria. Nucleic Acids Res. 32: Saarela, U., Siitonen, A., Kalkkinen, N., and Korhonen, T.K. 2001. W394–W399. matB, a common fimbrillin gene of Escherichia coli, expressed in a Bigelow, H.R., Petrey, D.S., Liu, J., Przybylski, D., and Rost, B. 2004. genetically conserved, virulent clonal group. J. Bacteriol. 183: 4727– Predicting transmembrane b-barrels in proteomes. Nucleic Acids Res. 4736. 32: 2566–2577. Powers, E.L. and Randall, L.L. 1995. Export of periplasmic galactose- Blattner, F.R., Plunkett, G., Bloch, C.A., Perna, N.T., Burland, V., Riley, binding protein in Escherichia coli depends on the chaperone SecB. J. M., Collado-Vides, J., Glasner, J.D., Rode, C.K., Mayhew, G.F., et al. Bacteriol. 177: 1906–1907. 1997. The complete genome sequence of Escherichia coli K-12. Science Randall, L.L. and Hardy, S.J. 2002. SecB, one small chaperone in the 277: 1453–1462. complex milieu of the cell. Cell. Mol. Life Sci. 59: 1617–1623. Casadaban, M.J. and Cohen, S.N. 1979. Lactose genes fused to exogenous Terzi, L., Pool, M.R., Dobberstein, B., and Strub, K. 2004. Signal recogni- promoters in one step using a Mu-lac bacteriophage: In vivo probe for tion particle Alu domain occupies a defined site at the ribosomal transcriptional control sequences. Proc. Nat. Acad. Sci. 76: 4530–4533. subunit interface upon signal sequence recognition. Biochemistry 43: Casadio, R., Fariselli, P., Finocchiaro, G., and Martelli, P.L. 2003. 107–117. Fishing new proteins in the twilight zone of genomes: The test von Heijne, G. 1999. Recent advances in the understanding of membrane case of outer membrane proteins in Escherichia coli K12, Escherichia protein assembly and structure. Q. Rev. Biophys. 32: 285–307. coli O157:H7, and other Gram-negative bacteria. Protein Sci. 12: Whitley, P., Nilsson, I., and von Heijne, G. 1994. De novo design of 1158–1168. integral membrane proteins. Nat. Struct. Biol. 1: 858–862. Chen, C.P., Kernytsky, A., and Rost, B. 2002. Transmembrane helix pre- Xu, Z., Knafels, J.D., and Yoshino, K. 2000. Crystal structure of the dictions revisited. Protein Sci. 11: 2774–2791. bacterial protein export chaperone secB. Nat. Struct. Biol. 7: 1172– Chen, Y., Song, J., Sui, S.F., and Wang, D.N. 2003. DnaK and DnaJ 1177. facilitated the folding process and reduced inclusion body formation of Yen, M.R., Peabody, C.R., Partovi, S.M., Zhai, Y., Tseng, Y.H., and magnesium transporter CorA overexpressed in Escherichia coli. Protein Saier, M.H. 2002. Protein-translocating outer membrane porins of Expr. Purif. 32: 221–231. Gram-negative bacteria. Biochim. Biophys. Acta 1562: 6–31.

www.proteinscience.org 889 THE JOURNAL OF BIOLOGICAL CHEMISTRY Vol. 279, No. 30, Issue of July 23, pp. 31026–31032, 2004 © 2004 by The American Society for Biochemistry and Molecular Biology, Inc. Printed in U.S.A. Targeting and Translocation of Two Lipoproteins in Escherichia coli via the SRP/Sec/YidC Pathway*

Received for publication, March 23, 2004, and in revised form, May 12, 2004 Published, JBC Papers in Press, May 12, 2004, DOI 10.1074/jbc.M403229200

Linda Fro¨ derberg‡§, Edith N. G. Houben¶, Louise Baars‡, Joen Luirink¶, and Jan-Willem de Gier‡ʈ From the ‡Department of Biochemistry and Biophysics, Arrhenius Laboratories, Stockholm University, SE-106 91 Stockholm, Sweden and the ¶Department of Microbiology, BioCentrum Amsterdam, De Boelelaan 1087, 1081 HV Amsterdam, The Netherlands

In Escherichia coli, two main protein targeting path- when it becomes exposed outside the ribosome. Upon contact of ways to the inner membrane exist: the SecB pathway for the SRP with its receptor, FtsY, the nascent IMP dissociates

the essentially posttranslational targeting of secretory from the SRP and enters the Sec translocase. Downloaded from proteins and the SRP pathway for cotranslational tar- The core of the Sec translocase consists of the IMPs, SecY geting of inner membrane proteins (IMPs). At the inner and SecE, and the peripheral subunit, SecA (1). SecY and SecE membrane both pathways converge at the Sec translo- form a protein-conducting channel (4), and SecA drives pro- case, which is capable of both linear transport into the teins in an ATP-dependent process through this channel (1). periplasm and lateral transport into the lipid bilayer. The Sec translocase catalyzes linear transport of secretory The Sec-associated YidC appears to assist the lateral proteins and of periplasmic domains of IMPs across the mem- www.jbc.org transport of IMPs from the Sec translocase into the lipid brane. In addition, TMs of IMPs are recognized in the Sec bilayer. It should be noted that targeting and transloca- translocase and laterally transferred into the lipid bilayer. The tion of only a handful of secretory proteins and IMPs Sec translocase-associated form of YidC appears to assist in have been studied. These model proteins do not include this lateral transfer of TMs (5, 6). YidC, which is present in lipoproteins. Here, we have studied the targeting and at Stockholm Universitetsbibliotek on August 6, 2007 translocation of two secretory lipoproteins, the murein excess over the Sec translocase, is also involved in the integra- lipoprotein and the bacteriocin release protein, using a tion of some small SRP/Sec-independent IMPs (7, 8). Thus far, combined in vivo and in vitro approach. The data indi- no evidence has been obtained that points to a role of YidC in cate that both murein lipoprotein and bacteriocin re- the targeting/translocation of secretory proteins across the in- 2 lease protein require the SRP pathway for efficient tar- ner membrane (7, 9). geting to the Sec translocase. Furthermore, we show It should be noted that in E. coli targeting and translocation that YidC plays an important role in the targeting/trans- of only a handful of secretory proteins and IMPs have been location of both lipoproteins. studied thoroughly. Hardly anything is known about the targeting and translocation of lipoproteins, which in most cases are secretory proteins. A lipoprotein is synthesized as a pre- In the bacterium Escherichia coli, the SecB pathway targets protein with an N-terminal signal sequence. Lipoproteins con- a subset of secretory proteins to the Sec translocase (1). The tain a conserved sequence, the “lipobox,” that includes the chaperone SecB keeps secretory proteins in a translocation- signal peptidase II (SPase II) cleavage site (10). The cysteine competent state. The SecB-preprotein complex is targeted at a located just after the SPaseII cleavage site is diacylglycerated late stage during translation or after translation to the Sec upon translocation; subsequently the signal sequence is clipped translocase in the inner membrane. The signal recognition off by SPaseII, yielding an apolipoprotein. Finally, the amino- particle (SRP)1 pathway targets IMPs to the same or a very modified cysteine is fatty acylated, giving rise to the mature similar Sec translocase in a cotranslational mechanism (2, 3). lipoprotein (11). Secretory lipoproteins can, depending on the The SRP, which consists of the Ffh protein and the 4.5 S RNA, sequence of the early mature region, remain associated to the binds to the first transmembrane segment (TM) of an IMP outer leaflet of the inner membrane or be transported by the Lol system to the outer membrane (12). To identify the components involved in the targeting and * This work was supported by grants from the Swedish Research Council, the Carl Tryggers Stiftelse, the European Molecular Biology translocation of secretory lipoproteins, we have analyzed the Organization (EMBO) Young Investigator Program (to J. W. dG.), and maturation of two model secretory lipoproteins. Maturation of the Dutch Research Council (to J.L.). The costs of publication of this lipoproteins has been studied in vivo using strains that are article were defrayed in part by the payment of page charges. This mutated in targeting and translocation factors and in vitro article must therefore be hereby marked “advertisement” in accordance using a translation/cross-linking system. One of the two model with 18 U.S.C. Section 1734 solely to indicate this fact. § Recipient of an EMBO short term fellowship. lipoproteins is the murein lipoprotein (Lpp), which is the most ʈ To whom correspondence should be addressed. Tel.: 46-8-164389 abundant protein in E. coli. Lpp is attached to the inner leaflet (laboratory)/162420 (office); Fax: 46-8-153679; E-mail: degier@dbb. of the outer membrane of E. coli and forms stable trimers (13). su.se. One of three Lpp molecules is covalently linked to the pepti- 1 The abbreviations used are: SRP, signal recognition particle; IMP, inner membrane protein; IMV, inverted membrane vesicle; IPTG, iso- doglycan layer (13). The other model lipoprotein is the propyl-1-thio-␤-D-galactopyranoside; LP, mature lipoprotein; OmpA, pCloCF13-encoded bacteriocin release protein (BRP) (14). The outer membrane protein A; TF, trigger factor; TM, transmembrane BRP is essential for the translocation of the bacteriocin cloacin segment; (Tmd)Phe, L-[3-(trifluoromethyl)-3-diazirin-3H-yl]phenylala- nine; U-PLP, unmodified prolipoprotein; HA, hemagglutinin; BRP, bacteriocin release protein; Lpp, murein lipoprotein; SPase, signal peptidase. 2 E. N. G. Houben and J. Luirink, unpublished observations.

31026 This paper is available on line at http://www.jbc.org BRP and LPP Targeting and Translocation 31027

DF13, a bactericidal protein, across the cell envelope (14). and the pEH3 vector in strain WAM121. For all experiments cells were Notably, the BRP signal sequence is very stable after cleavage grown to mid-log phase. Expression of the constructs was induced for 3 min with either L-arabinose (0.2%) or IPTG (1 mM). When indicated, the from the preprotein, in contrast to other signal sequences, and Ϫ SPaseII inhibitor globomycin (final concentration 100 ␮gml 1) was plays a yet undefined role in cloacin DF13 export (14). added 5 min before induction. Cells were labeled with [35S]methionine Our combined in vivo and in vitro studies indicate that the (60 ␮Ci/ml, Ci ϭ 37 GBq) for 30 s before precipitation with trichloro- SRP pathway plays an important role in the targeting of Lpp acetic acid (final concentration 10%). Subsequently, the samples were and BRP to the Sec translocase. Surprisingly, YidC is also washed with acetone, resuspended in 10 mM Tris/2% SDS, and immu- shown to function in the targeting/translocation of both secre- noprecipitated with anti-HA and anti-OmpA serum. Anti-HA immunoprecipitations were analyzed by means of Tricine SDS-PAGE (16.5% tory lipoproteins. peptide criterion gels from Bio-Rad) and anti-OmpA immunoprecipitations by means of standard SDS-PAGE. Gels were scanned by Fuji MATERIALS AND METHODS FLA-3000 phosphorimaging using the Image Reader V1.8J/Image Gauge Reagents, Enzymes, and Sera—All restriction enzymes, T4 DNA V 3.45 software. ligase, and alkaline phosphatase were purchased from Invitrogen. The E. coli Strains, Plasmids, and Growth Conditions for in Vitro Stud- Expand long template PCR kit was from Roche Applied Science. The ies—E. coli strain MC1061 grown in Luria Bertani medium supple-

QuikChange site-directed mutagenesis kit was from Stratagene, and mented with CaCl2 (10 mM) was used for all plasmid constructions. The the Megashort script T7 transcription kit was from Ambion Inc. QuikChange site-directed mutagenesis kit was used for the construc- [35S]methionine and protein A-Sepharose were from Amersham Bio- tion of all point mutations. Strain MRE600 was used to prepare a lysate sciences. Pansorbin was obtained from Merck. All other chemicals were for translation of in vitro synthesized mRNA and suppression of UAG supplied by Sigma. Antiserum against hemagglutinin (HA) tag was stop codons in the presence of (Tmd)Phe-tRNAsup. Inverted membrane Downloaded from purchased from Sigma and AbCam. Antisera against L23 and L29 were vesicles (IMVs) were prepared from strain MC4100 grown in Luria kind gifts from R. Brimacombe. The antisera against TF and SecA were Bertani medium. gifts from W. Wickner. Antisera against Ffh, YidC, and SecY were from Plasmid pC4Meth55LppTAG11 (Fig. 1) was constructed by plasmid our own collection. PCR and site-directed mutagenesis using pGEM42-Lpp as template E. coli Strains, Plasmids, and Growth Conditions for in Vivo Target- (17). Plasmid pC4Meth55BRPTAG10 was obtained by plasmid PCR ing and Translocation Studies—E. coli strain TOP10FЈ grown in Luria and site-directed mutagenesis using pJL28 as a template (16). Bertani medium was used for all plasmid constructions. In all lipopro- pC4Meth55LppTAG11 encodes truncated Lpp, and pC4Meth- tein expression vectors, the pCloDF13-encoded T1 terminator, which 55BRPTAG10 encodes truncated BRP. A methionine has been intro- www.jbc.org regulates the expression of the pCloDF13-encoded BRP, was cloned duced at position 16 in the Lpp signal sequence and at position 17 in the upstream of the genes encoding Lpp and BRP to prevent any back- BRP signal sequence, and both Lpp and BRP have been fused to a ground expression (15). Upon induction of expression, the T1 termina- C-terminal 4 ϫ methionine tag to improve labeling efficiency. Plasmid tor is “overruled.” The pCloDF13-encoded T1 terminator region was pC4Meth55LppTAG11 contains an amber mutation at position 11 in at Stockholm Universitetsbibliotek on August 6, 2007 amplified using pJL28 as a template (16). Primers were designed so the Lpp gene, and plasmid pC4Meth55BRPTAG10 contains an amber that both an NcoI site and a stop codon were introduced upstream and mutation (TAG) at position 10 in the BRP gene to enable tRNAsup an EcoRI site and a ribosome binding site were introduced downstream photocross-linking (Fig. 1). Where appropriate, ampicillin (100 ␮gmlϪ1) of the T1 terminator region. The T1 terminator region was cloned was added to the medium. NcoI-EcoRI into pET21d, yielding pET21d-T1. Plasmids pET21d-T1Lpp In Vitro Transcription, Translation, Targeting, and Cross-linking— and pET21d-T1BRP were obtained by plasmid PCR and site-directed Truncated mRNA was prepared as described previously from HindIII- mutagenesis using pGEM42-Lpp and pJL28, respectively, as templates linearized Lpp and BRP derivative plasmids. For photocross-linking, (17). A methionine codon was introduced at position 16 in the Lpp signal (Tmd)Phe was site-specifically incorporated into nascent chains by sup- sequence and position 17 in the BRP signal sequence (Fig. 1). In addi- pression of a UAG stop codon using (Tmd)Phe-tRNAsup in an E. coli in tion, a C-terminal 4 ϫ methionine tag was attached to both Lpp and vitro translation system containing [35S]methionine to label the nascent BRP to increase labeling efficiency. Subsequently, T1Lpp and T1BRP chains. This procedure has been described previously (5, 27). Targeting were NcoI-HindIII-cloned into pEH1 (18), pEH3 (18), and pBAD24 (19), to IMVs, photocross-linking, and carbonate extraction (to separate sol- yielding pEH1-T1Lpp, pEH3-T1Lpp, pBAD24-T1Lpp, pEH1-T1BRP, uble and peripheral membrane proteins from integral membrane pro- pEH3-T1BRP, and pBAD24-T1BRP. Finally, the genetic information teins) was carried out as described previously (27). Carbonate-soluble coding for an HA tag with a stop codon at its 3Ј prime end (20), preceded and -insoluble fractions were either trichloroacetic acid precipitated or by a flexible linker (Pro-Gly-Gly) was fused to the 4 ϫ methionine-tagged immunoprecipitated. The material used for immunoprecipitation was Lpp and BRP, using BamHI and HindIII sites. This yielded pEH1- 2-fold the amount used for trichloroacetic acid precipitation. Release of T1LppHA, pEH3-T1LppHA, pBAD24-T1LppHA, pEH1-T1BRPHA, nascent chains from the ribosome was provoked by incubating the pEH3-T1BRPHA, and pBAD24-T1BRPHA. The nucleotide sequences of nascent chains after translation for 10 min at 37 °C with EDTA (25 all constructs were verified by DNA sequencing. mM). Samples were analyzed using 15% Laemmli SDS-PAGE and phos- The Ffh conditional strain WAM121 was cultured in M9 minimal phorimaging as described previously (27). medium supplemented with 0.2% arabinose as described previously RESULTS (21). To deplete cells for Ffh, cells were grown to mid-log phase in the absence of arabinose. The 4.5 S RNA conditional strain FF283 was Model Lipoproteins—We have used the murein Lpp and the cultured in M9 minimal medium supplemented with 1 mM IPTG as pCloDF13-encoded BRP as model proteins to study the target- described previously (21). To deplete cells for 4.5 S RNA, cells were ing and translocation of secretory lipoproteins in E. coli (13, grown to mid-log phase in the absence of IPTG. The temperature- 14). To improve the labeling of Lpp and BRP with [35S]methi- sensitive amber suppressor SecA depletion strain BA13 and the control strain DO251 were cultured in M9 minimal medium at 30 °C as de- onine, both lipoproteins were slightly modified. A methionine scribed previously (22, 23). To deplete cells for SecA, cells were grown to was introduced in the signal sequences of both Lpp and BRP mid-log phase at 41 °C. The SecE depletion strain CM124 was cultured (Fig. 1). This does not have a significant impact on the pre- in M9 minimal medium supplemented with 0.2% glucose and 0.2% dicted hydrophobicity of the Lpp and BRP signal sequences, L-arabinose as described previously (24, 25). To deplete cells for SecE, and we felt confident that the introduction of a methionine cells were grown to mid-log phase in the absence of L-arabinose. The would affect Lpp and BRP signal sequence interactions only temperature-sensitive amber suppressor YidC depletion strain KO1672, along with its wild-type control strain KO1670 (26), were marginally at the most. Actually, in the ColA and ColN BRPs, cultured in M9 minimal medium at 30 °C overnight. To deplete KO1672 which are homologous to the pCloDF13 BRP, a methionine cells for YidC, cells were grown to mid-log phase at 42 °C. Where naturally occurs at this position. To further improve labeling, a Ϫ Ϫ appropriate, ampicillin (100 ␮gml 1), chloramphenicol (30 ␮gml 1), stretch of 4 methionines was attached to the very C terminus of ␮ Ϫ1 ␮ Ϫ1 kanamycin (50 gml ), and tetracyclin (12.5 gml ) were added to both Lpp and BRP. To be able to immunoprecipitate Lpp and the medium. BRP in the in vivo experiments, the 9-amino acid-long influ- In Vivo Assay for Targeting and Translocation—The model lipoproteins Lpp and BRP were expressed by L-arabinose induction from the enza virus HA epitope tag preceded by a flexible linker (Pro- pBAD24 vector in strains TOP10FЈ, FF283, BA13, DO251, KO1672, and Gly-Gly) was attached to their C termini (Fig. 1) (20). KO1670 and by IPTG induction from the pEH1 vector in strain CM124 The modified pCloDF13-encoded BRP causes lysis upon over- 31028 BRP and LPP Targeting and Translocation

FIG.1.The model lipoproteins Lpp and BRP. The amino acid sequence of BRP, Lpp, and their derivatives used in this study: wild-type BRP (WTBRP), in vitro construct 55BRPTAG10, in vivo construct BRP, wild-type Lpp (WTLPP), in vitro construct 54LPPTAG11, in vivo construct Lpp (LPP). The SPaseII cleavage site is indicated by an arrow, the lipid modifiable cystein by C, inserted/added methionines by M, and attached influenza virus hemagglutinin epitope tags with a black background. Between the methionine stretch and the HA tag there is a flexible linker (PGG). Amber codons and stop codons are marked with an asterisk. expression, just like the wild-type version of the protein, and the clipped off signal sequence is stable (results not shown). For the sake of clarity, we refer in the rest of this report to the modified lipoproteins as Lpp and BRP. Translocation of Lpp and BRP across the Inner Membrane Is Downloaded from Sec Translocase-dependent—To test whether the introduction of the extra methionines and the HA tag interfere with the in vivo processing and maturation of Lpp and BRP, the constructs were studied in the E. coli TOP10FЈ strain and Sec translocase mutant strains. Maturation of lipoproteins occurs in three steps: 1) the unmodified prolipoprotein (U-PLP) is converted into a diacylated prolipoprotein (M-PLP) upon translocation www.jbc.org across the inner membrane, 2) cleavage of the signal sequence by SPaseII yields the apolipoprotein, and 3) acylation of the FIG.2.Translocation of Lpp and BRP is affected by the SPa- lipoprotein gives rise to the mature lipoprotein (11). Both Lpp seII inhibitor globomycin and is Sec translocase-dependent. and BRP were expressed in the TOP10FЈ strain in the absence TOP10FЈ cells harboring either pBAD24-T1LppHA (top panel, lanes 1 at Stockholm Universitetsbibliotek on August 6, 2007 and presence of globomycin, which inhibits SPaseII. Inhibition and 2) or pBAD24-T1BRPHA (bottom panel, lanes 1 and 2) were cul- of SPaseII causes the accumulation of the M-PLP form of li- tured in M9 medium and pulse-labeled in the absence and presence of the SPaseII inhibitor, globomycin. CM124 (PAra-secE) cells harboring poproteins (Fig. 2, lanes 1 and 2) (28). This was confirmed in either pEH1-T1LppHA (top panel, lanes 3 and 4) or pEH1-T1BRPHA experiments where [3H]palmitate was used rather than (bottom panel, lanes 3 and 4) were cultured in M9 medium with (ϩSecE) [35S]methionine to label cells in the presence of globomycin or without (ϪSecE) L-arabinose. BA13 (SecA depletion strain, ϪSecA) ϩ (results not shown). Taken together, the modifications improve cells and DO251 (control strain, SecA) cells harboring either pBAD24- T1LppHA (top panel, lanes 5 and 6) or pBAD24-T1BRPHA (bottom labeling and facilitate immunoprecipitation of Lpp and BRP panel, lanes 5 and 6) were cultured in M9 medium at 41 °C (SecA without affecting their maturation. depletion conditions, ϪSecA). Cells were pulse-labeled and processed as Lpp and BRP were also expressed in SecE and SecA deple- described under ‘‘Materials and Methods.’’ Lpp and BRP were immu- tion strains. Both SecE and SecA are key components of the Sec noprecipitated with anti-HA antiserum. The processing of Lpp and BRP was not affected by Me2SO (DMSO) in which globomycin was dissolved. translocase. Upon depletion of SecE, SecY is rapidly degraded The processing of the SpaseII-independent OmpA was not affected in by the FtsH protease. Therefore, SecE depletion results in the the presence of globomycin, and depletion of SecE and SecA was loss of the SecY/E core of the Sec translocase (29). SecA deple- checked by monitoring the accumulation of OmpA (results not shown). tion results in the loss of the motor of the Sec translocase (22). Upon SecE and SecA depletion, the U-PLP form of both Lpp No effect of depletion of the SRP components could be detected, and BRP accumulate (Fig. 2, lanes 3–6). This points to a key confirming that the observed accumulation of U-PLP is not role of the Sec translocase in the translocation of Lpp and BRP because of more general secondary effects of SRP depletion. across the inner membrane, corroborating previous studies Together, the results indicate that a functional SRP pathway is using Sec-conditional mutant strains (30–33). required for efficient targeting of the BRP and, albeit to a lesser Efficient in Vivo Targeting of Lpp and BRP Requires SRP— extent, of the Lpp. How are Lpp and BRP targeted to the Sec translocase? The Efficient in Vivo Translocation of Lpp and BRP Requires SecB and the SRP pathways are the two main targeting path- YidC—All IMPs studied so far require YidC for efficient assem- ways to the Sec translocase (34). In the absence of SecB, tar- bly into the inner membrane. It has been suggested that YidC geting of both Lpp and BRP is not significantly affected (results assists the transfer of TMs from the Sec translocase into the not shown). lipid bilayer (5). So far, no evidence has been obtained pointing We next investigated the role of the SRP in the targeting of to a role of YidC in the translocation of secretory proteins (6, 7, Lpp and BRP to the Sec translocase. The E. coli SRP consists of 27, 37, 38). However, the unexpected role of the SRP in the the protein component Ffh and the RNA component 4.5 S RNA. targeting of BRP and Lpp prompted us to evaluate the role of Both Ffh and 4.5 S RNA are essential for viability, and deple- YidC in the translocation of these proteins using a tempera- tion of either of the SRP components compromises the SRP ture-sensitive strain that is conditional for YidC expression targeting pathway, thereby preventing the targeting of many (Fig. 4). To our surprise, depletion of YidC by growth at the IMPs (3, 35, 36). Targeting of Lpp and BRP was studied both non-permissive temperature resulted in the accumulation of under 4.5 S RNA and Ffh depletion conditions (Fig. 3, A and B). unmodified precursor forms of BRP (most pronounced) and Lpp Depletion of 4.5 S RNA (Fig. 3A) and of Ffh (Fig. 3B) both (less pronounced). Again, the processing of pro-OmpA that is resulted in accumulation of the unmodified precursor forms of translocated independent of YidC (7) was monitored as a con- BRP (most pronounced) and Lpp (less pronounced). As a control and appeared unaffected. The combined data suggest that trol, the processing of pro-OmpA, an outer membrane protein YidC plays a differential role in the translocation of both that is targeted by SecB, was monitored in the same samples. lipoproteins. BRP and LPP Targeting and Translocation 31029

signal sequence in the nascent Lpp and BRP species, a single amber stop codon (TAG) was introduced in the center of the hydrophobic core in the Lpp (position 11) and BRP (position 10) signal sequences (Fig. 1). The amber stop codons were suppressed during the in vitro translation by addition of (Tmd- )Phe-tRNAsup, an amber suppressor tRNA that is amino-acylated with the photo cross-linker (Tmd)Phe. In all constructs, the TAGs were efficiently suppressed by (Tmd)Phe-tRNAsup (data not shown). Purified inverted IMVs were added from the start of the translation reaction to allow cotranslational membrane targeting and interaction of the translation intermedi- ates with the membrane. After the translation/insertion reaction, one half of each sample was irradiated with UV light to induce cross-linking; the other half was kept in the dark to serve as a control. The samples were extracted with carbonate to separate soluble and peripherally membrane-associated material from membrane-integrated components. Cross-linking

FIG.3.The SRP pathway is required for efficient targeting of partners were identified by immunoprecipitation. Without UV Downloaded from Lpp and BRP. A, FF283 (P -4.5 S RNA) cells were cultured in M9 IPTG irradiation, no cross-linking products were detected using both medium with (ϩ4.5 S RNA) or without (Ϫ4.5 S RNA) IPTG. Cells were pulse-labeled and processed as described under ‘‘Materials and Meth- Lpp and BRP nascent chains (data not shown). Upon transla- ods.’’ Lpp and BRP were immunoprecipitated with anti-HA antiserum tion but prior to UV irradiation, the samples were divided in and OmpA with OmpA antiserum. OmpA processing was not affected two, and EDTA was added to one aliquot to provoke the release upon 4.5 S RNA depletion. B, WAM121 (P -ffh) cells harboring either ara of the nascent chains from the ribosome. This allowed us to pEH3T1LppHA or pEH3T1BRPHA were cultured in M9 medium with

(ϩFfh) or without (ϪFfh) arabinose. Cells were pulse-labeled and pro- assess the importance of the context of the ribosome for cross- www.jbc.org cessed as described under ‘‘Materials and Methods.’’ Lpp and BRP were linking to the truncated Lpp and BRP species. EDTA has been immunoprecipitated with anti-HA antiserum and OmpA with OmpA shown to disassemble ribosomes (37). antiserum. OmpA processing was not affected upon Ffh depletion. Both nascent Lpp and BRP were efficiently targeted to the ϳ

IMVs, judging from the relatively high ( 50%) carbonate re- at Stockholm Universitetsbibliotek on August 6, 2007 sistance. When the carbonate supernatant of UV-irradiated samples was analyzed, Lpp nascent chains were shown to cross-link Ffh, albeit inefficiently (Fig. 5A, lane 3), and the chaperone trigger factor (TF, lane 4). The ϳ30-kDa cross-linking adducts represented cross-linking to a breakdown product of Ffh, as observed before (39). Cross-linking to both Ffh and TF appeared dependent on the context of the ribosome (lanes 8, 9) consistent with earlier studies (39, 40). In contrast, cross-linking to SecA was hardly detectable unless the nascent chains were released from the ribosomes prior to cross-linking (lanes 7, 10). Strong cross- linking specific for the ribosome-associated Lpp was observed to the ribosomal proteins L23 and L29 (lanes 5, 6). L23 and L29 are located near the exit site of the large ribosomal tunnel that runs from the peptidyl transferase center to the surface of the large ribosomal subunit (41). In the carbonate pellet, cross- linking to SecY and (weakly) to YidC was observed that ap- FIG.4.YidC is required for efficient targeting of Lpp and BRP. peared dependent on the ribosomal context (lanes 13, 14, 16, KO1670 (control strain, ϩYidC) cells and KO1672 (YidC depletion strain, ϪYidC) cells were cultured in M9 medium at 42 °C (YidC de- 17), again consistent with earlier studies (5, 37). In addition, in pletion conditions) and 30 °C (non-depletion conditions). Cells were the pellet fractions SecA cross-linking was detectable only upon pulse-labeled and processed as described under ‘‘Materials and Meth- release of nascent Lpp from the ribosome. Together, the data ods.’’ Lpp and BRP were immunoprecipitated with anti-HA antiserum suggest that nascent Lpp leaves the ribosome via the major exit and OmpA with OmpA antiserum. OmpA processing was not affected upon YidC depletion. YidC depletion was monitored by means of West- tunnel near L23 and L29. Most likely, the signal sequence of a ern blotting and resulted, as expected, in a strong PspA response (Ref. small fraction of nascent Lpp contacts the SRP. However, the 56 and results not shown). majority of nascent Lpp is close to TF. Consistent with this explanation, both the SRP and TF dock near L23/L29 on the Nascent Lpp and BRP Synthesized in Vitro Cross-link to Ffh, ribosome, probably in a mutually exclusive manner. Upon SecA, SecY, and YidC—To study the targeting and transloca- forced release of the nascent chains from the ribosome, the tion of Lpp and BRP in more detail, we have used an in vitro signal peptide loses contact with L23, L29, TF, and Ffh and is translation/photo cross-linking approach. In this assay, the free to bind SecA, part of which is carbonate-resistant. Nascent interactions of nascent (ribosome-associated) polypeptides with Lpp is primarily targeted to SecY but also contacts YidC. cytosolic and membrane components are fixed and analyzed. Qualitatively, very similar results were obtained when nas- [35S]Methionine-radiolabeled nascent chains of Lpp and BRP cent BRP was used instead of Lpp (Fig. 5B). However, cross- were synthesized in an E. coli cell-free extract from truncated linking to Ffh appeared much more prominent at the expense mRNA to a length of 55 amino acids. Assuming that the ribo- of cross-linking to TF (lanes 1, 3, 7), especially when consider- some covers ϳ35 amino acids, the Lpp and BRP signal se- ing the small amount of SRP present in cells as compared with quences are expected to be exposed just outside the ribosome TF (42, 43). Upon treatment with EDTA, cross-linking to Ffh (37). To specifically probe the molecular environment of the was no longer detectable (lane 8), but another unknown factor 31030 BRP and LPP Targeting and Translocation

FIG.5.Interactions of nascent Lpp and BRP. A, 55LppTAG11 was trans-

lated in the presence of IMVs and (Tmd)- Downloaded from Phe-tRNAsup as described under ‘‘Materi- als and Methods.’’ After translation, the nascent chains were treated with EDTA when indicated, UV irradiated, and subsequently extracted with sodium carbonate (Supernatant, lanes 1–10; Pellet, lanes 11–18). UV-irradiated fractions were immunoprecipitated using anti- www.jbc.org serum against Ffh, TF, L23, L29, SecA, SecY, and YidC (lanes 3–10, 13–18). All cross-linking adducts are indicated with Ͼ. B, in vitro translation, EDTA treatment, cross-linking, and carbonate ex- at Stockholm Universitetsbibliotek on August 6, 2007 traction of nascent 55BRPTAG10 were carried out as described under panel A.

of about the same molecular mass (ϳ50 kDa) was cross-linked these in vitro results, BRP and, to a lesser extent, Lpp depend (lane 2). Other notable differences were that ribosome-associ- on the presence of the SRP and YidC for efficient targeting to ated BRP also detectably cross-linked SecA (lane 1, 6), and and translocation across the inner membrane in vivo. there was stronger cross-linking of targeted nascent BRP to In vitro, Lpp and BRP nascent chains with a length of 55 YidC (lane 11). Finally, a cross-linking adduct of 75 kDa was amino acids, carrying a UV-inducible cross-linker in the middle observed in the carbonate pellet fractions both before and after of the signal sequence, are also cross-linked to the ribosomal release of nascent BRP from the ribosome, which remains to be components L23 and L29 and to the ribosome-associated chap- identified (lanes 11, 12). The more prominent contacts with the erone TF. L23 and L29 are located near the exit of the pre- SRP and YidC suggest a more important role for these factors sumed ribosomal tunnel (41) and have been cross-linked to in the targeting and translocation of BRP, corroborating the short nascent chains of other origin before (39, 44). L23 has in vivo data. recently been shown to function as an attachment site for both DISCUSSION TF and SRP (39, 45, 46). Cross-link studies have identified TF Here, we have studied in E. coli the targeting and translo- as the first chaperone to interact generically with nascent cation of two secretory lipoproteins, Lpp and BRP, using a polypeptides (47) unless they carry a particularly hydrophobic combined in vivo and in vitro approach. Surprisingly, the signal targeting signal that has a high affinity for the SRP (39, 40, 48). peptides of both nascent BRP and, to a lesser extent, Lpp show The mechanism that underlies the interplay between TF and cross-linking to the targeting factor SRP and to the Sec-asso- SRP at the nascent chain exit site and how this interplay ciated YidC, which are thought to function in the membrane influences the mode of membrane targeting and insertion of a targeting and integration of integral IMPs. Consistent with particular nascent protein are still unresolved issues (39, 44). BRP and LPP Targeting and Translocation 31031

The SRP is primarily used for the targeting of IMPs that are in vivo depletion of YidC affects the maturation of BRP and thought to benefit from a cotranslational insertion mechanism Lpp. In the context of the Sec translocon, YidC is considered to to prevent aggregation of hydrophobic domains in the cyto- facilitate the transfer of TMs of IMPs from the Sec translocase plasm. Then why does the SRP appear to play such an impor- into the lipid bilayer without being essential for this process. tant role in the targeting of the secretory lipoprotein BRP? The role of YidC in the targeting/translocation of Lpp and BRP First, the BRP signal sequence is very hydrophobic, more hy- is enigmatic. It is possible that YidC facilitates the lateral drophobic than many signal anchor sequences of SRP-depend- movement of the Lpp and BRP signal sequences into the lipid ent IMPs and prone to aggregation if unprotected. Interaction bilayer or that YidC chaperones Lpp and BRP to the SPaseII/ of the signal sequence with the SRP and cotranslational tar- Lol system. It is also possible that there is a direct connection geting of BRP may be the best way to prevent uncontrolled between SRP dependence and YidC dependence. It has been insertion of the hydrophobic BRP signal sequence into the shown that the chloroplast homologues of SRP and YidC par- inner membrane and aggregation of BRP in the cytoplasm. ticipate in targeting complexes (53). Therefore, it cannot be Recently, it has been suggested that basic amino acids in the N excluded that YidC plays a role in the SRP cycle, perhaps in the region of a signal sequence contribute to SRP binding, probably reception of the SRP or FtsY. Interestingly, membranes iso- through the formation of salt bridges between the 4.5 S RNA lated from cells depleted of YidC show decreased levels of the and positively charged amino acids in the N region (49). Strik- lipoprotein CyoA, which in contrast to Lpp and BRP is an ingly, the BRP signal sequence has 3 lysines in its N region, integral IMP (54, 55). This may point to a more general and not

which may enhance even more the affinity of the already very yet understood role of YidC in the targeting/translocation of Downloaded from hydrophobic BRP signal sequence for the SRP. These features lipoproteins. In conclusion, our results indicate that the SRP/ may compensate for the relatively short time window in which Sec/YidC pathway is used by the secretory lipoproteins Lpp the SRP can interact with nascent BRP, given that the BRP is and BRP. a very small protein. It should be noted that the BRP signal REFERENCES peptide is peculiar in the sense that it is stable in the inner 1. Manting, E. H., and Driessen, A. J. (2000) Mol. Microbiol. 37, 226–238 membrane, whereas other cleaved signal peptides are rapidly 2. Herskovits, A. A., Bochkareva, E. S., and Bibi, E. (2000) Mol. Microbiol. 38, www.jbc.org degraded upon cleavage from the precursor protein. Appar- 927–939 3. de Gier, J. W., and Luirink, J. (2001) Mol. Microbiol. 40, 314–322 ently, the BRP signal peptide is recognized as a signal anchor 4. Van den Berg, B., Clemons, W. M., Jr., Collinson, I., Modis, Y., Hartmann, E., sequence of an IMP: it binds the SRP in the cytosol, inserts at Harrison, S. C., and Rapoport, T. A. (2004) Nature 427, 36–44 5. Urbanus, M. L., Scotti, P. A., Froderberg, L., Saaf, A., de Gier, J. W., Brunner, the Sec translocon, and is subsequently transferred to the lipid J., Samuelson, J. C., Dalbey, R. E., Oudega, B., and Luirink, J. (2001) at Stockholm Universitetsbibliotek on August 6, 2007 bilayer as a stably folded unit assisted by YidC (see below). EMBO Rep. 2, 524–529 The signal sequence of Lpp is not very hydrophobic, and its N 6. Beck, K., Eisner, G., Trescher, D., Dalbey, R. E., Brunner, J., and Mu¨ ller, M. (2001) EMBO Rep. 2, 709–714 region contains only one positively charged amino acid. How- 7. Samuelson, J. C., Chen, M., Jiang, F., Moller, I., Wiedmann, M., Kuhn, A., ever, it does show (weak) SRP cross-linking and SRP depend- Phillips, G. J., and Dalbey, R. E. (2000) Nature 406, 637–641 8. Serek, J., Bauer-Manz, G., Struhalla, G., Van Den Berg, L., Kiefer, D., Dalbey, ence in vivo. There are recent precedents of relatively non- R., and Kuhn, A. (2004) EMBO J. 23, 294–301 hydrophobic signal peptides that are yet able to funnel 9. Sijbrandi, R., Urbanus, M. L., Ten Hagen-Jongman, C. M., Bernstein, H. D., passenger proteins into the SRP pathway (50). Therefore, it is Oudega, B., Otto, B. R., and Luirink, J. (2003) J. Biol. Chem. 278, 4654–4659 not unlikely that there are yet unknown features of a signal 10. Juncker, A. S., Willenbrock, H., Von Heijne, G., Brunak, S., Nielsen, H., and sequence that can provoke SRP binding. There may be an Krogh, A. (2003) Protein Sci. 12, 1652–1662 11. Sankaran, K., and Wu, H. C. (1994) J. Biol. Chem. 269, 19701–19706 important biological reason for a preference of Lpp for the 12. Hara, T., Matsuyama, S., and Tokuda, H. (2003) J. Biol. Chem. 278, SRP-targeting pathway. Lpp is the most abundant protein in E. 40408–40414 coli and, therefore, highly expressed. When the mature part of 13. Shu, W., Liu, J., Ji, H., and Lu, M. (2000) J. Mol. Biol. 299, 1101–1112 14. van der Wal, F. J., Luirink, J., and Oudega, B. (1995) FEMS Microbiol. Rev. 17, Lpp is expressed in the cytoplasm it forms stable trimers (13). 381–399 It is conceivable that cotranslational targeting of Lpp via the 15. van den Elzen, P. J., Walters, H. H., Veltkamp, E., and Nijkamp, H. J. (1983) Nucleic Acids Res. 11, 2465–2477 SRP pathway prevents the formation of Lpp trimers in the 16. Luirink, J., Duim, B., de Gier, J. W., and Oudega, B. (1991) Mol. Microbiol. 5, cytoplasm, which would make it incompetent for translocation. 393–399 The partial effect of SRP depletion suggests that Lpp can also 17. Valent, Q. A., Kendall, D. A., High, S., Kusters, R., Oudega, B., and Luirink, J. (1995) EMBO J. 14, 5494–5505 travel via SecB. The lack of SecB did not significantly affect the 18. Hashemzadeh-Bonehi, L., Mehraein-Ghomi, F., Mitsopoulos, C., Jacob, J. P., kinetics of Lpp translocation (32).3 However, pre-Lpp could be Hennessey, E. S., and Broome-Smith, J. K. (1998) Mol. Microbiol. 30, 676–678 detected in cytosolic aggregates isolated from a SecB null 19. Siegele, D. A., and Hu, J. C. (1997) Proc. Natl. Acad. Sci. U. S. A. 94, 3 strain. A similar flexibility in targeting has recently been 8168–8172 demonstrated for the autotransporter Hbp (9). Therefore, it is 20. Chen, Y. T., Holcomb, C., and Moore, H. P. (1993) Proc. Natl. Acad. Sci. U. S. A. 90, 6508–6512 not unlikely that when the capacity of the SRP pathway is 21. de Gier, J. W., Mansournia, P., Valent, Q. A., Phillips, G. J., Luirink, J., and insufficient, the SecB pathway gets a more prominent role in von Heijne, G. (1996) FEBS Lett. 399, 307–309 22. Qi, H. Y., and Bernstein, H. D. (1999) J. Biol. Chem. 274, 8993–8997 the targeting of Lpp and, perhaps, also BRP. Unfortunately, we 23. Froderberg, L., Houben, E., Samuelson, J. C., Chen, M., Park, S. K., Phillips, have not succeeded in studying the targeting of Lpp and BRP in G. J., Dalbey, R., Luirink, J., and De Gier, J. W. (2003) Mol. Microbiol. 47, an Ffh depletion/SecB null mutant background. Therefore, 1015–1027 24. Traxler, B., and Murphy, C. (1996) J. Biol. Chem. 271, 12394–12400 other modes of targeting, like spontaneous or mRNA targeting, 25. de Gier, J. W., Scotti, P. A., Sa¨a¨f, A., Valent, Q. A., Kuhn, A., Luirink, J., and cannot be excluded at present (3, 51). von Heijne, G. (1998) Proc. Natl. Acad. Sci. U. S. A. 95, 14646–14651 26. Hansen, F. G., Hansen, E. B., and Atlung, T. (1985) Gene 38, 85–93 Cross-linking of Lpp and BRP nascent chains to SecA and 27. Scotti, P. A., Urbanus, M. L., Brunner, J., de Gier, J. W., von Heijne, G., van SecY and the almost completely blocked maturation of both der Does, C., Driessen, A. J., Oudega, B., and Luirink, J. (2000) EMBO J. lipoproteins under SecA and SecE depletion conditions clearly 19, 542–549 28. Inukai, M., Takeuchi, M., Shimizu, K., and Arai, M. (1978) J. Antibiot. (Tokyo) show that both Lpp and BRP are translocated by the Sec 31, 1203–1205 translocase across the inner membrane, corroborating previous 29. Akiyama, Y., Kihara, A., Tokuda, H., and Ito, K. (1996) J. Biol. Chem. 271, 31196–31201 in vivo studies (30–33, 52). Surprisingly, both BRP and, to 30. Hayashi, S., and Wu, H. C. (1985) J. Bacteriol. 161, 949–954 some extent, Lpp nascent chains are cross-linked to YidC, and 31. Oudega, B., Mol, O., van Ulsen, P., Stegehuis, F., van der Wal, F. J., and Luirink, J. (1993) J. Bacteriol. 175, 1543–1547 32. Sugai, M., and Wu, H. C. (1992) J. Bacteriol. 174, 2511–2516 33. Watanabe, T., Hayashi, S., and Wu, H. C. (1988) J. Bacteriol. 170, 4001–4007 3 L. Baars and J.-W. de Gier, manuscript in preparation. 34. Valent, Q. A., Scotti, P. A., High, S., de Gier, J. W., von Heijne, G., Lentzen, G., 31032 BRP and LPP Targeting and Translocation

Wintermeyer, W., Oudega, B., and Luirink, J. (1998) EMBO J. 17, Specking, A., Ban, N., Deuerling, E., and Bukau, B. (2002) Nature 419, 2504–2512 171–174 35. Ribes, V., Romisch, K., Giner, A., Dobberstein, B., and Tollervey, D. (1990) Cell 47. Patzelt, H., Rudiger, S., Brehmer, D., Kramer, G., Vorderwulbecke, S., Schaf- 63, 591–600 fitzel, E., Waitz, A., Hesterkamp, T., Dong, L., Schneider-Mergener, J., 36. Phillips, G. J., and Silhavy, T. J. (1992) Nature 359, 744–746 Bukau, B., and Deuerling, E. (2001) Proc. Natl. Acad. Sci. U. S. A. 98, 37. Houben, E. N., Urbanus, M. L., Van Der Laan, M., Ten Hagen-Jongman, C. M., 14244–14249 Driessen, A. J., Brunner, J., Oudega, B., and Luirink, J. (2002) J. Biol. 48. Valent, Q. A., de Gier, J. W., von Heijne, G., Kendall, D. A., ten Hagen- Chem. 277, 35880–35886 Jongman, C. M., Oudega, B., and Luirink, J. (1997) Mol. Microbiol. 25, 38. Drew, D., Fro¨derberg, L., Baars, L., and de Gier, J. W. (2003) Biochim Biophys 53–64 Acta 1610, 3–10 49. Peterson, J. H., Woolhead, C. A., and Bernstein, H. D. (2003) J. Biol. Chem. 39. Ullers, R. S., Houben, E. N., Raine, A., ten Hagen-Jongman, C. M., Ehrenberg, 50. Schierle, C. F., Berkmen, M., Huber, D., Kumamoto, C., Boyd, D., and Beck- M., Brunner, J., Oudega, B., Harms, N., and Luirink, J. (2003) J. Cell Biol. with, J. (2003) J. Bacteriol. 185, 5706–5713 161, 679–684 51. Bernstein, H. D., and Hyndman, J. B. (2001) J. Bacteriol. 183, 2187–2197 40. Beck, K., Wu, L. F., Brunner, J., and Mu¨ ller, M. (2000) EMBO J. 19, 134–143 52. Tian, G., Wu, H. C., Ray, P. H., and Tai, P. C. (1989) J. Bacteriol. 171, 41. Ban, N., Nissen, P., Hansen, J., Moore, P. B., and Steitz, T. A. (2000) Science 1987–1997 289, 905–920 53. Moore, M., Goforth, R. L., Mori, H., and Henry, R. (2003) J. Cell Biol. 162, 42. Lill, R., Crooke, E., Guthrie, B., and Wickner, W. (1988) Cell 54, 1013–1018 1245–1254 43. Jensen, C. G., and Pedersen, S. (1994) J. Bacteriol. 176, 7148–7154 54. Chepuri, V., Lemieux, L., Au, D. C., and Gennis, R. B. (1990) J. Biol. Chem. 44. Eisner, G., Koch, H. G., Beck, K., Brunner, J., and Muller, M. (2003) J. Cell 265, 11185–11192 Biol. 163, 35–44 55. Ma, J., Katsonouri, A., and Gennis, R. B. (1997) Biochemistry 36, 11298–11303 45. Gu, S. Q., Peske, F., Wieden, H. J., Rodnina, M. V., and Wintermeyer, W. 56. van der Laan, M., Urbanus, M. L., Ten Hagen-Jongman, C. M., Nouwen, N., (2003) RNA 9, 566–573 Oudega, B., Harms, N., Driessen, A. J., and Luirink, J. (2003) Proc. Natl. 46. Kramer, G., Rauch, T., Rist, W., Vorderwulbecke, S., Patzelt, H., Schulze- Acad. Sci. U. S. A. 100, 5801–5806 Downloaded from www.jbc.org at Stockholm Universitetsbibliotek on August 6, 2007 Supplemental Material can be found at: http://www.jbc.org/cgi/content/full/M509929200/DC1

Defining the Role of the Escherichia coli Chaperone SecB Using Comparative Proteomics*□S Received for publication, September 8, 2005, and in revised form, December 9, 2005 Published, JBC Papers in Press, December 13, 2005, DOI 10.1074/jbc.M509929200 Louise Baars‡, A. Jimmy Ytterberg§, David Drew‡, Samuel Wagner‡, Claudia Thilo¶, Klaas Jan van Wijk§, and Jan-Willem de Gier‡1 From the ‡Department of Biochemistry and Biophysics, Arrhenius Laboratories, Stockholm University SE-106 91 Stockholm, Sweden, the §Department of Plant Biology, Cornell University, Ithaca, New York 14853, and the ¶Center for Infectious Medicine, Karolinska Institute, Karolinska University Hospital Huddinge, SE-141 86 Stockholm, Sweden

To improve understanding and identify novel substrates of the OmpF, GBP, and OmpA), whereas four secretory proteins (PhoA, Lpp, cytoplasmic chaperone SecB in Escherichia coli, we analyzed a secB RbsB, and ␤-lac) do not seem to require SecB (9–12, 55). SecB also has null mutant using comparative proteomics. The secB null mutation the capacity to assist the chaperone DnaK in the folding of proteins, as Downloaded from did not affect cell growth but caused significant differences at the shown in vitro with luciferase as a model substrate (12). This indicates proteome level. In the absence of SecB, dynamic protein aggregates that SecB has the potential to assist the folding of cytoplasmic proteins. containing predominantly secretory proteins accumulated in the The successful complementation of a DnaK/trigger factor (TF)2 double cytoplasm. Unprocessed secretory proteins were detected in radio- mutant strain by overexpression of SecB, and cross-linking of SecB to labeled whole cell lysates. Furthermore, the assembly of a large frac- nascent chains of both secretory and cytoplasmic proteins in SecB-

tion of the outer membrane proteome was slowed down, whereas its enriched lysates support this notion (13). www.jbc.org steady state composition was hardly affected. In response to aggre- SecB does not bind to signal sequences and peptide library screens gation and delayed sorting of secretory proteins, cytoplasmic chap- suggested a very loosely defined SecB binding “motif” (12). This motif, erones DnaK, GroEL/ES, ClpB, IbpA/B, and HslU were up-regu- which is ϳ9 residues long, is enriched in aromatic and basic residues,

lated severalfold, most likely to stabilize secretory proteins during whereas acidic residues are disfavored. It theoretically occurs every at Stockholm Universitetsbibliotek on August 6, 2007 their delayed translocation and/or rescue aggregated secretory pro- 20–30 residues in both secretory and cytoplasmic proteins and is too teins. The SecB/A dependence of 12 secretory proteins affected by unspecific to facilitate genome-wide prediction of SecB substrates (10– the secB null mutation (DegP, FhuA, FkpA, OmpT, OmpX, OppA, 12). Thus experimentation is needed to identify novel SecB substrates. TolB, TolC, YbgF, YcgK, YgiW, and YncE) was confirmed by “clas- To characterize the role of SecB in more detail and identify additional sical” pulse-labeling experiments. Our study more than triples the SecB substrates, we analyzed a secB null mutant using comparative number of known SecB-dependent secretory proteins and shows proteomics. This analysis included flow cytometry, pulse labeling that the primary role of SecB is to facilitate the targeting of secretory combined with cell fractionation, one- and two-dimensional gel elec- proteins to the Sec-translocase. trophoresis, and mass spectrometry (MS), complemented by immunoblotting. The comparative proteomics approach allowed us to investigate protein mistargeting, aggregation, and translocation kinetics, and The periplasmic and outer membrane proteins in the Gram-negative to determine changes in the proteome composition. Our analysis bacterium Escherichia coli need to cross the cytoplasmic membrane showed that, although the secB null mutation did not affect cell growth, to reach their final destination. The vast majority of these secretory there are significant differences at the proteome level. Most differences proteins are translocated through the cytoplasmic membrane via the pointed to protein targeting defects, resulting in a protein folding/ag- Sec-translocase (1, 2). The core of the Sec-translocase is comprised gregation problem in the cytoplasm. Careful analysis of the (sub)pro- of integral membrane proteins SecY and SecE, which form a protein teome(s) of the secB null mutant strain combined with a classical pulse- conducting channel (3). The peripheral subunit SecA drives poly- labeling approach enabled us to more than triple the number of known peptide chains in an ATP-dependent manner into and through the SecB-dependent secretory proteins. Sec-translocase (1). EXPERIMENTAL PROCEDURES It is generally assumed that secretory proteins in E. coli are targeted to the Sec-translocase by the cytoplasmic protein SecB in a mostly post- Strains and Culture Conditions—We used E. coli strain EK413, which translational fashion (4–8). However, direct evidence for SecB depend- is a MC4100 derivative that is araϩ (a kind gift from Ken-ichi Nish- ence is only established for six secretory proteins (PhoE, LamB, MBP, iyama), harboring plasmid pE63 as wild-type. Plasmid pE63 harbors the gpsA gene, which encodes for sn-glycerol-3-phosphate dehydrogenase, under control of an arabinose inducible promotor and has a pSC101 * This work was supported by grants from the Swedish Research Council, Carl Tryggers origin of replication and a ␤-lactamase resistance marker (14). Using P1 Stiftelse, Marianne and Marcus Wallenberg Foundation, and the EMBO Young Inves- tigator Programme (to J.-W. d. G.), a grant from The Swedish Foundation for Inter- transduction, we moved the secB null mutation secB8 (15) from strain national Cooperation in Research and Higher Education (STINT) (to J.-W. d. G. and HS101/pE63 into EK413/pE63, yielding an EK413/pE63-derived secB K. J. v. W.), and proteomics infrastructure was supported by a grant from New York State Office of Science, Technology and Academic Research (to K. J. v. W.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. 2 The abbreviations used are: TF, trigger factor; MS, mass spectrometry; CHAPS, 3-[(3- Section 1734 solely to indicate this fact. cholamidopropyl)dimethylammonio]-1-propanesulfonic acid; bis-Tris, 2-[bis(2-hy- □ S The on-line version of this article (available at http://www.jbc.org) contains supple- droxyethyl)amino]-2-(hydroxymethyl)propane-1,3-diol; Tricine, 2-{[2-hydroxy-1,1- mental Tables S1–S4. bis(hydroxymethyl)ethyl]amino}ethanesulfonic acid; Ibp, inclusion body associated 1 To whom correspondence should be addressed. Tel.: 46-8-162420; Fax: 46-8-153679; protein; MALDI-TOF, matrix-assisted laser desorption ionization time-of-flight; IPG, E-mail: [email protected]. immobilized pH gradient; SRP, signal recognition particle.

10024 JOURNAL OF BIOLOGICAL CHEMISTRY VOLUME 281•NUMBER 15•APRIL 14, 2006 Defining the Role of E. coli SecB null mutant strain. As expected, the secB null mutant is unable to form of cells. Lysis was achieved by incubation at 100 °C in a water bath for 2 single colonies on LB plates in the absence of arabinose; i.e. when GspA min. The samples were then cooled on ice for 10 min before addition of is not expressed (14). Hereafter, we will refer to this EK413/pE63-de- 4 ␮l of DNase/RNase solution (1 mg/ml DNase I, 0.25 mg/ml RNase A, rived secB null mutant strain and EK413/pE63 as the secB null mutant 476 mM Tris-HCl (pH 8), 24 mM Tris base, 50 mM MgCl2 and deionized and the control strains, respectively. water) per 1 A600 unit of cells. The samples were incubated on ice for 10 Cells were cultured in standard M9 medium supplemented with thi- min and were then immediately used for isoelectric focusing as amine (10 mM), all amino acids but methionine and cysteine, glucose described below. (0.2% w/v), arabinose (0.2% w/v), and ampicillin (100 ␮g/ml). Overnight Preparation of Radiolabeled Membranes—Cells corresponding to cultures were diluted 1:50 in pre-warmed medium and cultured at 200 A600 units were cultured as described above. At the time of harvest- ing, an aliquot of 2 A units of cells was labeled with [35S]methionine 37 °C. Growth was monitored by measuring the A600 with a Shimadzu 600 UV-1601 spectrophotometer. Under these conditions, we did not (60 ␮Ci/ml, Ci ϭ 37 GBq) for 1 min. An excess of cold methionine (final concentration 1 mg/ml) was added and cells were collected by centrif- observe differences in growth (as monitored by A600 measurements) between the secB null mutant and the control (results not shown). For all ugation either directly after labeling or after a 10-min chase. The remaining unlabeled cells were washed and collected by centrifugation. the experiments, cells were harvested at an A600 of 1.0 (i.e. in the early exponential phase). Before breaking the cells, labeled and unlabeled cells from the same Flow Cytometry—Analysis of the secB null mutant and the control by culture were pooled back together resulting in a mixture of labeled and Downloaded from means of flow cytometry was done using a FACSCalibur (BD Bio- unlabeled cells with a ratio of 1:100 that was then used for membrane sciences) instrument. Cultures of the secB null mutant and the control isolations. Carbonate-washed total membranes (i.e. a mixture of inner were immediately diluted in ice-cold phosphate-buffered saline to a and outer membranes) were isolated essentially as described by Molloy final concentration of ϳ106 cells per ml, and analyzed with an average et al. (23), with the exception that we used sonication rather than French flow rate of 400 events/s. Forward and side scatters were measured and pressing to break cells. Protein concentrations were determined with the BCA assay (Pierce) according to the instructions of the manufacturer. used for comparison of cell morphology of the secB null mutant and control www.jbc.org (16). Propidium iodide staining was performed to assess viability (16). Two-dimensional Gel Electrophoresis—The analysis of stained two- dimensional electrophoresis gels of whole cell lysates was first done on Immunoblot Analysis—The protein accumulation of SecB, SecY, gels with a low protein load (0.5 A units of cells) to avoid saturation SecE, SecA, Ffh, PspA, TF, GroEL, DnaK, and IbpB (it should be noted 600

and allow analysis of highly abundant proteins, and then on gels with a at Stockholm Universitetsbibliotek on August 6, 2007 that the IbpB antiserum cross-reacts with IbpA) in the secB null mutant high protein load (1 A unit of cells) for the analysis of low abundant strain and the control strain were determined by immunoblot analysis. 600 proteins. 1 A unit of cells was used for the analysis of [35S]methi- Cells were cultured as described above. Cells (0.2 A units) or inner 600 600 onine-labeled whole cell lysates. Whole cell lysates were solubilized in 9 membranes (5 ␮g of protein) isolated by sucrose gradient centrifugation M urea, 4% (w/v) CHAPS, 2 mM tributylphosphine, 0.5% (v/v) Triton (17, 18) were solubilized in Laemmli solubilization buffer. Proteins were X-100, 5% glycerol, 2% (v/v) immobilized pH gradient gel (IPG) buffer separated by SDS-PAGE. Blotting, immunodecoration, detection, and for pH 4–7 (Amersham Biosciences) and bromphenol blue. For analysis quantification of blots were done as described previously (19). of the outer membrane proteome, 350 ␮g of protein was solubilized in 7 Protein Translocation Assays in Vivo—Protein translocation assays M urea, 2 M thiourea, 1% (w/v) ASB-14, 2 mM tributylphosphine, 5% were done with 1 ml of culture each. Cells were labeled with [35S]me- glycerol, 2% (v/v) IPG buffer for pH 4–7 (Amersham Biosciences) and thionine (60 ␮Ci/ml, Ci ϭ 37 GBq) for 45 s and subsequently precipi- bromphenol blue (23). Unsolubilized material was removed by centrif- tated in 10% trichloroacetic acid. Trichloroacetic acid-precipitated ugation at 14,000 ϫ g for 30 min. The clarified protein solution was used samples were washed with acetone, resuspended in 10 mM Tris-HCl to re-swell Immobilin DryStrips, pH 4–7 (Amersham Biosciences), (pH 7.5), 2% SDS, and immunoprecipitated with antisera to OmpA and overnight at room temperature. Isoelectric focusing was subsequently ␤ -lactamase, followed by standard SDS-PAGE analysis (21). Gels were performed at 20 °C in a Multiphor II apparatus (Amersham Bio- scanned in a Fuji FLA-3000 phosphorimager and quantified using the sciences); whole cell samples at 80 kVh and membrane samples at 60 Image Gauge software (version 3.4). Potential SecB-dependent secre- kVh at a maximum 3,500 V. Proteins were separated in the second tory proteins were C-terminal hemagglutinin-tagged and expressed by dimension on 10% duracrylamide (Genomic Solutions) gels (10% acryl- ␤ isopropyl 1-thio- -D-galactopyranoside induction from the pEH1 vec- amide monomer and 1% bisacrylamide) containing 1 M Tris-HCl (pH tor as described previously (20), and the protein translocation assay was 8.45), 0.1% (w/v) SDS, and 20% (v/v) glycerol. After focusing, proteins in performed as described above except that antiserum to the hemagglu- the IPG strips were reduced and alkylated, as described before (24). The tinin tag was used for immunoprecipitations. strips were loaded on top of the second dimension gel by submerging Preparation of Whole Cell Lysates for Two-dimensional Gel Electro- the strips in warm agarose solution (1% (w/v) low melting agarose, 0.2% phoresis—Cells were cultured as described above. Radiolabeled cells SDS, 150 mM bis-Tris, 80 mM HCl and bromphenol blue). Electrophore- 35 were cultured in the presence of [ S]methionine (60 ␮Ci/ml, Ci ϭ 37 sis was performed with Tricine-SDS buffer system (25) in a DALTON GBq) for 1 min followed by the addition of an excess of cold methionine tank (Amersham Biosciences) at 30–60 mA/gel for ϳ48 h, until the dye (final concentration 1 mg/ml). Cells were collected by centrifugation at front reached the bottom of the gel. Gels used for comparative analysis 10,000 ϫ g for 10 min at 4 °C, and were subsequently washed twice in were stained with high sensitivity silver stain (26) and gels containing ice-cold M9 minimal medium. For labeled cells, cold methionine was radiolabeled proteins were dried on filter paper. Preparative gels used included in the wash steps. Cell pellets were snap-frozen and stored at for identification of proteins by mass spectrometry were stained with Ϫ80 °C. To prepare samples for isoelectric focusing, cells were lysed Coomassie Brilliant Blue R-250 or with mass spectrometry compatible essentially as described by VanBogelen and Neidhardt (22). Frozen cell silver stain. pellets were thawed on ice and then quickly resuspended in 40 ␮lof Several proteins were found in multiple spots at different pI values, solubilization solution (0.3% (w/v) SDS, 200 mM dithiothreitol, 28 mM but with the same molecular weight. This was also observed in the outer

Tris-HCl (pH 8), 22 mM Tris base and deionized water) per 1 A600 unit membrane maps of E. coli constructed by Molloy et al. (23). Most of

APRIL 14, 2006•VOLUME 281•NUMBER 15 JOURNAL OF BIOLOGICAL CHEMISTRY 10025 Defining the Role of E. coli SecB Downloaded from www.jbc.org FIGURE 1. Analysis of the E. coli secB null mutant by flow cytometry and immunoblotting. A, flow cytometric properties of secB null mutant (SecBϪ) and control (SecBϩ) cells. In the first 2 panels, the size of the population (forward scatter FSC) is plotted versus granularity (side scatter, SSC) for both SecBϩ and SecBϪ. To facilitate comparison of those 2 parameters for SecBϩ and SecBϪ, histograms for size and granularity are shown in the third and fourth panels. One representative experiment of four is shown. Cells were cultured and flow cytometry was performed as described under “Experimental Procedures.” B, Western blot analysis of SecB and components of the Sec-translocase (SecA, -Y, -E), Ffh, a

constituent of the SRP-targeting pathway, and PspA, of which the expression is up-regulated when the electrochemical potential is affected. Left panel, cells (0.2 A600 units) were at Stockholm Universitetsbibliotek on August 6, 2007 separated by means of SDS-PAGE and subsequently subjected to immunoblot analysis with antibodies to SecB, SecA, Ffh, and PspA. Right panel, inner membranes (5 ␮g of protein) were separated by means of SDS-PAGE, and subsequently subjected to immunoblot analysis with antibodies to SecA, SecY, and SecE. these “trains of spots” are because of modifications induced during sam- ham, MA), followed by automatic internal calibration using tryptic pep- ple preparation (27), likely because of stepwise deamidation of residues tides from autodigestion. The latest version of the NBCI non-redundant Asn and Gln, resulting in loss of 1 dalton and net loss of one positive data base (downloaded locally) were searched automatically with the 3 charge. resulting peptide mass lists, using the search engine ProFound (29), as Image Analysis and Statistics—Stained gels were scanned using a part of Knexus (30). Criteria for positive identification by MALDI-TOF GS-800 densitometer from Bio-Rad. Radiolabeled gels were scanned in MS peptide mass fingerprinting were at least four matching peptides a Fuji FLA-3000 phosphorimager. Spots were detected, quantified, with an error distribution within Ϯ25 ppm and at least 15% sequence matched, and compared using the two-dimensional analysis software coverage. During the search, we only allowed one missed cleavage and PDQuest (Bio-Rad). The analyses of silver-stained and radiolabeled partially oxidized methionines. In the more complex samples, the pep- outer membrane proteins were done on the same set of gels. In all cases, tides were also analyzed by nano-LC-ESI-MS/MS in automated mode each analysis set consists of at least three gels in each replicate group (i.e. on a quadruple/orthogonal acceleration TOF tandem mass spectrome- secB null mutant and the control). All gels in a set represented independ- ter (Q-TOF; Micromass, Manchester, UK) (see Ref. 31 for details). The ent samples (i.e. samples from different bacterial colonies, cultures, and spectra were used to search the SwissProt 42.10 data base with the Mascot membrane preparations), which were subjected to two-dimensional search engine. All significant MS/MS identifications by Mascot were man- electrophoresis and image analysis in parallel, i.e. en group. Spot quantities were normalized using the “total density in gel image” method to ually verified for spectral quality and matching y and b ion series. compensate for non-expression related variations in spot quantities Isolation of Protein Aggregates—Protein aggregates were isolated between gels. The PDQuest software was set to detect differences that essentially as described (32). 100 ml of culture with an A600 of 1.0 was were found to be statistically significant using the Student’s t test and a used for each aggregate isolation. The protein content of total cells and 99 (whole cell lysates) or 95% (outer membrane) level of confidence, aggregates was determined with the BCA assay according to the instruc- including qualitative differences (“on-off responses”) present in all gels tions of the manufacturer (Pierce). Aggregates were analyzed by SDS- in a group. Saturated spots were excluded from the analysis. PAGE using 24-cm long 8–16% acrylamide gradient gels. Proteins were Protein Identification by Mass Spectrometry and Bioinformatics— stained with Coomassie Brilliant Blue R-250 and identified by mass Stained protein spots or bands were excised, washed, digested with spectrometry as described before. modified trypsin and peptides extracted manually or automatically For radiolabeling of aggregates, 100 A600 units of cells were labeled (ProPic and Progest, Genomic Solutions, Ann Arbor, MI), and peptides with [35S]methionine (2500 ␮Ci/ml, Ci ϭ 25 GBq) for 30 s and chased were applied to the MALDI target plates as described previously (28). for 1, 3, and 15 min by addition of an excess of cold methionine (final The mass spectra were obtained automatically by MALDI-TOF MS in concentration 1 mg/ml). Aggregates were isolated as described above, reflectron mode (Voyager-DE-STR; PerSeptive Biosystems, Framing- solubilized in 10 mM Tris-HCl (pH 7.5), 2% SDS, and subsequently processed using an OmpA antiserum as described under “Protein 3 V. Zabrouskov et al., unpublished results. Translocation Assays” (see above).

10026 JOURNAL OF BIOLOGICAL CHEMISTRY VOLUME 281•NUMBER 15•APRIL 14, 2006 Defining the Role of E. coli SecB Downloaded from www.jbc.org at Stockholm Universitetsbibliotek on August 6, 2007

FIGURE 2. Analysis of whole cell lysates of the secB null mutant by two-dimensional electrophoresis. A, comparative two-dimensional electrophoresis gel analysis of highly Ϫ ϩ abundant proteins in whole cell lysates of the secB null mutant (SecB ) and its control (SecB ). Cells were harvested when the cultures had reached an A600 of 1.0. 0.5 A600 units of cells were solubilized and proteins were separated by two-dimensional electrophoresis. Proteins were visualized by silver stain and differences between secB null mutant and control gels were analyzed using the PDQuest software (Bio-Rad). At least four independent samples from each strain were used for the analysis. Differential protein expression between the secB null mutant and the control was analyzed using the Student’s t test and a 99% level of confidence (see “Experimental Procedures”). Proteins were identified by mass spectrometry from spots excised from gels stained with Coomassie or mass spectrometry compatible silver stain (Table 1 and supplemental Table 1). Annotated spots have been matched onto the silver-stained gels shown here using the PDQuest software (Bio-Rad). The levels of the highly abundant chaperones DnaK and GroEL are ϳ50% higher in the secB null mutant (supplemental Table 1). In contrast, the level of TF is unaffected in the secB null mutant. B, quantitative immunoblotting of DnaK, GroEL, and TF in secB null mutant (SecBϪ) and control (SecBϩ) cells. SDS-PAGE and blotting were performed as described under “Experimental Procedures.” The levels of the chaperones DnaK and GroEL are increasedinthesecB null mutant, whereas the level of TF is unaffected. C, comparative two-dimensional electrophoresis gel analysis of low abundant proteins in whole cell lysates of the secB null mutant Ϫ ϩ (SecB ) and the control (SecB ). Proteins from 1 A600 unit of cells were visualized by silver staining and analyzed as described above. 46 additional spots are significantly changed using the criteria described above (supplemental Table 1). Spots that are down-regulated in the secB null mutant are indicated in the “SecBϩ” gel and spots that are up-regulated are indicated in the “SecBϪ” gel. Spots where proteins have been successfully identified are labeled with both the spot number and gene name (Table 1 and supplemental Table 1). The level of GroES, co-chaperone, and regulator of GroEL, is 50% increased in the secB null mutant. The level of the chaperone ClpB is doubled and the level of the chaperone/protease HslU is tripled. D, zooms of two-dimensional electrophoresis gels with radiolabeled whole lysates from the secB null mutant (SecBϪ) and the control (SecBϩ) visualized by autoradiography. Processed (m ϭ mature) and precursor (p) forms of the secretory proteins OmpT, OmpA, and OppA are indicated.

APRIL 14, 2006•VOLUME 281•NUMBER 15 JOURNAL OF BIOLOGICAL CHEMISTRY 10027 Defining the Role of E. coli SecB

TABLE 1 Proteins identified in differentially regulated spots in two-dimensional electrophoresis gels of whole cell lysates of the secB null mutant and the control Whole cells were analyzed by two-dimensional electrophoresis (Fig. 2, A and C). Differentially regulated spots were excised from silver- or Coomassie-stained gels. Proteins were identified by MALDI-TOF MS and/or by nano-LC-ESI-MS/MS as described under “Experimental Procedures” (supplemental Table I). Spots 1–2 are from whole cell maps loaded with 0.5 A600 units of cells (Fig. 2A), and the remaining spots are from gels loaded with 1 A600 unit of cells (Fig. 2C). Change (secB null Spot no.a Gene name(s)b Protein name Localizationc mutant/control)d -Fold 1 dnaK, grpF, groP, seg Chaperone protein DnaK Cytoplasmic 1.5 2 groEL, groL, mopA 60-kDa Chaperonin GroEL Cytoplasmic 1.6 3 crr, csr, iex, tgs, treD PTS system, glucose-specific IIA component Cytoplasmic 0.3 5 ompT Protease VII Outer membrane 0.3 10 fhuA, tonA Ferrichrome-iron receptor Outer membrane 0.5 11 luxS S-Ribosylhomocysteinase Cytoplasmic 0.5 14 deoB, drm, thyR Phosphopentomutase Cytoplasmic 0.5 15 ribH, ribE Riboflavin synthase ␤ chain Cytoplasmic 0.5 17 oppA Periplasmic oligopeptide-binding protein Periplasmic 0.5 18 oppA Periplasmic oligopeptide-binding protein Periplasmic 0.6 22 oppA Periplasmic oligopeptide-binding protein Periplasmic 0.7

26 oppA Periplasmic oligopeptide-binding protein Periplasmic 0.7 Downloaded from 27 kdgA, eda, hga KHG/KDPG aldolase Cytoplasmic 0.7 27 ssb, exrB, lexC Single-strand binding protein, helix-destabilizing protein Cytoplasmic 0.7 31 groES, mopB 10-kDa Chaperonin, GroES protein Cytoplasmic 1.5 33 rplL 50 S ribosomal protein L7/L12, L8 Cytoplasmic 1.7 35 clpB Chaperone ClpB Cytoplasmic 2.3 38 hslU, htpI ATP-dependent hsl protease ATP-binding subunit HslU Cytoplasmic 3.8 44 yjfR Hypothetical protein YjfR PSORT:cytoplasmic Ͼ100 Ͼ

46 yjfR Hypothetical protein YjfR PSORT:cytoplasmic 100 www.jbc.org 47 yjfR Hypothetical protein YjfR PSORT: cytoplasmic Ͼ100 a The numbering corresponds to the spots in two-dimensional electrophoresis gel images shown in Fig. 2, A and C. b Names in bold are used to label the corresponding spots in the gel shown in Fig. 2, A and C. c Localization according to the SwissProt database. The localization of unknown proteins was predicted using PSORT. d The PDQuest software was used to detect and calculate the -fold change, i.e. the ratio of the average intensity of spots in secB null mutant gels to the average intensity of matched at Stockholm Universitetsbibliotek on August 6, 2007 spots in the control gels (supplemental Table I).

RESULTS targets inner membrane proteins to the Sec-translocase but may have Characterization of the secB Null Mutant Strain—Using P1 transduc- some overlap with the SecB targeting pathway (20, 35–37). It is not tion we moved the secB null mutation secB8 (15) from strain HS101/ known if the SRP targeting pathway can compensate for the absence of pE63 (14) into strain EK413/pE63 (a MC4100 derivative that is araϩ;a SecB. Immunoblot analysis of secB null mutant and control showed that kind gift from Ken-ichi Nishiyama), yielding an EK413/pE63-derived Ffh levels were unchanged in the absence of SecB. It has been shown that secB null mutant strain. Hereafter, we will refer to the secB null mutant expression of PspA is up-regulated when the electrochemical potential strain and EK413/pE63 as the secB null mutant and control, respectively. is affected (38). Because the electrochemical potential plays an impor- The secB null mutant and control were cultured aerobically in M9 min- tant role in protein translocation we analyzed the levels of the PspA imal medium. Under these conditions, we did not observe any differ- protein by immunoblot analysis. In contrast to several other Sec ences in growth, as monitored by A600 measurements. In addition, pro- mutants (38), there is no PspA response in the secB null mutant (Fig. 1B). pidium iodide staining (16) did not point to differences in viability Analysis of Whole Cell Lysates of the secB Null Mutant by Two-di- between the mutant and the control (results not shown). Early log- mensional Electrophoresis—To identify potential SecB substrates and phase cells were used in all the experiments described in this study. The compensatory mechanisms and/or stress responses in the secB null morphology of cells was analyzed by means of flow cytometry (16, 33). mutant, we used a proteomics approach. Whole cell lysates of the Interestingly, we detected a small increase of both the forward scatter mutant and the control were analyzed by two-dimensional electro- and side scatter of secB null mutant cells (Fig. 1A). This indicates that phoresis, using IPG strips with a pI range from 4 to 7. To allow for secB null mutant cells are slightly bigger than control cells and most quantitative analysis of highly abundant proteins, gels were loaded with likely contain extra internal structures (i.e. extra membranes and/or limited amounts of protein, such that staining of highly abundant pro- protein aggregates). teins was not saturated. The comparative analysis was based on 4 gels To verify the phenotype of the secB null mutant strain, we monitored per strain (an independent culture was used for each gel). Gels were the targeting of the established SecB-dependent outer membrane pro- stained with silver, scanned, and images were analyzed and compared tein OmpA (14) and SecB-independent periplasmic protein ␤-lactamase (34), using pulse-chase radiolabeling experiments in combination using the PDQuest software (Bio-Rad). Significance was determined with immunoprecipitations. As expected, the translocation of OmpA using Student’s t test (for details see “Experimental Procedures”). This was hampered in the secB null mutant, as evidenced by accumulation of analysis demonstrated that the levels of both DnaK and GroEL were Ͻ precursor protein, whereas the translocation of ␤-lactamase was not increased by about 50% (p 0.01) in the secB null mutant (Fig. 2A, Table affected (results not shown). The levels of SecA, -Y, and -E were deter- 1, and supplemental Table 1). Higher levels of DnaK and GroEL are mined by Western blotting for the secB null mutant and control, consistent with increased synthesis rates of these proteins in a SecB because SecB delivers proteins to the SecYE protein-conducting chan- knock-out strain (39). The level of TF, the first cytoplasmic chaperone nel through interaction with SecA (1). Protein levels of the SecAYE- that interacts with ribosome-associated nascent peptides, was not translocase components did not change in the absence of SecB (Fig. 1B). affected. These results were confirmed by immunoblot analysis Ffh is a core component of the SRP targeting pathway, which mainly (Fig. 2B).

10028 JOURNAL OF BIOLOGICAL CHEMISTRY VOLUME 281•NUMBER 15•APRIL 14, 2006 Defining the Role of E. coli SecB

To detect differences in accumulation of low abundant proteins, we repeated the two-dimensional electrophoresis proteome analysis with higher protein loading (Fig. 2C). In total 48 spots were found to be significantly (with p Ͻ 0.01) altered. Although most of these spots were weakly stained, we were able to identify 16 non-redundant proteins by MS (Table 1; image analysis and MS data are summarized in supplementary Table 1). Some proteins were identified in multiple spots located in “trains” next to each other and in total 20 spots could be annotated. In addition to DnaK and GroEL, the levels of the chaperones GroES, ClpB, and HslU were increased by ϳ50, ϳ200, and ϳ400%, respectively. GroES is the regulator and co-chaperone of GroEL and the GroEL/ES chaperone system is essential for proper folding and maturation of proteins in the cytoplasm (40). The chaperone ClpB has been shown to act in concert with DnaK and inclusion body proteins (Ibp) to extract and refold proteins from aggregates (41). The protein HslU can function either on its own as a chaperone (42) or in a complex together with HslV (ClpY) as a protease (43). However, we did not find any spots Downloaded from that were differentially regulated in the region of the gel where HslV should migrate. Notably, the genes encoding the DnaK, GroEL/ES, ClpB, and HslU proteins are all regulated by transcription factor ␴32 (44, 45). The ␴32-induced response, better known as the “heat shock response,” is activated in response to protein misfolding/aggregation in www.jbc.org the cytoplasm (44, 45). The levels of the processed forms of the outer membrane proteins ferrichrome-iron receptor (FhuA), the protease OmpT, and the

periplasmic oligopeptide-binding protein (OppA) were decreased by at Stockholm Universitetsbibliotek on August 6, 2007 ϳ50, ϳ70, and ϳ30%, respectively, in the secB null mutant strain. Inter- estingly, the homolog of OppA in Salmonella typhimurium has been shown to bind tightly to E. coli SecB in vitro (46). No precursors of secretory proteins were identified in the silver-stained gels of whole cell lysates. To study kinetic effects of the secB null mutation, we repeated the comparative proteome analysis with cells labeled with [35S]methionine. In these radioactive gels, 37 radiolabeled spots were significantly changed (p Ͻ 0.01) in the secB null mutant. Interestingly, several of these spots were not present in the silver-stained gels. Based on pI and molecular weight they match the precursors of secretory proteins, such FIGURE 3. Characterization of aggregates isolated from E. coli secB null mutant. A, as OmpA, OmpT, and OppA (Fig. 2D). aggregates from the secB null mutant (SecBϪ) and the control (SecBϩ) were isolated from 100-ml cultures in M9 minimal media. The protein content of the aggregates was Isolation and Characterization of Protein Aggregates from the secB analyzed by SDS-PAGE on a 24-cm long 8–16% gradient gel stained by Coomassie Bril- Null Mutant—The heat shock response in the secB null mutant pointed liant Blue R-250 and proteins were identified by mass spectrometry as described under to a problem of protein folding and aggregation in the cytoplasm of cells “Experimental Procedures” (Table 2 and supplemental Table 2). Proteins that were identified with a signal sequence are marked with an asterisk (*). Known or predicted local- lacking SecB. In addition, the flow cytometry experiments suggested the izations are indicated by: om, outer membrane; om lp, outer membrane lipoprotein; cyt, presence of extra internal structures, i.e. extra internal membranes cytoplasmic; per, periplasmic; or sec, secretory (predicted by PSORT to be either periplasmic or outer membrane protein). B, quantitative immunoblotting of IbpA/B in secB null and/or protein aggregates, in the secB null mutant. Indeed, protein mutant (SecBϪ) and control (SecBϩ) whole cell lysates. The expression levels of IbpA/B aggregates containing around 0.5% of total cellular protein could be are considerably higher in the secB null mutant. C, OmpA was isolated from aggregates prepared from secB null mutant (SecBϪ) control (SecBϩ) cells. Cells were labeled with isolated from the secB null mutant, but were virtually absent in the [35S]methionine and chased with cold methionine for different times as indicated in the control strain (Fig. 3A). The aggregates were dissolved in Laemmli sol- figure. Aggregates were prepared and OmpA was subsequently isolated by means of ubilization buffer, and proteins were separated by SDS-PAGE, followed immunoprecipitation as described under “Experimental Procedures.” To be able to distinguish between the precursor (pro-OmpA) and processed forms of OmpA, the immu- by identification using nano-LC electrospray tandem mass spectrome- noprecipitation was also performed on [35S]methionine-labeled whole cells from E. coli ϩ Ϫ try (nano-LC-ESI-MS/MS). Fourteen secretory proteins and five cyto- strain CM124 (PAra-secE) cultured in the presence (SecE ) and absence (SecE ) of arabinose. In SecEϩ cells there is a fully functional Sec-translocase and OmpA is translocated plasmic proteins were identified (Fig. 3A, Table 2, and supplemental across the cytoplasmic membrane and subsequently processed, whereas in SecEϪ cells Table 2). The inclusion body protein IbpA, also part of the heat shock there is not a functional Sec-translocase and the precursor form of OmpA accumulates in regulon, was among the cytoplasmic proteins identified in the aggre- the cytoplasm (54). gates (Table 2) (41, 47). Immunoblotting of total cell extracts showed Furthermore, MS/MS data revealed that at least two of the secretory that the levels of the chaperones IbpA/B are indeed strongly up-regu- proteins, OmpA and the murein lipoprotein (Lpp or MulI), identified in lated in the secB null mutant (Fig. 3B). In contrast, immunoblotting the aggregates contained an uncleaved signal sequence (results not showed that the levels of the periplasmic chaperone Skp and protease shown), again pointing to aggregation of proteins in the cytoplasm DegP are not changed in the secB null mutant (data not shown). This rather than in the periplasm. indicates that no significant protein misfolding/aggregation occurred in To study the localization and dynamics of the aggregates in more the periplasm/outer membrane (38, 48). detail, we isolated OmpA by immunoprecipitation from aggregates iso-

APRIL 14, 2006•VOLUME 281•NUMBER 15 JOURNAL OF BIOLOGICAL CHEMISTRY 10029 Defining the Role of E. coli SecB

TABLE 2 Identification of proteins in aggregates isolated from the secB null mutant and the control Protein bands were excised from Coomassie-stained one-dimensional gels loaded with aggregates isolated from the secB null mutant and the control (Fig. 3A). Proteins were identified by MALDI-TOF MS and/or nano-LC-ESI-MS/MS as described under “Experimental Procedures.” Band no.a Gene name(s)b Protein name(s) Localizationc 1 glgB 1,4-␣-Glucan branching enzyme Cytoplasmic 2 tolC, mtcB, mukA, refI Outer membrane protein TolC Outer membrane 3 glgA Glycogene synthase Cytoplasmic 4 degP, htrA Heat shock protein HtrA Periplasmic 4 tolB TolB protein Periplasmic 5 yncE Hypothetical protein YncE Unknown, PSORT predicts periplasmic or outer membrane 6 ompA, tolG, tut, con Outer membrane protein A Outer membrane 6 yncE Hypothetical protein YncE Unknown, PSORT predicts periplasmic or outer membrane 7 ydgH Hypothetical protein YdgH Unknown, PSORT predicts periplasmic or outer membrane 8 fkpA FKBP-type peptidyl-prolyl cis-trans isomerase FkpA Periplasmic 9 ybgF Hypothetical protein YbgF Unknown, PSORT predicts periplasmic or outer membrane 10 ompA, tolG, tut, con Outer membrane protein A, Outer membrane Outer membrane protein II* Downloaded from 11 lpp, mulI, mlpA Major outer membrane lipoprotein precursor Outer membrane lipoprotein (murein-lipoprotein) 12 infC Initiation factor 3 Cytoplasmic 13 ompX Outer membrane protein X Outer membrane 14 dps Starvation-inducible DNA-binding protein Cytoplasmic 15 ibpA, hslT, htpN 16-kDa heat shock protein A Cytoplasmic 15 ygiW Hypothetical Protein YgiW Unknown, PSORT predicts periplasmic or outer membrane www.jbc.org 16 lpp, mulI, mlpA Major outer membrane lipoprotein precursor, Outer membrane lipoprotein murein-lipoprotein 16 ycgK Hypothetical protein YcgK Unknown, PSORT predicts periplasmic or outer membrane

a The numbering corresponds to the bands in the one-dimensional electrophoresis gel shown in Fig. 3A. at Stockholm Universitetsbibliotek on August 6, 2007 b The gene names and synonyms. Names in bold are used to label the corresponding bands in the gel shown in Fig. 3A. c Localization according to the SwissProt database. The localization of unknown proteins was predicted using PSORT. lated from [35S]methionine-labeled cells (Fig. 3C). Only the precursor the insertion kinetics of the outer membrane proteome in the same set form of OmpA could be isolated from secB null mutant aggregates, of gels. To facilitate identification of proteins by MS, we generated pre- which is another indication for their cytoplasmic localization. Interest- parative gels stained with Coomassie Brilliant Blue in parallel to the ingly, the OmpA signal disappeared during the chase with non-radio- radioactive silver-stained gels. Twenty-five proteins were identified active methionine, indicating that after extraction from the aggregates, (Fig. 4A, Table 3, and supplemental Table 3). Fourteen of those are pro-OmpA is either degraded or remobilized for translocation. established integral (␤-barrel) outer membrane proteins, four are Analysis of the Outer Membrane Proteome of the secB Null Mutant— peripheral outer membrane lipoproteins, and one is a potential integral Our analysis of labeled whole cell lysates suggested that the secB null outer membrane protein. The remaining six proteins are peripheral mutation affects the targeting of secretory proteins by slowing down the membrane proteins interacting with the inner or outer membrane and delivery of the precursor to the Sec-translocase. Furthermore, secretory integral inner membrane proteins with one predicted transmembrane proteins were identified in aggregates isolated from the secB null segment. mutant. Therefore, we decided to compare the secretomes (periplasm The comparative analysis of the silver-stained gels showed that levels and outer membrane) of the secB null mutant and the control using of iron-siderophore transporter FhuA, the ferrichrome-iron receptor comparative two-dimensional electrophoresis analysis of radiolabeled FhuE, and peptidoglycan-associated lipoprotein (Pal) were decreased and unlabeled cells. Attempts to analyze the periplasmic proteome 2-fold in the absence of SecB. In contrast, the level of the ferri-enter- failed, because the isolated periplasmic fractions were insufficiently obactin receptor FepA was doubled, possibly compensating for the pure for conclusive comparative proteome analysis. Importantly, we decrease of the FhuA and FhuE levels. These significant changes in were successful at quantitatively comparing the steady state and assem- levels of key players in the strictly regulated iron metabolism of the bly kinetics of the outer membrane proteomes of the secB null mutant E. coli cell had apparently no effect on growth of the secB null mutant and the control. under the experimental conditions used. However, it is very well con- The outer membrane proteins are ␤-barrel proteins and they can be ceivable that under other conditions similar changes could have signif- well resolved on two-dimensional electrophoresis gels when using a icant effects. Steady state levels of all the other identified outer mem- mixture of the non-ionic detergent ASB-14 and thiourea (23). The inner brane proteins were unchanged. membrane contains ␣-helical proteins, which typically precipitate dur- Importantly, the analysis of the two-dimensional electrophoresis ing isoelectric focusing and are thus not resolved on the two-dimen- autoradiograms of [35S]methionine-labeled proteins revealed that many sional electrophoresis gels. Therefore, two-dimensional electrophoresis outer membrane proteins, such as BtuB, FhuA, FhuE, FadL, OmpT, gels from the total membrane fraction visualize predominantly outer OmpX, and TolC appear considerably slower in the outer membrane of membrane proteins and peripheral membrane proteins (23). Mem- the secB null mutant than in the outer membranes of the control. This branes isolated from [35S]methionine-labeled cells were used for the shows that SecB helps to improve efficiency of secretion, rather than analysis. The gels were first stained with silver and then used for auto- being strictly essential. In Fig. 4B, this is shown in more detail for two radiography. This allowed us to study the steady state composition and examples, FhuA and TolC. In the secB null mutant, both FhuA and TolC

10030 JOURNAL OF BIOLOGICAL CHEMISTRY VOLUME 281•NUMBER 15•APRIL 14, 2006 Defining the Role of E. coli SecB Downloaded from www.jbc.org at Stockholm Universitetsbibliotek on August 6, 2007

FIGURE 4. Analysis of the outer membrane proteome from the E. coli secB null mutant. A, outer membrane proteomes of the E. coli secB null mutant (SecBϪ) and the control strain (SecBϩ) were visualized by two-dimensional electrophoresis. 350 ␮g of total membrane proteins was used for each gel. Proteins were visualized by silver stain and differences between secB null mutant and control gels were analyzed using the PDQuest software. At least three independent samples from each strain were used for the analysis. Differences in protein expression between the secB null mutant and the control were evaluated using the Student’s t test and a 99% level of confidence (see “Experimental Procedures”). Saturated spots were excluded from the analysis. One representative gel from each strain is shown. Proteins were identified by mass spectrometry from spots excised from Coomassie-stained gels (Table 3 and supplemental Table 3). Identified proteins are indicated by gene name. Annotated spots were matched from the Coomassie-stained gels to the silver-stained gels shown in the figure using the PDQuest software. B, zooms of autoradiographs of two-dimensional electrophoresis gels with radiolabeled outer membranes of the secB null mutant (SecBϪ) and control (SecBϩ). Membranes were isolated from cells harvested directly after labeling or after a 10-min chase as described under “Experimental Procedures.” Note that the FepA signal is, just like in the silver-stained gels, higher in the secB null mutant than in the control.

can only be detected after a 10-min chase rather than directly after growth, but flow cytometry experiments showed that the secB null labeling as in the control strain. mutation did affect cell morphology; cells are slightly bigger and seem to Identification of Novel SecB-dependent Secretory Proteins—SecB contain internal structures. dependence of 12 potential SecB substrates identified in the protein The comparative proteome analysis of the secB null mutant resulted aggregates or in the two-dimensional electrophoresis gels was directly in three main observations and conclusions: 1) absence of SecB results in monitored using a pulse-labeling approach (Fig. 5 and supplemental aggregation of secretory proteins and increased levels of cytoplasmic chap- Table 4). erones; 2) SecB is not required for targeting and translocation of secretory Cells with and without SecB were labeled with [35S]methionine and proteins per se, but rather is needed to improve efficiency of the cytoplasmic the precursor/processed forms of the secretory proteins tested (DegP, targeting process and delivery to the Sec-translocase; and 3) SecB depend- FhuA, FkpA, OmpT, OmpX, OppA, TolB, TolC, YbgF, YcgK, YgiW and ence was established for an additional 12 secretory proteins. Below we YncE) were subsequently immunoprecipitated and analyzed by SDS- explain and discuss these main conclusions and observations in more detail. PAGE and autoradiography. Strikingly, translocation of all these secre- Absence of SecB Affects Protein Homeostasis in the Cytoplasm—The tory proteins was hampered in the secB null mutant, as shown by the two-dimensional electrophoresis analysis of radiolabeled whole cell accumulation of precursors compared with control. This shows that all lysates showed that precursors of secretory proteins accumulate in the these proteins indeed need SecB for efficient translocation across the secB null mutant. Furthermore, the secB null mutation induced the ␴32- cytoplasmic membrane. In addition, similar pulse-labeling experiments regulated heat shock response, which is diagnostic for protein aggrega- in the presence of the SecA inhibitor azide showed that translocation of tion/misfolding in the cytoplasm. these 12 proteins is also SecA dependent (results not shown). Indeed, cytoplasmic protein aggregates, which contained mainly secretory proteins, were for the first time isolated from secB null mutant DISCUSSION cells. The amount of aggregated proteins in secB null mutant cells To characterize the role of the chaperone SecB in E. coli in more (around 0.5% of the total protein) is about half as much as in single detail, we studied a secB null mutant strain using a comparative pro- mutant cells that lack cytoplasmic chaperones like DnaK and TF (32, 49, teomics approach complemented with Western blotting and pulse- 50). This suggests that DnaK and TF play a different role than SecB in chase experiments. The secB null mutation did not significantly affect protein homeostasis.

APRIL 14, 2006•VOLUME 281•NUMBER 15 JOURNAL OF BIOLOGICAL CHEMISTRY 10031 eiigteRl fE oiSecB coli E. of Role the Defining 10032 ORA FBOOIA CHEMISTRY BIOLOGICAL OF JOURNAL

TABLE 3 Quantification of silver-stained and ͓35 S͔methionine-labeled two-dimensional electrophoresis outer membrane maps ͓35 S͔Methionine-labeled membranes from the secB null mutant and the control were analyzed by two-dimensional electrophoresis (Fig. 4, A and B). Silver-stained and ͓35 S͔methionine-labeled spots were matched and quantified as described under “Experimental Procedures.” Proteins were identified by MALDI-TOF mass spectrometry from Coomassie-stained protein spots. Details are given in supplementary Table 3. Ratio silver stain (secB null Ratio ͓35 S͔Met (secB Gene name(s)a Protein name Localizationb mutant/control)c null mutant/control)c acrA, mtcA,lir Acriflavine resistance protein A Inner membrane lipoprotein No significant change No significant change btuB, bfe, cer, dcrC Vitamin B12 receptor Outer membrane No significant change 0.23 cirA, cir, feuA Colicin I receptor Outer membrane No significant change No significant change fadL, ttr Long-chain fatty acid transport protein Outer membrane No significant change 0.27 fepA, fep, feuB Ferrienterobactin receptor Outer membrane 2.16 1.28 fhuA, tonA Ferrichrome-iron receptor Outer membrane 0.46 0.15 fhuE FhuE receptor Outer membrane 0.52 0.16 ftsZ, sifB, sulB Cell division protein FtsZ Cytoplasm (attaches to the inner membrane during No significant change No significant change cell division) metQ D-Methionine-binding lipoprotein metQ Probably attached to membrane by lipid anchor No significant change Not detected nlpA Lipoprotein-28 Inner membrane lipoprotein No significant change No significant change nlpB, dapX Lipoprotein-34 Outer membrane lipoprotein No significant change No significant change ompA, tolG, tut, con Outer membrane protein A Outer membrane No significant change No significant change opmC, meoA, par Outer membrane protein C Outer membrane No significant change No significant change ompT OmpT, Omptin, Protease A Outer membrane No significant change 0.34 ompX Outer membrane protein X Outer membrane No significant change 0.28 ostA, imp Organic solvent tolerance protein Outer membrane No significant change Not detected pal, excC Peptidoglycan-associated lipoprotein Outer membrane lipoprotein 0.49 No significant change ppiD Peptidyl-prolyl cis-trans isomerase D Inner membrane No significant change Not detected tolC, mtcB, mukA, refI TolC Outer membrane No significant change 0.30 tsx, nupA Nucleoside-specific channel-forming protein tsx Outer membrane No significant change No significant change yaeT Outer membrane protein assembly factor YaeT Outer membrane No significant change No significant change vacJ VacJ lipoprotein Outer membrane lipoprotein No significant change No significant change ybhC Hypothetical protein in bioA 5Ј region, Putative Probably attached to the OM with lipid anchor No significant change Not detected

OUE21NME 15• 281•NUMBER VOLUME lipoprotein YbhC ybiL Probable tonB-dependent receptor YbiL Potentially outer membrane No significant change No significant change yfgM Hypothetical protein YfgM PSORT predicts inner membrane No significant change No significant change a The gene names and synonyms. Names in bold are used to label the corresponding spots in the two-dimensional electrophoresis gel images in Fig. 4, A and B. b Localization according to the SwissProt database. The localization of unknown proteins was predicted using PSORT. c -Fold change of significantly (p Ͻ 0.01 in silver-stained gels and p Ͻ 0.05 in ͓35 S͔methionine-labeled gels) changed spots calculated as the ratio of the average intensity of spots in secB null mutant gels to the average intensity of matched spots in the control gels. Note that not all identified proteins were detected in ͓35 S͔methionine-labeled gels.

PI 4 2006 14, APRIL

Downloaded from from Downloaded www.jbc.org at Stockholm Universitetsbibliotek on August 6, 2007 2007 6, August on Universitetsbibliotek Stockholm at Defining the Role of E. coli SecB

cating that the translocase capacity was not compromised. Further- more, the secB null mutation did not induce a PspA response, indicating that the electrochemical potential, which plays an important role in protein secretion, is not affected by the absence of SecB. Thus the phenotype of the secB null mutation is a direct consequence of the absence of SecB. This urged us to take a closer look at the secretome of the secB null mutant. Our analysis of the outer membrane proteome showed that the steady state levels of most proteins were unaffected. However, assembly of a considerable number of outer membrane proteins into the outer membrane was delayed in the secB null mutant. Recently, an outer membrane protein complex consisting of the proteins YeaT, NlpB, YfgL, and YfiO has been shown to act as an insertion machinery for outer membrane proteins (52, 53). In our outer membrane two-dimensional electrophoresis gels, we identified YeaT and lipoprotein NlpB. The steady state levels of these components were unaffected in the secB null strain. The analysis of radiolabeled proteins showed no significant Downloaded from differences in the levels of YeaT and NlpB. This suggests that the outer membrane protein insertion capacity is not affected in the secB null mutant, which is consistent with the absence of cell envelope stress responses. The characterization of the outer membrane proteome of the secB null mutant along with the observation that the secB null mutation does not cause any significant protein misfolding/aggregation in www.jbc.org the periplasm/outer membrane indicates that the bottleneck created by the absence of SecB is at the level of sorting of outer membrane proteins across the inner membrane rather than their sorting in the cell envelope, at Stockholm Universitetsbibliotek on August 6, 2007 and that SecB is not essential for targeting of these proteins to the Sec-translocase but rather facilitates their targeting. The levels of a few processed secretory proteins (FhuA, FhuE, OmpT, and OppA) go drastically down in the absence of SecB. Notably, none of these proteins were detected in the aggregates. It is tempting to speculate that in the absence of SecB the precursor forms of these proteins are FIGURE 5. Identification of SecB-dependent secretory proteins. Targeting of the more prone to proteolysis, thereby lowering their levels. potential SecB substrates, DegP, FhuA, FkpA, OmpT, OmpX, OppA, TolB, TolC, YbgF, YcgK Comparative Proteome Analysis as a Platform for the Identification of YgiW, and YncE, was monitored in the secB null mutant (SecBϪ) and the control (SecBϩ) using the classical pulse approach, combined with specific immunoprecipitations (see SecB Substrates—We used a pulse-labeling approach to directly moni- “Experimental Procedures”). To facilitate immunoprecipitations of the potential SecB tor the SecB dependence of a subset of aggregated secretory proteins, as substrates, a hemagglutinin tag was attached to the C terminus of each protein. The well as secretory proteins that were affected in two-dimensional elec- SecB-independent secretory protein ␤-lactamase (AmpC) was used as a control to monitor if the hemagglutinin tag could confer SecB dependence. m, mature; p, precursor. trophoresis maps of whole cell lysates and the outer membrane of the secB null mutant. Strikingly, translocation of all tested potential SecB substrates was indeed SecB-dependent. This clearly demonstrates that The pro-OmpA isolated from [35S]methionine-labeled aggregates the comparative proteomics approach is an excellent platform for the disappears in a chase, showing that the protein aggregates are dynamic. identification of SecB-dependent secretory proteins. Thus far, bioinfor- Proteins extracted from aggregates may either be degraded or get a matic analysis of these novel and previously established SecB substrates second chance to be translocated. Based on recent observations that 4 has not lead to the identification of a common denominator, stressing Ibp/ClpB/DnaK-mediated reactivation of aggregated proteins plays an the importance of experimentation in protein targeting research. important role in viability of the E. coli cell (41, 51), we suggest that the Recently, pulse-labeling experiments showed that the SRP pathway is aggregates in the secB null mutant are actively reactivated for translo- required for efficient targeting of the murein lipoprotein Lpp to the cation rather than being degraded. Sec-translocase (20). However, the identification of the precursor form The composition of the aggregates and whole cell lysates of the secB of Lpp in the aggregates in the current study strongly suggests that null mutant do not point to a significant role of SecB in the folding of targeting of at least a small fraction of Lpp (which is the most highly cytoplasmic proteins. However, a recent study suggests that SecB can expressed protein in E. coli) also depends on SecB. Our observations play a significant role in the folding of cytoplasmic proteins under spe- thus further strengthen the idea that selected proteins can be targeted cialized conditions, e.g. in the absence of both the chaperones DnaK and by both the SRP and SecB targeting pathways (20, 35–37). TF (13). Conclusions—The analysis of a secB null mutant using comparative SecB Improves Secretion Efficiency but Is Not Required for Secretion proteomics clearly points to a primary role of the chaperone SecB in per se—The majority of the proteins we identified in the aggregates facilitating targeting of secretory proteins to the Sec-translocase, and isolated from the secB null mutant are secretory proteins, and, corre- has enabled us to more than triple the number of known SecB sub- spondingly, the two-dimensional electrophoresis analysis of whole cell strates. This shows for the first time that comparative proteomics is a lysates indicated that deletion of secB affects the targeting kinetics of secretory proteins. The secB null mutation did not cause any changes in the levels of the Sec-translocase core components SecA, -Y, or -E, indi- 4 L. Baars and J. W. de Gier, unpublished results.

APRIL 14, 2006•VOLUME 281•NUMBER 15 JOURNAL OF BIOLOGICAL CHEMISTRY 10033 Defining the Role of E. coli SecB very powerful tool to study protein targeting pathways and their sub- 757–761 strates in E. coli. 28. Peltier, J. B., Emanuelsson, O., Kalume, D. E., Ytterberg, J., Friso, G., Rudella, A., Liberles, D. A., Soderberg, L., Roepstorff, P., von Heijne, G., and van Wijk, K. J. (2002) Plant Cell 14, 211–236 Acknowledgments—Bacterial strains and plasmid pE63 were a kind gift from 29. Zhang, W., and Chait, B. T. (2000) Anal. Chem. 72, 2482–2489 Ken-ichi Nishiyama; Ben de Kruijff, Annemieke van Dalen, Arnold Driessen, 30. Field, H. I., Fenyo, D., and Beavis, R. C. (2002) Proteomics 2, 36–47 Jan Tommassen, Axel Mogk, Bernd Bukau, and Joen Luirink are thanked for 31. Friso, G., Giacomelli, L., Ytterberg, A. J., Peltier, J. B., Rudella, A., Sun, Q., and Wijk, antisera. Gunnar von Heijne, Joen Luirink, and Dirk-Jan Slotboom are K. J. (2004) Plant Cell 16, 478–499 thanked for valuable discussions. 32. Tomoyasu, T., Mogk, A., Langen, H., Goloubinoff, P., and Bukau, B. (2001) Mol. Microbiol. 40, 397–413 33. Davey, H. M., and Kell, D. B. (1996) Microbiol. Rev. 60, 641–696 REFERENCES 34. Laminet, A. A., Kumamoto, C. A., and Pluckthun, A. (1991) Mol. Microbiol. 5, 117–122 1. Manting, E. H., and Driessen, A. J. (2000) Mol. Microbiol. 37, 226–238 35. Sijbrandi, R., Urbanus, M. L., Ten Hagen-Jongman, C. M., Bernstein, H. D., Oudega, 2. Mori, H., and Ito, K. (2001) Trends Microbiol. 9, 494–500 B., Otto, B. R., and Luirink, J. (2003) J. Biol. Chem. 278, 4654–4659 3. Van den Berg, B., Clemons, W. M., Jr., Collinson, I., Modis, Y., Hartmann, E., Harri- 36. Peterson, J. H., Woolhead, C. A., and Bernstein, H. D. (2003) J. Biol. Chem. 278, son, S. C., and Rapoport, T. A. (2004) Nature 427, 36–44 46155–46162 4. Hartl, F. U., Lecker, S., Schiebel, E., Hendrick, J. P., and Wickner, W. (1990) Cell 63, 37. Schierle, C. F., Berkmen, M., Huber, D., Kumamoto, C., Boyd, D., and Beckwith, J. 269–279 (2003) J. Bacteriol. 185, 5706–5713 5. Wickner, W., Driessen, A. J., and Hartl, F. U. (1991) Annu. Rev. Biochem. 60, 101–124 38. Darwin, A. J. (2005) Mol. Microbiol. 57, 621–628

6. Fekkes, P., den Blaauwen, T., and Driessen, A. J. (1995) Biochemistry 34, 39. Wild, J., Walter, W. A., Gross, C. A., and Altman, E. (1993) J. Bacteriol. 175, Downloaded from 10078–10085 3992–3997 7. Fekkes, P., van der Does, C., and Driessen, A. J. (1997) EMBO J. 16, 6105–6113 40. Kerner, M. J., Naylor, D. J., Ishihama, Y., Maier, T., Chang, H. C., Stines, A. P., Geor- 8. Fekkes, P., and Driessen, A. J. (1999) Microbiol. Mol. Biol. Rev. 63, 161–173 gopoulos, C., Frishman, D., Hayer-Hartl, M., Mann, M., and Hartl, F. U. (2005) Cell 9. Randall, L. L., and Hardy, S. J. (2002) Cell Mol. Life Sci. 59, 1617–1623 122, 209–220 10. Dekker, C., de Kruijff, B., and Gros, P. (2003) J. Struct. Biol. 144, 313–319 41. Mogk, A., Deuerling, E., Vorderwulbecke, S., Vierling, E., and Bukau, B. (2003) Mol. 11. Xu, Z., Knafels, J. D., and Yoshino, K. (2000) Nat. Struct. Biol. 7, 1172–1177 Microbiol. 50, 585–595 12. Knoblauch, N. T., Rudiger, S., Schonfeld, H. J., Driessen, A. J., Schneider-Mergener, J., 42. Seong, I. S., Oh, J. Y., Lee, J. W., Tanaka, K., and Chung, C. H. (2000) FEBS Lett. 477, www.jbc.org and Bukau, B. (1999) J. Biol. Chem. 274, 34219–34225 224–229 13. Ullers, R. S., Luirink, J., Harms, N., Schwager, F., Georgopoulos, C., and Genevaux, P. 43. Seol, J. H., Yoo, S. J., Shin, D. H., Shim, Y. K., Kang, M. S., Goldberg, A. L., and Chung, (2004) Proc. Natl. Acad. Sci. U. S. A. 101, 7583–7588 C. H. (1997) Eur. J. Biochem. 247, 1143–1150 14. Shimizu, H., Nishiyama, K., and Tokuda, H. (1997) Mol. Microbiol. 26, 1013–1021 44. Arsene, F., Tomoyasu, T., and Bukau, B. (2000) Int. J. Food Microbiol. 55, 3–9

15. Kumamoto, C. A., and Beckwith, J. (1985) J. Bacteriol. 163, 267–274 at Stockholm Universitetsbibliotek on August 6, 2007 45. Rosen, R., and Ron, E. Z. (2002) Mass Spectrom. Rev. 21, 244–265 16. Hewitt, C. J., and Nebe-Von-Caron, G. (2004) Adv. Biochem. Eng. Biotechnol. 89, 197–223 46. Smith, V. F., Hardy, S. J., and Randall, L. L. (1997) Protein Sci. 6, 1746–1755 17. Osborn, M. J., Gander, J. E., and Parisi, E. (1972) J. Biol. Chem. 247, 3973–3986 47. Carrio, M. M., and Villaverde, A. (2003) FEBS Lett. 537, 215–221 18. Osborn, M. J., Gander, J. E., Parisi, E., and Carson, J. (1972) J. Biol. Chem. 247, 48. Duguay, A. R., and Silhavy, T. J. (2004) Biochim. Biophys. Acta 1694, 121–134 3962–3972 49. Deuerling, E., Patzelt, H., Vorderwulbecke, S., Rauch, T., Kramer, G., Schaffitzel, E., 19. Froderberg, L., Rohl, T., van Wijk, K. J., and de Gier, J. W. (2001) FEBS Lett. 498, 52–56 Mogk, A., Schulze-Specking, A., Langen, H., and Bukau, B. (2003) Mol. Microbiol. 47, 20. Froderberg, L., Houben, E. N., Baars, L., Luirink, J., and De Gier, J. W. (2004) J. Biol. Chem. 279, 31026–31032 1317–1328 21. Froderberg, L., Houben, E., Samuelson, J. C., Chen, M., Park, S. K., Phillips, G. J., 50. Vorderwulbecke, S., Kramer, G., Merz, F., Kurz, T. A., Rauch, T., Zachmann-Brand, Dalbey, R., Luirink, J., and De Gier, J. W. (2003) Mol. Microbiol. 47, 1015–1027 B., Bukau, B., and Deuerling, E. (2004) FEBS Lett. 559, 181–187 22. VanBogelen, R. A., and Neidhardt, F. C. (1999) Methods Mol. Biol. 112, 21–29 51. Weibezahn, J., Tessarz, P., Schlieker, C., Zahn, R., Maglica, Z., Lee, S., Zentgraf, H., 23. Molloy, M. P., Herbert, B. R., Slade, M. B., Rabilloud, T., Nouwens, A. S., Williams, Weber-Ban, E. U., Dougan, D. A., Tsai, F. T., Mogk, A., and Bukau, B. (2004) Cell 119, K. L., and Gooley, A. A. (2000) Eur. J. Biochem. 267, 2871–2881 653–665 24. Peltier, J. B., Friso, G., Kalume, D. E., Roepstorff, P., Nilsson, F., Adamska, I., and van 52. Wu, T., Malinverni, J., Ruiz, N., Kim, S., Silhavy, T. J., and Kahne, D. (2005) Cell 121, Wijk, K. J. (2000) Plant Cell 12, 319–341 235–245 25. Schagger, H., Aquila, H., and Von Jagow, G. (1988) Anal. Biochem. 173, 201–205 53. Ruiz, N., Falcone, B., Kahne, D., and Silhavy, T. J. (2005) Cell 121, 307–317 26. Oakley, B. R., Kirsch, D. R., and Morris, N. R. (1980) Anal. Biochem. 105, 361–363 54. Traxler, B., and Murphy, C. (1996) J. Biol. Chem. 271, 12394–12400 27. Berven, F. S., Karlsen, O. A., Murrell, J. C., and Jensen, H. B. (2003) Electrophoresis 24, 55. Powers, E. L., and Randall, L. L. (1995) J. Bacteriol. 177, 1906–1907

10034 JOURNAL OF BIOLOGICAL CHEMISTRY VOLUME 281•NUMBER 15•APRIL 14, 2006

Effects of SecE depletion on the inner and outer membrane proteomes of E. coli

Louise Baars1, Samuel Wagner1, David Wickström1, Mirjam Klepsch1, A. Jimmy Ytterberg2§, Klaas J. van Wijk2 and Jan-Willem de Gier1*

1Stockholm University, Center for Biomembrane Research, Department of Biochemistry and Biophysics, SE-106 91 Stockholm, Sweden. 2Cornell University, Department of Plant Biology, 332 Emerson Hall, Ithaca, NY 14853, USA. §Present address: University of California Los Angeles, Department of Chemistry and Biochemistry, Box 951569, Los Angeles, CA 90095-1569, USA.

*Corresponding author: Jan-Willem de Gier Tel: +46-8-162420 Fax: +46-8-153679 E-mail: [email protected]

Keywords: E. coli, membrane protein, secretory protein, Sec-translocon, SecE, proteomics

Abstract

It is generally assumed that the Sec-translocon is required for the translocation of most secretory proteins across, and the insertion of most integral membrane proteins into, the Escherichia coli inner membrane. To date, protein translocation and insertion has been studied using focused approaches and a very limited set of model substrates. To study the Sec-translocon dependence of secretory and inner membrane proteins in a global way, a comparative sub-proteome analysis of cells depleted of the essential translocon component SecE, and cells with normal levels of SecE, was carried out. The steady-state proteomes and the proteome dynamics were evaluated using 1- and 2D gel analysis, followed by mass spectrometry based protein identification and extensive immunoblotting. The analysis showed that SecE depletion 1) leads to cytosolic aggregation of secretory proteins, as well as the induction of the cytoplasmic σ32-stress response, 2) reduced the accumulation of outer membrane proteins, with the exception of OmpA, Pal and FadL, and 3) had a strong differential effect on the accumulation levels of inner membrane proteins - steady state levels and insertion of some integral inner membrane proteins were reduced, while others were not affected or even increased upon SecE depletion. The inner membrane proteins that were not affected or increased upon SecE depletion did not contain large translocated domains and/or consisted of only one or two transmembrane segments. Our study suggests that several secretory and inner membrane proteins can either make use of Sec-translocon independent pathways or have superior access to the Sec-translocon.

2 Introduction

The genome of the Gram-negative bacterium Escherichia coli harbors around four thousand open reading frames (ORFs) (1)(2). Around 25% of these ORFs encode inner membrane proteins and around 10% encode secretory (i.e., periplasmic and outer membrane) proteins (3)(4). It is generally assumed that the Sec-translocon is required for the translocation of most secretory proteins across, and the insertion of most integral membrane proteins into, the inner membrane (5)(6). The targeting of secretory proteins to the Sec-translocon is mostly post-translational and can be facilitated by the cytoplasmic chaperone SecB (7)(8)(9). Inner membrane proteins are targeted to the Sec-translocon via the SRP-pathway in a co-translational fashion (7)(6). In E. coli, a small number of proteins are translocated across, or integrated into, the inner membrane via the Twin- Arginine protein Transport (TAT)-pathway (10)(11)(12). The core of the Sec-translocon consists of the integral membrane proteins SecY, SecE, and SecG (13). SecY and SecE, but not SecG, are essential for viability (13). The crystal structure of the SecYEβ-complex from the Archaeon Methanococcus jannaschii shows that in E. coli, the ten transmembrane segments of SecY can be divided in two halves (transmembrane segment 1-5 and 6-10) that are clamped together by the third and essential transmembrane segment of SecE (14). Recent evidence suggests that although a SecYEG heterotrimer serves as the protein translocation channel, multiple SecYEG heterotrimers may cooperate in protein translocation/insertion (15)(16)(17). SecA, an ATPase that is associated with the Sec-translocon, drives the stepwise translocation of secretory proteins and large periplasmic loops of inner membrane proteins across the inner membrane (13). (13)The Sec-translocon associated proteins SecD, SecF, and YajC form a complex that facilitates protein translocation, but are not required for viability (13). The SecDFYajC complex is thought to mediate the interplay between the SecYEG-protein conducting channel and YidC, an essential inner membrane protein which appears to be involved in the transfer of transmembrane segments from the Sec-translocon into the lipid bilayer (18)(6)(19)(20). Evidence is accumulating that YidC by itself can also mediate the insertion of a subset of membrane proteins (6)(20). The notion that most secretory and inner membrane proteins require the Sec- translocon for translocation and/or insertion is based on studies using focused approaches and a limited number of model proteins, like the outer membrane protein OmpA and the inner membrane protein FtsQ (see e.g., (21)(17)). To study the Sec-translocon

3 dependence of secretory and inner membrane proteins in a more global way, we have performed a comparative sub-proteome analysis of cells depleted of SecE and cells expressing normal levels of SecE. This approach allowed us to investigate protein mis- localization, aggregation and changes in the composition of the outer and inner membrane proteomes in cells with strongly reduced Sec-translocon levels (22). Our analysis showed that upon SecE depletion, secretory proteins aggregate in the cytoplasm and the cytoplasmic σ32-stress response is induced. This response is activated upon protein misfolding/aggregation in the cytoplasm (23). Interestingly, the effects of reduced Sec-translocon levels on the proteomes of the outer and inner membranes were different. Both steady-state levels and translocation efficiencies of most outer membrane proteins were reduced. The integral inner membrane proteins showed a differential response to SecE depletion. The abundance of approximately half of the identified integral inner membrane proteins was reduced, whereas the abundance of the other inner membrane proteins was either unaffected or increased. Notably, all inner membrane proteins that were unaffected or increased upon SecE depletion lack large periplasmic domains and/or contain only one or two transmembrane segments. The 'global' analysis of cells with reduced Sec-translocon levels provides several testable hypotheses and new substrates to further discover guiding principles for protein translocation and insertion.

Experimental procedures Strains and culture conditions In E. coli strain CM124, the chromosomal copy of the gene encoding SecE is inactivated and placed on a plasmid under control of the promoter of the araBAD operon (24). CM124 was cultured in standard M9 minimal medium supplemented with thiamine (10 mM), all amino acids (0.7 mg/ml) except for methionine and cysteine, glucose (0.2% w/v), arabinose (0.2% w/v) and ampicillin (100 μg/ml) at 37ºC in an Innova 4330 (New Brunswick Scientific) shaker at 180 rpm. Overnight cultures were washed in fresh

medium without arabinose and then diluted to OD600= 0.035 in fresh medium without arabinose to deplete cells of SecE (‘SecE depleted cells’) or medium containing 0.2 % arabinose to induce expression of SecE (‘control cells’). Cells were cultured for six

hours. Growth was monitored by measuring the OD600 with a Shimadzu UV-1601 spectrophotometer.

4 SDS-PAGE, 1D Blue Native-PAGE and immunoblot analysis Immunoblot analysis was used to monitor the protein levels of SecE, SecY, SecG, SecA,

SecD, SecF, YidC, FtsQ, Lep, Fob, Foc, DegP, Skp, OmpA, OmpF, PhoE, IbpA/B, SecB,

Ffh, and PspA in whole cell lysates and/or inner membranes. Whole cells (0.1 OD600 unit), purified inner membranes (5 µg of protein) and aggregates (from 2 OD600 units of cells) were solubilized in Laemmli solubilization buffer and separated by sodium dodecyl sulfate (SDS)-PAGE. Proteins were transferred from the polyacrylamide gels to a polyvinylidene fluoride (PVDF) membrane (Millipore). Membranes were blocked and decorated with antisera to the components listed above essentially as described before (25). Proteins were detected with HRP-conjugated secondary antibodies (Bio-Rad) using the ECL system (according to the instructions of the manufacturer, GE Healthcare) and a Fuji LAS 1000-Plus CCD camera. Blots were quantified using the Image Gauge 3.4 software (Fuji). Experiments were repeated with three independent samples. To monitor the abundance of the SecYEG-protein conducting channel, inner membrane vesicles were subjected to 1D Blue Native (BN)-PAGE (26) followed by immunoblot analysis using antibodies to SecY, SecE, and SecG. Inner membrane pellets (20 μg of protein) were solubilized in buffer containing 750 mM 6-aminocaproic acid, 50 mM Bis-Tris-HCl (pH 7.0 at 4°C) and freshly prepared 0.5% (w/v) n-dodecyl-β-D- maltopyranoside (DDM). After removal of unsolubilized material by centrifugation (100.000 x g, 30 minutes), Serva Blue G was added to a final concentration of 0.5% (w/v) and the samples were loaded onto the first dimension gel. The 0.02% Serva Blue G cathode buffer of the BN-PAGE was exchanged to a 0.002% Serva Blue G cathode buffer after 1/3 of the run in order to prevent excessive binding of Coomassie dye to the PVDF membrane in the subsequent transfer step. Ferritin (440 & 880 kDa), aldolase (158 kDa) and albumin (66 kDa) (GE Healthcare) were used as molecular weight markers. Proteins were transferred to PVDF membrane, detected by antisera to SecY, SecE, and SecG and quantified as described above.

Protein translocation assay Translocation of OmpA was monitored essentially as described previously (27). Cultures 35 corresponding to 0.4 OD600 unit were labeled with [ S]methionine (60 µCi/ml, 1 Ci=37 GBq) for 30 seconds followed by precipitation in 10% trichloroacetic acid (TCA), either directly or after a chase with cold methionine (final concentration 0.5 mg/ml) for 3 and 10 minutes. TCA-precipitated samples were washed with acetone, resuspended in 10 mM

5 Tris-HCl (pH 7.5), 2% SDS and immunoprecipitated with antiserum to OmpA. The OmpA precipitate was subjected to standard SDS-PAGE analysis. Gels were scanned in a Fuji FLA-3000 phosphorimager and quantified as described above.

Flow cytometry and microscopy Analysis of SecE depleted and control cells using flow cytometry was carried out using a FACSCalibur (BD Biosciences) instrument. To assess viability, cells were incubated in the dark at room temperature with 30 μM of propidium iodide (PI) for 15 minutes (28). For staining of the inner membrane, cells were cultured at 37ºC for 30 minutes with 2 µM of the membrane-specific fluorophore FM4-64 (Invitrogen) (29). Cultures were diluted in ice cold PBS to a final concentration of approximately 106 cells per ml. A low flow rate was used throughout data collection with an average of 250 events per second. Forward and side scatter acquisition was used for comparison of cell morphology (9). Data acquisition was performed using CellQuest software (BD Biosciences) and data were analyzed with FloJo software (Tree Star). For microscopy, cells were mounted on a slide and immobilized in 1% low melting temperature agarose. Microscopy was performed on a Zeiss Axioplan2 fluorescence microscope equipped with an Orca-ER camera (Hamamatsu). Images were processed with the AxioVision 4.5 software from Zeiss.

Isolation and analysis of protein aggregates Protein aggregates were extracted from whole cells essentially as described before (30).

Cells corresponding to 75 OD600 units were used for each aggregate extraction. The protein content of cell lysates and aggregate extracts was determined with the BCA assay according to the instructions of the manufacturer (Pierce). Aggregates were analyzed by SDS-PAGE using 24 cm long 8-16% acrylamide gradient gels. Gels were stained with Coomassie Brilliant Blue R-250 and proteins were identified by mass spectrometry (MS) as described below. The aggregate fraction was also subjected to in solution digest followed by nano-liquid chromatography electrospray tandem MS (nanoLC-ESI-MS/MS) essentially as described before (31).

Isolation of inner and outer membranes Inner and outer membranes were isolated essentially as described before (32). Membrane fractions used for immunoblot analysis were prepared from non-radio-labeled cultures.

6 Membrane fractions used for analysis by 2D-gel electrophoresis (outer membranes) or 2D-BN-PAGE (inner membranes) were prepared from a mixture of labeled and unlabeled

cells as outlined in the Supplemental figure 1. Cells corresponding to 1000 OD600 units

were cultured as described above. An aliquot of 10 OD600 units of cells was labeled with [35S]methionine (60 µCi/ml, 1 Ci=37 GBq) for 1 minute, followed by a chase of 10 minutes with cold methionine (final concentration 5 mg/ml). Labeled cells were subsequently collected by centrifugation and cell pellets were snap-frozen in liquid nitrogen. The rest of the cells (990 OD600 units) was harvested by centrifugation and washed once with buffer K (50 mM triethanolamine (TEA), 250 mM sucrose, 1 mM ethylenediaminetetraacetic acid (EDTA), 1 mM dithiothreitol (DTT), pH 7.5). The cell pellets were snap-frozen in liquid nitrogen and stored at -80°C. Before breaking the cells, labeled and unlabeled cells from the same culture were pooled in a 1:100 ratio. The resulting mixture was resuspended in 8 ml buffer K supplemented with 0.1 mg/ml Pefabloc and 5 µg/ml DNAse and lysed by two cycles of French press (18,000 psi). The lysate was cleared of unbroken cells by centrifugation at 8000 x g for 20 minutes, and the total membrane fraction was collected by centrifugation at 100.000 x g for 1 hour. The membrane pellet was resuspended in 1 ml of buffer M (50 mM TEA, 1 mM EDTA, 1 mM DTT, pH 7.5) and loaded on top of a six-step sucrose gradient (from bottom to top); 0.5 ml 55%, 1.5 ml 50%, 1.5 ml 45%, 2.5 ml 40%, 2.5 ml 35%, 2.5 ml 30% (w/w sucrose in buffer M). After centrifugation at 210,000 x g for 15 hours, the inner membrane and outer membrane fractions were collected from the 35% and 45 % sucrose layers, respectively. The collected fractions were diluted in TEA buffer (50 mM TEA, 1 mM DTT, pH 7.5) to a sucrose concentration below 10%. Membranes were collected by centrifugation at 170.000 x g for 1 hour and subsequently resuspended in buffer L (50 mM TEA, 250 mM Sucrose, 1 mM DTT, pH 7.5). The inner membrane fraction was snap-frozen in liquid nitrogen and the outer membrane fraction was washed in 0.1 mM sodium carbonate as described before (9). Protein concentrations were determined using the BCA assay. Samples were stored at -80°C.

2-dimensional gel-electrophoresis (2DE) 35 Whole cell lysates (1 OD600 unit) and [ S]methionine labeled outer membranes (185 μg of protein) isolated by density centrifugation were analysed by 2DE using iso-electric focusing in the first dimension and SDS-PAGE in the second dimension (9). Gels used for comparative analysis of whole cells lysates were stained with high sensitivity silver

7 stain (33). Gels used for comparative analysis of the outer membrane proteome and all gels used for MS based identification of proteins were stained with colloidal Coomassie (34). Most proteins in the outer membrane gels gave rise to multiple spots with the same molecular mass but different pI. This phenomenon was also observed in the outer membrane map of E. coli constructed by Molloy et al. (35). Most of these 'trains of spots' are caused by modifications induced during sample preparation (36), likely due to stepwise deamidation of the asparagine and glutamine residues, resulting in loss of 1 Dalton and net loss of one positive charge (37).

Analysis of cytoplasmic membrane fractions by 2D BN-PAGE Comparative 2D Blue Native electrophoresis was performed as described previously (32). In short, [35S]methionine labeled inner membranes (100 µg of protein) were solubilized in 0.5% (w/v) DDM and subjected to Blue-Native electrophoresis in the first dimension and denaturing SDS-PAGE in the second dimension. For calibration, ferritin (440 & 880 kDa), aldolase (158 kDa) and albumin (66 kDa) (GE Healthcare) were used as molecular weight markers. Gels were stained with Coomassie Brilliant Blue R-250 (9).

Image analysis and statistics Stained gels were scanned using a GS-800 densitometer from BioRad. Radio-labeled gels were scanned in a Fuji FLA-3000 phosphorimager. Spots were detected, matched and quantified using PDQuest software version 8.0 (BioRad). The analysis of Coomassie stained and [35S]methionine labeled outer and inner membrane proteins was done on the same set of gels. In all cases, each analysis set consisted of at least three gels in each replicate group (i.e., SecE depleted cells and the control). Each gel in a set represented an independent sample (i.e., from a different bacterial colony, culture and membrane preparation). Independent samples were subjected to 2DE or 2D BN-PAGE and image analysis in parallel, i.e., en groupe. Quantities of stained spots were normalized using the ‘total intensity of valid spots’ method to compensate for non-expression related variations in spot quantities between gels (there were no significant variations in the total spot quantity between the two groups; SecE depletion and control). Since protein aggregates can co-sediment with outer membranes during density gradient centrifugation, an additional normalization step was required for the analysis of the outer membrane gels (38)(39). First, to distinguish between outer membrane spots and contaminating aggregate spots, the outer membrane fraction from SecE depleted and control cells were

8 subjected to aggregate extraction as described above. The resulting extracts were analysed by 2DE and proteins in the aggregates were visualized by staining with colloidal Coomassie. Spots detected in the gels of the aggregate extract were removed from the outer membrane analysis set if the intensity of a spot in the aggregate gel was more than 5% of the intensity of the spot detected in the outer membrane gels. The quantities of the remaining spots were normalized using the ‘total quantity of valid spot’ method to correct for the contribution of protein aggregates on protein loading. Spots detected by means of phosphorimaging were normalized using the correction value calculated from the corresponding Coomassie stained gel to allow correction for errors in protein loading while retaining differences in labeling efficiency between the control and cells depleted of SecE. PDQuest was set to detect differences that were found to be statistically significant using the student t-test and a 95% level of confidence, including qualitative differences (“on – off responses’’) present in all gels in a group. Saturated spots were excluded from the analysis.

Mass spectrometry based identification of proteins Coomassie stained protein spots or bands were excised, washed, digested with modified trypsin and peptides were extracted manually or automatically (ProPic and Progest, Genomic Solutions, Ann Arbor, Mi). Peptides were applied to the MALDI target plate as described previously (40). Mass spectra were obtained automatically by MALDI-TOF MS in reflectron mode (Voyager-DE-STR; PerSeptive Biosystems, Framingham, MA), followed by automatic internal calibration using tryptic peptides from autodigestion. The spectra were analyzed for monoisotopic peptide peaks (m/z range 850-5000) using the software MoverZ from Genomic Solutions (http://65.219.84.5/moverz.html) with a signal to noise ratio threshold of 3.0. Matrix and/or auto-proteolytic trypsin fragments were not removed. Spectral annotations (in particular assignments of mono isotopic masses) were verified by manual inspection for a large number of measurements. The resulting peptide mass lists were used to search the SwissProt 45.0 database (release 10/04) for E. coli with Mascot (v2.0) in automated mode (www.matrixscience.com), using the following search parameters/criteria: significant protein MOWSE score at P<0.05; no missed cleavages allowed; variable methionine oxidation; fixed carbamidomethylation of cysteins; minimum mass accuracy 50 ppm. The search results pages were extracted and analyzed by an additional in-house filter (Sun and van Wijk, unpublished) applying the following three criteria for positive identification: i) Minimum MOWSE score ≥50;ii) ≥four

9 matching peptides with an error distribution within ±25 ppm; iii) ≥15% sequence coverage. False positive rates were less than 1%, as determined by searching with the .pkl list against the E. coli database (SwissProt 48.1) mixed with a randomized version of the E. coli database, generated using a Perl script from Matrix science.

Results Characterization of the SecE depletion strain CM124 The E. coli strain CM124 was used to study protein translocation and insertion upon depletion of SecE. In CM124, the chromosomal copy of secE is inactivated and a copy of secE is placed on a plasmid under control of the promoter of the araBAD operon (41). Cells were cultured aerobically in M9 minimal medium in the presence of arabinose to induce expression of SecE (these cells will be further referred to as ‘control cells’) and in the absence of arabinose to deplete cells of SecE. Growth was monitored by measuring

the OD600 (Figure 1A). As expected, growth of CM124 cells cultured in the absence of arabinose was much slower than in control cultures. For all experiments, cells were harvested six hours after inoculation. At this time-

point, control cells are in the mid-log phase with an OD600 of 0.8, and SecE depleted cells

reached an OD600 of 0.3-0.4. Inner membranes were isolated from control cells and SecE depleted cells and the levels of SecE, SecY, SecG, SecA, SecD, SecF, and YidC were analysed by immunoblotting. The level of SecE in the membrane of depleted cells was less than 10% of the SecE detected in membranes prepared from control cells. It should be noted that the level of SecE in the membrane of the control cells was similar to the level of SecE in the strain CM124 is derived from (data not shown). The levels of SecY and SecG in SecE depleted membranes were reduced to 50% and 80%, respectively (Figure 1B). SecY is degraded by the FtsH protease in the absence of SecE (22). Since the Sec-translocon is composed of SecY, SecE, and SecG in a 1:1:1 ratio (14), this indicates that the SecE depleted membrane must contain pools of SecY and SecG, which are not in a SecYEG-complex. The abundance of SecYEG-heterotrimers in SecE depleted membranes was monitored by BN-PAGE combined with immunoblotting (Figure 1C). For this purpose, inner membranes prepared from SecE depleted and non- depleted control cells were solubilized in 0.5% DDM (42). Upon depletion of SecE, the amount of the SecYEG heterotrimer was strongly reduced. Interestingly, a complex that most likely represents a SecYG heterodimer could be detected in membranes of SecE depleted cells but not in control membranes.

10 The amount of SecA was almost doubled in SecE depleted membranes (Figure 1B)(43). The levels of SecD, SecF, and YidC were reduced to approximately 50% upon SecE depletion. In addition, the steady state levels of the well studied model proteins

FtsQ, Lep, Fob, and Foc were monitored by immunoblotting (Figures 1B)(21)(44)(45)(46)(47). As expected, the accumulation levels of the Sec-translocon dependent inner membrane proteins FtsQ and Lep were strongly reduced upon SecE

depletion, whereas the level of the Sec-translocon independent protein Foc was only

slightly increased. To our surprise, the level of Fob was somewhat increased. This is

contradictory to previous work proposing that translocation of Fob is dependent on the Sec-translocon (45). Upon SecE depletion, translocation of OmpA was delayed but not abolished (Figure 1D). Notably, the intensity of the total OmpA signal was clearly increased in SecE depleted cells compared to control cells, indicating that OmpA expression was induced. We do not have a ready explanation for this observation.

Flow cytometry and microscopy The integrity of the inner membrane of SecE depleted and control cells was monitored using propidium iodide (PI) staining combined with flow cytometry (28). 9.0% (+/- 2.0%) of SecE depleted cells stained fluorescently red with PI, compared to 1.0% (+/- 0.3%) of the control cells, indicating that SecE depletion did not have a major impact on the integrity of the inner membrane. Furthermore, we detected a small increase of both the forward scatter and side scatter of cells depleted for SecE (Figure 2A). This indicates that SecE depleted cells are slightly bigger than control cells and most likely contain extra internal structures (i.e., extra membranes and/or protein aggregates). Light microscopy showed that SecE depleted cells were slightly elongated compared to the control cells (Figure 2A). The inner membranes of SecE depleted cells and control cells were stained with the fluorescent dye FM4-64 and cells were analyzed by flow cytometry (29). The fluorescence of SecE depleted cells was enhanced approximately four times compared to the control cells (Figure 2B). This is in keeping with the observation that SecE depletion induces the formation of endoplasmic membranes (48).

11 SecE depletion leads to accumulation of secretory proteins in the cytoplasm and the induction of the σ32-stress response Whole cell lysates of SecE depleted and control cells were compared by 2DE and immunoblot analysis. The comparative 2DE analysis was based on four biological replicates. Proteins were separated by denaturing immobilized pH gradient (IPG) strips (pH 4-7) in the first dimension and by Tricine-SDS-PAGE in the second dimension. Gels were stained with silver or colloidal Coomassie and spot volumes were compared using PDQuest. This analysis demonstrated that the volumes of 28 spots were significantly (P<0.05) changed in the lysates of SecE depleted cells compared to the control; the intensity increased in 13 spots and decreased in 15 spots. The affected spots were excised and used for protein identification by matrix assisted laser desorption/ionization mass spectrometry (MALDI-TOF MS) and peptide mass finger printing (PMF) (Figure 3A, Table 1). Spot statistics and MS data are provided in Supplemental table 1. The effects of SecE depletion on protein accumulation levels are shown as fold changes (SecE depletion/control) in Figure 3B. Accumulation levels of a number of secretory proteins (β-lactamase, DppA, FliY, LivJ, PotD, OmpC, OmpT, OppA, RbsB, TolB, UshA, YbiS, YehZ, YggE, YodA, and ZnuA) were reduced in SecE depleted cells. Based on the pI and molecular weight, most of these spots corresponded to the mature forms of the proteins (Table 1). MetQ, OmpC, OmpA, and RbsB were identified in spots that were stronger in lysates of SecE depleted cells. The OmpC spot corresponded to a degraded form of the protein while the increased RbsB spot most likely corresponded to the precursor form. The OmpA spot likely corresponded to the precursor form of OmpA, since we observed a peptide mass matching with a predicted tryptic peptide within the signal sequence. To study the effect of SecE depletion in more detail, the accumulation levels of the periplasmic proteins DegP and Skp and the outer membrane proteins OmpA, OmpF, and, PhoE were monitored by immunoblotting (Figure 3C). Upon SecE depletion, the precursor form of all these proteins was detected, indicating accumulation in the cytoplasm due to hampered translocation. In the case of Skp, DegP, and PhoE, this was accompanied by a decrease of the mature form of the proteins. Interestingly, the total levels of DegP and Skp were not significantly affected upon SecE depletion. This suggests that no extracytoplasmic stress responses are activated upon SecE depletion (49). The accumulation levels of the mature form of OmpA and OmpF were unaffected by the SecE depletion.

12 Upon SecE depletion, accumulation levels of the σ32-inducible, cytoplasmic chaperones DnaK, GroEL, GroES and ClpB were increased (Figure 3B). The up- regulation of DnaK and GroEL was confirmed by Western-blotting (results not shown). Since inclusion body proteins IbpA/B, SecA, SecB, Ffh, and phage shock protein A (PspA) were not identified in the 2D gels, we monitored their accumulation levels by immunoblotting (Figure 3D). The level of the heat shock chaperones IbpA/B, also part of the σ32-regulon, was increased. Inclusion body proteins associate with protein aggregates and facilitate the extraction of proteins from aggregates by ClpB and DnaK (50). In agreement with the membrane blotting experiments (see above), the total level of SecA was increased in SecE depleted cells, consistent with insufficient Sec-translocon capacity (43). Accumulation levels of SecB and Ffh, components involved in the targeting of secretory and inner membrane proteins to the Sec-translocon, respectively, were both unaffected upon SecE depletion. This indicates that the protein targeting capacity is not affected upon SecE depletion. The level of PspA was monitored since the electrochemical potential plays an important role in protein translocation and the expression of PspA is up- regulated when it is affected. Just like in several other translocation and insertion-mutant strains, a considerable PspA response was detected in SecE depleted cells (51). Taken together, the up-regulation of SecA and the accumulation of the unprocessed forms of secretory proteins indicate that protein translocation across the cytoplasmic membrane is strongly hampered. Furthermore, the accumulation levels of the σ32-regulated chaperones DnaK, GroEL/ES, ClpB, and IbpA/B are all increased upon SecE depletion, suggesting that reduced Sec-translocon levels lead to protein miss- folding/aggregation in the cytoplasm.

Accumulation of cytoplasmic protein aggregates in SecE depleted cells Protein aggregates were extracted from whole cells depleted of SecE. The aggregates from SecE depleted cells contained 2.6% of the total cellular protein compared to 0.3% in the control. The protein composition of the aggregates was analysed by 1D gel electrophoresis combined with MALDI TOF MS PMF (Figure 4, Supplemental table 2) and by nanoLC-ESI-MS/MS of solubilized aggregates digested with trypsin (Supplemental table 2). In total, 61 proteins were identified in aggregates isolated from cells depleted of SecE; 19 secretory proteins, 5 inner membrane proteins, 36 cytoplasmic proteins and one protein with a localization that could not be unambiguously predicted (Table 2). Among the identified cytoplasmic proteins were the chaperones IbpA, DnaK

13 and DnaJ. The MS/MS analysis revealed that at least four of the secretory proteins, OmpA, Lpp, β-lactamase and SlyB, contained an uncleaved signal sequence (results not shown), indicating that these proteins aggregate in the cytoplasm rather than the periplasm. The identified inner membrane proteins ElaB and YqjD contain one predicted transmembrane segment, while YhjK contains two. The Penicillin-binding protein 5 (DacA), Penicillin-binding protein 6 (DacC) are probably attached to the inner membrane via a C-terminal amphiphilic α-helix (52). It is possible that the number of identified inner membrane proteins is somewhat under-represented due to experimental problems associated with MS based identification of α-helical membrane proteins (53). Intensities of the bands in the gel shown in Figure 4 were quantified to get an idea of the relative abundance of different classes of proteins in the aggregates isolated from SecE depleted cells. 70% represented secretory proteins and 18% represented cytoplasmic proteins. The cytoplasmic chaperones DnaK and IbpA together constituted 2% of the total band intensity. The protein content of the remaining bands is unknown.

Effect of SecE depletion on the outer and inner membrane proteomes To study the effect of SecE depletion on the insertion and composition of the inner and the outer membrane proteomes, cells were labeled with [35S]methionine. Outer and inner membranes were subsequently isolated using a combination of French-press and sucrose gradient centrifugation (Supplemental figure 1). The outer membrane proteome was analysed by 2DE using IEF in the first dimension and SDS-PAGE in the second dimension (9). The inner membrane proteome was analysed with backed 2D BN-PAGE that allows relative quantification (32). Gels were stained with colloidal Coomassie and [35S]methionine labeled proteins were detected by phosphorimaging. Spot intensities were quantified and compared using PDQuest. Each analysis set contained at least three biological replicates and the threshold for acceptance was 95% significance determined by the student-t test. Spots were excised and used for protein identification by MALDI- TOF MS and PMF.

Outer membrane proteome – MS analysis of the Coomassie stained spots in the 2D gels of the outer membrane proteome resulted in the identification of 39 different proteins from 51 spots (Figure 5A, Supplemental table 3). 40 of these spots could be matched to spots detected by phosphorimaging (Figure 5B). We found that the outer membrane fraction of SecE depleted cells was contaminated with aggregates that can co-sediment

14 with outer membranes during density gradient centrifugation (38)(39). The spots corresponding to aggregated proteins were identified and removed from the analysis set as described in the ‘Experimental procedures’ (Supplemental table 3, Supplemental figure 2). The bar diagram in Figure 5C shows the average fold-change of the spot intensities (SecE depletion/control) for proteins detected by Coomassie (black) or phosphorimaging (grey). Statistically significant (P<0.05) fold-changes are indicated by bold numbers in Table 3 and Supplemental table 3. The steady-state levels and insertion efficiencies of most outer membrane proteins were reduced upon SecE depletion. Three outer membrane proteins were unaffected or slightly increased upon SecE depletion; OmpA, Pal, and FadL. After the 10 minutes chase, the level of [35S]methionine labeled OmpA detected in the outer membrane of SecE depleted cells was not significantly affected. Furthermore, the steady-state level of OmpA was increased by approximately 20% (Figure 5A, Table 3, Supplemental table 3). This was in agreement with the pulse chase analysis shown in Figure 1D, which demonstrates that translocation of OmpA was slowed down but not abolished upon SecE depletion.

Inner membrane proteome - MS analysis identified proteins in 85 spots of the Coomassie stained 2D BN-PAGE gels. 28 additional spots could be annotated with the help of our previously published reference map (Figure 7A, Table 4, Supplemental table 4)(32). 56 proteins were integral membrane proteins and 27 were proteins located at the cytoplasmic side of the inner membrane, mostly as part of membrane localized complexes. In addition, five secretory proteins were identified in the 2D BN-PAGE gels. The bar diagram in Figure 6C shows the effect of the SecE depletion as fold- changes calculated from the average spot intensities (SecE depletion/control) of stained and [35S]methionine labeled proteins. Statistically significant (P<0.05) fold-changes are indicated by bold numbers in Table 4 and Supplemental table 4. Since several proteins were identified in more than one spot, we also calculated the effect of the SecE depletion on the total level of each identified integral membrane protein (Supplemental table 5). The total levels of approximately 30 integral membrane proteins were reduced in the membrane of SecE depleted cells. Notably, the steady levels of all components of the FtsH-HflKC protease complex were significantly reduced (Table 4, Supplemental table 4). It is possible that this could affect the stability of the inner membrane proteome, although it should be noted that FtsH-HflKC mediated proteolysis has only been shown

15 for a few membrane proteins (54). The cytochrome bo3 terminal oxidase subunits CyoA and CyoB were reduced by 75% and 85%, respectively. The biogenesis of CyoA, which is a lipid modified integral membrane protein, has recently been shown to depend on both YidC and the Sec-translocon (55)(56)(57). Interestingly, the accumulation levels of a surprisingly large number of integral membrane proteins (approximately 30) were not significantly affected or even increased by SecE depletion (Supplemental table 5). Among the significantly increased proteins were; Aas, AtpF, MscS, MgtA, NarI, YbbK, YhcB, YhjG, and YajC. We tested if the effect of SecE depletion could be correlated to different membrane protein properties. Our analysis showed that protein abundance, topology, hydrophobicity, and the energy

required for membrane integration of the first transmembrane domain (ΔGapp) (58) do not correlate with the effect on total protein levels upon SecE depletion (data not shown). However, when the fold-changes (Coomassie staining) were plotted against the number of amino acids in the largest translocated domain of each protein, we found that almost all proteins with large periplasmic domains are sensitive to SecE depletion (Figure 7A, Supplemental table 5). In contrast, almost all proteins that were positively affected by SecE depletion do not contain any large periplasmic domains. A closer look at the proteins that do not follow this trend (Aas, YhjG, YbbK, and YhcB) revealed that they consist of only one or two transmembrane segments. This prompted us to perform a combined analysis of the effect of number of transmembrane segments and the size of periplasmic domains. Based on the plot shown in figure 7A, we divided the proteins into two groups; proteins with large translocated domains (≥60 amino acids) and proteins with small translocated domains (≤60 amino acids). The 60 amino acid cut-off for Sec- dependence is in agreement with previous studies (59). The effect of SecE depletion on these two groups was plotted against the number of transmembrane segments (Figure 7B, Supplemental table 5). This clearly demonstrated that proteins that do not have large periplasmic domains are overrepresented among the proteins that are either unaffected (fold change ≥0.75≤1.25) or positively affected (fold change ≥1.25) by SecE depletion. The few exceptions are proteins that contain only one or two transmembrane segments. Collectively, our analysis suggests that many proteins that do not contain large periplasmic domains and/or contain one or two transmembrane segment(s) can insert efficiently into the membrane when the levels of the Sec-translocon are reduced.

16 Discussion So far, the Sec-translocon dependence of protein translocation and insertion in E. coli has been studied using focused approaches and a very limited set of model substrates. To characterize the Sec-translocon dependence of secretory and inner membrane proteins in a global way, we have performed a comparative sub-proteome analysis of E. coli cells depleted of the Sec-translocon component SecE. Depletion of SecE resulted in 90% (10- fold) reduction in SecE and a 50%-70% (2-fold) reduction in the SecY,G translocon components. 1D BN-PAGE combined with immunoblotting showed that the level of the SecYEG-complex was strongly reduced upon SecE-depletion. Furthermore, accumulation of SecA was increased 1.7 fold. This indicates that translocation of the translocation monitor SecM is hampered due to insufficient Sec-translocon capacity, leading to increased expression of SecA (43). SecE depletion did not affect the accumulation levels of SecB or Ffh, which are both involved in protein targeting. Our analysis of the subproteomes of cells with strongly reduced Sec-translocon levels resulted in three main observations and conclusions. Reduced Sec-translocon levels 1) resulted in the accumulation of secretory proteins in the cytoplasm, the formation of protein aggregates and a σ32-response, 2) negatively affected levels of all constituents of the outer membrane proteome, with the exception of OmpA, Pal, and FadL, and 3) had differential effects on inner membrane proteins - steady state levels and insertion of some integral inner membrane proteins were reduced, while others were not affected or even increased in the membranes of SecE depleted cells. The proteins that were not affected or increased upon SecE depletion did not contain large translocated domains and/or consisted of only one or two transmembrane segments. Below these main observations and conclusions are explained and discussed in more detail.

Reduced Sec-translocon levels induce the formation of protein aggregates Upon SecE depletion, secretory proteins accumulate in the cytoplasm, leading to aggregate formation and the induction of a σ32-response. It was estimated that secretory proteins made up 70% of the total protein in the aggregates. The MS/MS analysis revealed that at least three secretory proteins, OmpA, Lpp, and β-lactamase, contained an uncleaved signal sequence, pointing to aggregation in the cytoplasm rather than the periplasm. Sequence analysis of all the aggregated secretory proteins with the aggregation propensity prediction program Tango showed that their signal sequences are more aggregation prone than the mature part of these proteins ((60), (data not shown)).

17 This suggests that secretory proteins are more prone to aggregation in the cytoplasm than in the periplasm. The accumulation of secretory proteins upon SecE depletion induced a σ32-response, leading to increased levels of the cytoplasmic chaperones IbpA/B, DnaK, GroEL and ClpB. These chaperones protect proteins from aggregation and are also involved in the disaggregation, refolding and degradation of aggregated proteins (61)(62)(63). IbpA/B, DnaK, GroEL and the Lon protease were among the proteins that were identified in the aggregate fraction. Thus, aggregated secretory proteins may either be actively reactivated for translocation or degraded. Recently, we have shown that in an E. coli secB null mutant in the cytoplasm aggregated OmpA is extracted from aggregates (9). Although our analysis did not provide any evidence for aggregate formation in the periplasm we cannot exclude this. Only a few inner membrane proteins were identified in the aggregates. This may be explained by the efficient degradation by SsrA mRNA dependent tagging of stalled nascent chains of co-translationally targeted membrane proteins, and subsequent turnover by proteases (64)(65). Sec-translocon independent membrane insertion mechanisms could also explain the low abundance of inner membrane proteins in the aggregates (see below).

SecE depletion reduces the insertion and steady-state levels of most outer membrane proteins A direct correlation between the effects on the outer membrane proteome and Sec- translocon dependence is difficult to make since key players involved in outer membrane protein biogenesis – like YaeT and Skp (66) – are affected by SecE depletion. Nevertheless, the analysis of the outer membrane proteome indicates that translocation of proteins across the inner membrane is hamped but not blocked upon SecE depletion. The components of the outer membrane proteome were differentially affected by the depletion of SecE. OmpA, FadL, and Pal were unaffected or increased in the outer membrane upon SecE depletion, while most other outer membrane proteins were reduced to different extents. One explanation is superior accessof these proteins tothe Sec- translocon. Signal peptide based selective modulation of protein translocation occurs in the ER during stress (67). If such a mechanism for modulation of translocation efficiency exists in E. coli, it should become apparent upon lowering Sec-translocon levels. We were not able to identify any signal sequence characteristics (e.g., hydrophobicity, charge

18 distribution, length) that correlated with the differential effects on the constituents of the outer membrane proteome upon depletion of SecE (data not shown). Using a small number of model proteins it has been shown that DnaK can keep outer membrane proteins, but not periplasmic proteins, in a prolonged export competent state upon depletion of SecA (68). This suggests that affinity towards DnaK and other chaperones could also affect the translocation efficiency of secretory proteins during SecE depletion. However, it should be noted that both periplasmic and outer membrane proteins were identified in aggregates from SecE depleted cells (Figure 4, Table 2). We were not able to extend our analysis to include the periplasmic proteome, since it was not possible to isolate sufficiently pure periplasmic fractions from SecE depleted cells. Thus, it is not clear if the outer membrane proteome is in fact less affected than the periplasmic proteome or if the chaperone mediated protection of secretory proteins is independent of the final destination of the protein. It is conceivable, that proteins that under normal conditions use the Sec- translocon can cross the membrane via alternative pathways upon SecE depletion. It has been shown that there are secretory proteins that are promiscuous; i.e., can use both the Sec- and TAT-protein translocation pathways (12). In this respect it should be noted that the TAT-pathway is still operationalupon SecE depletion (69). Interestingly, our Blue Native analysis revealed that SecA dimers accumulated at the inner membrane of SecE depleted cells. Impaired Sec-translocon function results in increased expression of SecA, mediated by the secretion monitor SecM (43). SecA is the peripheral subunit of the Sec- translocon and responsible for the ATP dependent translocation of secretory proteins and large periplasmic domains of integral membrane proteins. The increased levels of SecA may enhance translocation efficiency when the pressure on the translocon is particularly high. Recently, it has been shown that the Sec-translocon catalyzes the monomerization of the SecA dimer (70). This could explain the accumulation of SecA dimers at the membrane observed upon SecE depletion. It has also been proposed that the SecA dimer by itself can act as an alternative translocase for secretory proteins (71). If this is indeed the case, it could mean that the SecA dimer functions as a backup translocon when Sec- translocon capacity is not sufficient.

SecE depletion has differential effects on the inner membrane proteome Depletion of SecE resulted in reduced steady state levels and integration of approximately 30 integral membrane proteins, while another 30 were not significantly

19 affected or even increased. The immunoblot and 2D BN-PAGE analysis showed that

FtsQ, Lep, and the cytochrome bo3 subunit CyoA were all strongly reduced in the inner membrane of SecE depleted cells. These proteins have been shown to require both the Sec-translocon and YidC for proper assembly into the inner membrane (21)(44)(56)(55). A surprisingly large number of inner membrane proteins were either unaffected or even increased in the membrane of SecE depleted cells. Among the unaffected proteins was the Foc subunit of the Fo sector of the ATP synthase. This is in keeping with previous studies showing that Foc is inserted into the inner membrane in a Sec-translocon

independent but YidC dependent fashion (45)(46)(47). The efficient insertion of Foc demonstrates that the YidC pathway was operational, although YidC levels were 50% reduced upon SecE depletion. Interestingly, the level of Foa and Fob, also components of the Fo sector of the ATP synthase, were slightly increased upon SecE depletion (Table 4

and Figure 1C, respectively). It has been proposed that insertion of both Foa and Fob is dependent on the Sec-translocon (45). However, it should be noted that integration was studied in cells depleted of SecDF rather than SecE/Y (45). Furthermore, Fob integration

was examined using a Fob variant with a N-terminal T7 tag, which may affect the biogenesis requirements of Fob. The analysis of the inner membrane proteome of SecE depleted and control cells allowed us to search for common features among proteins that were reduced, unaffected or increased in the membrane of SecE depleted cells. We found no correlations between the effect of SecE depletion and properties of the first transmembrane segment (e.g.,

hydrophobicity, ΔGapp for insertion (58)) as could be expected if the differences were due to different affinities towards the residual translocons. However, we found that the inner membrane proteins that were either unaffected or increased upon depletion of SecE, do not contain any large periplasmic domains and/or consist of only one or two transmembrane segments (Figure 7, Supplemental table 5). Interestingly, all the proteins that so far have been shown to integrate via the Sec-translocon independent/YidC

dependent pathway (M13, Pf3, Foc, MscL and a C-terminally truncated ProW variant) share similar features (20). Thus, it is tempting to speculate that the proteins that are unaffected or increased in the membrane of SecE depleted cells are potential substrates of the YidC only pathway. However, it should be noted that membrane integration of the E. coli inner membrane protein KdpD, which consists of four transmembrane segments and exceptionally big cytoplasmic N- and C-terminal domains, is not affected by either SecE or YidC depletion (72). Based on these observations it has been proposed that an inner

20 membrane assembly pathway, which is independent of both the Sec-translocon and YidC, may exist in E. coli (72). It is also possible that some integral membrane proteins, just like some secretory proteins, are promiscuous; i.e., use the insertion pathway that is available. Clearly, the observation that the levels of such a large number of inner membrane proteins are not affected or go up upon SecE depletion is intriguing and warrants further investigations. For instance, it would be interesting to use a proteomics approach to analyze membrane protein biogenesis in YidC depletion and SecE/YidC double depletion backgrounds.

In conclusion, substantial protein translocation and insertion activity was still observed in SecE depleted cells. This suggests that the significance of Sec-translocon independent translocation/insertion and pathway promiscuity in outer and inner membrane protein biogenesis have been underestimated. Our study provides several testable hypotheses and new substrates to further discover guiding principles for protein translocation and insertion in the model organism E. coli.

Acknowledgements

Claudia Wagner is thanked for assistance with the flow cytometry experiments. Dirk-Jan Scheffers and Joen Luirink are thanked for critically reading the manuscript. This research was supported by grants from the Swedish Research Council, the Carl Tryggers Stiftelse, the Marianne and Marcus Wallenberg Foundation and the SSF supported Center for Biomembrane Research to JWdG, and a grant from The Swedish Foundation for International Cooperation in Research and Higher Education (STINT) to JWdG and KJvW. Proteomics infrastructure was supported by a grant from NYSTAR to KJvW.

21 References

1. Blattner, F. R., Plunkett, G., Bloch, C. A., Perna, N. T., Burland, V., Riley, M., Collado-Vides, J., Glasner, J. D., Rode, C. K., Mayhew, G. F., Gregor, J., Davis, N. W., Kirkpatrick, H. A., Goeden, M. A., Rose, D. J., Mau, B., and Shao, Y. (1997) Science 277, 1453-1462 2. Baba, T., Ara, T., Hasegawa, M., Takai, Y., Okumura, Y., Baba, M., Datsenko, K. A., Tomita, M., Wanner, B. L., and Mori, H. (2006) Mol Syst Biol 2, 2006 0008 3. Daley, D. O., Rapp, M., Granseth, E., Melen, K., Drew, D., and von Heijne, G. (2005) Science 308, 1321-1323 4. Rey, S., Acab, M., Gardy, J. L., Laird, M. R., deFays, K., Lambert, C., and Brinkman, F. S. (2005) Nucleic Acids Res 33, D164-168 5. Osborne, A. R., Rapoport, T. A., and van den Berg, B. (2005) Annu Rev Cell Dev Biol 21, 529-550 6. Luirink, J., von Heijne, G., Houben, E., and de Gier, J. W. (2005) Annu Rev Microbiol 59, 329-355 7. Driessen, A. J. M., Manting, E. H., and van der Does, C. (2001) Nature Structural Biology 8, 492-498 8. Randall, L. L., and Hardy, S. J. (2002) Cell Mol Life Sci 59, 1617-1623 9. Baars, L., Ytterberg, A. J., Drew, D., Wagner, S., Thilo, C., van Wijk, K. J., and de Gier, J. W. (2006) J Biol Chem 281, 10024-10034 10. Hatzixanthis, K., Palmer, T., and Sargent, F. (2003) Mol Microbiol 49, 1377-1390 11. Lee, P. A., Tullman-Ercek, D., and Georgiou, G. (2006) Annu Rev Microbiol 60, 373-395 12. Tullman-Ercek, D., DeLisa, M. P., Kawarasaki, Y., Iranpour, P., Ribnicky, B., Palmer, T., and Georgiou, G. (2007) J Biol Chem 282, 8309-8316 13. Manting, E. H., and Driessen, A. J. (2000) Mol Microbiol 37, 226-238 14. Van den Berg, B., Clemons, W. M., Jr., Collinson, I., Modis, Y., Hartmann, E., Harrison, S. C., and Rapoport, T. A. (2004) Nature 427, 36-44 15. Mitra, K., Schaffitzel, C., Shaikh, T., Tama, F., Jenni, S., Brooks, C. L., 3rd, Ban, N., and Frank, J. (2005) Nature 438, 318-324 16. Mitra, K., Frank, J., and Driessen, A. (2006) Nat Struct Mol Biol 13, 957-964 17. Osborne, A. R., and Rapoport, T. A. (2007) Cell 129, 97-110 18. Nouwen, N., and Driessen, A. J. (2002) Mol Microbiol 44, 1397-1405 19. Xie, K., Kiefer, D., Nagler, G., Dalbey, R. E., and Kuhn, A. (2006) Biochemistry 45, 13401-13408 20. Kiefer, D., and Kuhn, A. (2007) Int Rev Cytol 259, 113-138 21. Facey, S. J., and Kuhn, A. (2004) Biochim Biophys Acta 1694, 55-66 22. Akiyama, Y., Kihara, A., Tokuda, H., and Ito, K. (1996) J Biol Chem 271, 31196- 31201 23. Arsene, F., Tomoyasu, T., and Bukau, B. (2000) Int J Food Microbiol 55, 3-9 24. Traxler, B., and Murphy, C. (1996) J Biol Chem 271, 12394-12400 25. Froderberg, L., Rohl, T., van Wijk, K. J., and de Gier, J. W. (2001) FEBS Lett 498, 52-56

22 26. Schagger, H., and von Jagow, G. (1991) Anal Biochem 199, 223-231 27. Froderberg, L., Houben, E., Samuelson, J. C., Chen, M. Y., Park, S. K., Phillips, G. J., Dalbey, R., Luirink, J., and de Gier, J. W. L. (2003) Mol Microbiol 47, 1015-1027 28. Hewitt, C. J., and Nebe-Von-Caron, G. (2004) Adv Biochem Eng Biotechnol 89, 197-223 29. Fishov, I., and Woldringh, C. L. (1999) Mol Microbiol 32, 1166-1172 30. Tomoyasu, T., Mogk, A., Langen, H., Goloubinoff, P., and Bukau, B. (2001) Mol Microbiol 40, 397-413 31. Peltier, J. B., Ytterberg, A. J., Sun, Q., and van Wijk, K. J. (2004) J Biol Chem 279, 49367-49383 32. Wagner, S., Baars, L., Ytterberg, A. J., Klussmeier, A., Wagner, C. S., Nord, O., Nygren, P. A., van Wijk, K. J., and de Gier, J. W. (2007) Mol Cell Proteomics 33. Oakley, B. R., Kirsch, D. R., and Morris, N. R. (1980) Anal Biochem 105, 361- 363 34. Neuhoff, V., Arold, N., Taube, D., and Ehrhardt, W. (1988) Electrophoresis 9, 255-262 35. Molloy, M. P., Herbert, B. R., Slade, M. B., Rabilloud, T., Nouwens, A. S., Williams, K. L., and Gooley, A. A. (2000) Eur J Biochem 267, 2871-2881. 36. Berven, F. S., Karlsen, O. A., Murrell, J. C., and Jensen, H. B. (2003) Electrophoresis 24, 757-761 37. Zabrouskov, V., Han, X., Welker, E., Zhai, H., Lin, C., van Wijk, K. J., Scheraga, H. A., and McLafferty, F. W. (2006) Biochemistry 45, 987-992 38. Laskowska, E., Bohdanowicz, J., Kuczynska-Wisnik, D., Matuszewska, E., Kedzierska, S., and Taylor, A. (2004) Microbiology 150, 247-259 39. Marani, P., Wagner, S., Baars, L., Genevaux, P., de Gier, J. W., Nilsson, I., Casadio, R., and von Heijne, G. (2006) Protein Sci 15, 884-889 40. Peltier, J. B., Emanuelsson, O., Kalume, D. E., Ytterberg, J., Friso, G., Rudella, A., Liberles, D. A., Soderberg, L., Roepstorff, P., von Heijne, G., and van Wijk, K. J. (2002) Plant Cell 14, 211-236 41. Traxler, B., and Murphy, C. (1996) J Biol Chem 271, 12394-12400 42. Bessonneau, P., Besson, V., Collinson, I., and Duong, F. (2002) Embo J 21, 995- 1003 43. Nakatogawa, H., Murakami, A., and Ito, K. (2004) Curr Opin Microbiol 7, 145- 150 44. Dalbey, R. E., and Chen, M. (2004) Biochim Biophys Acta 1694, 37-53 45. Yi, L., Celebi, N., Chen, M., and Dalbey, R. E. (2004) J Biol Chem 279, 39260- 39267 46. van Bloois, E., Jan Haan, G., de Gier, J. W., Oudega, B., and Luirink, J. (2004) FEBS Lett 576, 97-100 47. van der Laan, M., Bechtluft, P., Kol, S., Nouwen, N., and Driessen, A. J. (2004) J Cell Biol 165, 213-222 48. Herskovits, A. A., Shimoni, E., Minsky, A., and Bibi, E. (2002) Journal of Cell Biology 159, 403-410 49. Ruiz, N., and Silhavy, T. J. (2005) Curr Opin Microbiol 8, 122-126

23 50. Mogk, A., Schlieker, C., Friedrich, K. L., Schonfeld, H. J., Vierling, E., and Bukau, B. (2003) J Biol Chem 278, 31033-31042 51. Darwin, A. J. (2005) Mol Microbiol 57, 621-628 52. Gittins, J. R., Phoenix, D. A., and Pratt, J. M. (1994) FEMS Microbiol Rev 13, 1- 12 53. Wu, C. C., and Yates, J. R. (2003) Nature Biotechnology 21, 262-267 54. Ito, K., and Akiyama, Y. (2005) Annu Rev Microbiol 59, 211-231 55. van Bloois, E., Haan, G. J., de Gier, J. W., Oudega, B., and Luirink, J. (2006) J Biol Chem 281, 10002-10009 56. du Plessis, D. J., Nouwen, N., and Driessen, A. J. (2006) J Biol Chem 281, 12248- 12252 57. Celebi, N., Yi, L., Facey, S. J., Kuhn, A., and Dalbey, R. E. (2006) J Mol Biol 357, 1428-1436 58. Hessa, T., Kim, H., Bihlmaier, K., Lundin, C., Boekel, J., Andersson, H., Nilsson, I., White, S. H., and von Heijne, G. (2005) Nature 433, 377-381 59. Andersson, H., and von Heijne, G. (1993) Embo J 12, 683-691 60. Fernandez-Escamilla, A. M., Rousseau, F., Schymkowitz, J., and Serrano, L. (2004) Nat Biotechnol 22, 1302-1306 61. Carrio, M. M., and Villaverde, A. (2003) FEBS Lett 537, 215-221 62. Carrio, M. M., and Villaverde, A. (2005) J Bacteriol 187, 3599-3601 63. Weibezahn, J., Tessarz, P., Schlieker, C., Zahn, R., Maglica, Z., Lee, S., Zentgraf, H., Weber-Ban, E. U., Dougan, D. A., Tsai, F. T., Mogk, A., and Bukau, B. (2004) Cell 119, 653-665 64. Withey, J. H., and Friedman, D. I. (2003) Annu Rev Microbiol 57, 101-123 65. Choy, J. S., Aung, L. L., and Karzai, A. W. (2007) J Bacteriol 66. Ruiz, N., Kahne, D., and Silhavy, T. J. (2006) Nat Rev Microbiol 4, 57-66 67. Kang, S. W., Rane, N. S., Kim, S. J., Garrison, J. L., Taunton, J., and Hegde, R. S. (2006) Cell 127, 999-1013 68. Qi, H. Y., Hyndman, J. B., and Bernstein, H. D. (2002) Journal of Biological Chemistry 277, 51077-51083 69. Cristobal, S., Scotti, P., Luirink, J., von Heijne, G., and de Gier, J. W. L. (1999) J Biol Chem 274, 20068-20070 70. Alami, M., Dalal, K., Lelj-Garolla, B., Sligar, S. G., and Duong, F. (2007) Embo J 26, 1995-2004 71. Chen, Y., Tai, P. C., and Sui, S. F. (2007) J Struct Biol 72. Facey, S. J., and Kuhn, A. (2003) European Journal of Biochemistry 270, 1724- 1734

24 FIGURE LEGENDS

Figure 1. Effect of SecE depletion on growth, steady state levels of Sec-components, model inner membrane proteins, and OmpA translocation. A. Effect of SecE depletion on cell growth. Growth of CM124 cultured in the presence (control) and

absence (SecE depletion) of 0.2% arabinose was monitored by measuring the OD600. B. Quantification of the steady-state levels of SecE, SecY, SecG, SecA, SecD, SecF, YidC as well as the model substrates FtsQ, Lep, Fob, and Foc in the inner membrane of SecE depleted and control cells. Inner membranes from SecE depleted and control cells were subjected to SDS-PAGE followed by immunoblot analysis with antibodies to the components listed above. The bar diagram indicates the average fold-change of the intensities of the bands upon SecE depletion compared to the control. The quantification is based on three independent samples. C. Analysis of the integrity and abundance of the SecYEG-complex. Inner membranes from SecE depleted and control cells were subjected to Blue-Native PAGE analysis followed by detection of SecE, SecY, and SecG by immunoblotting. The SecYEG trimer is indicated by < and the putative SecYG complex is indicated by <<. D. Effect of the depletion of SecE on the translocation of the major outer membrane protein OmpA. SecE depleted and control cells were labeled with [35S]methionine for 30 seconds and after adding cold methionine, chased for 3 and 10 minutes. OmpA was immunoprecipitated, subjected to standard SDS-PAGE analysis and labeled material was detected by phosphorimaging. The bars in the diagram indicate the percentage of the precursor and mature form of OmpA detected in the SecE depleted cells as compared to the mature OmpA detected in the control cells.

Figure 2. Flow cytometric properties of control and SecE depleted cells. SecE depleted and control cells were analysed by flow cytometry. A. Size of the population (forward scatter, FSC) plotted versus granularity (side scatter, SSC) for SecE depleted and control cells. Insets show microscopy pictures of a representative cell for the SecE depleted and control cultures. Cell length is indicated with scale bars. B. Histograms representing the fluorescence of cultures stained with the membrane-specific fluorophore FM4-64.

25 Figure 3. Analysis of whole cell lysates of SecE depleted and control cells by 2DE and immunoblotting. A. Comparative 2DE analysis of total lysates from SecE depleted

and control cells. Proteins from 1 OD600 unit of solubilized cells were separated by 2DE. Proteins were visualized by silver staining and differences between SecE depleted and control cells were analyzed using PDQuest. 28 spots were significantly (P<0.05) affected by SecE depletion. Proteins were identified by MALDI-TOF MS PMF from spots excised from gels stained with Coomassie (Table 1, Supplemental table 1). If several proteins were identified in the same spot, the first gene name listed corresponds to the one with the highest Mascot MOWSE score. Primary gene names were taken from SwissProt (www.expasy.org). Annotated spots were matched onto the silver stained gels shown using PDQuest. B. Bar graph showing the fold-changes of spots that are significantly (P<0.05) affected by SecE depletion. Fold-changes were calculated as the average spot intensities in SecE depleted samples/average spot intensities in control samples. C. Quantification of the precursor (p) and mature (m) forms of secretory proteins in SecE depleted and control cells by immunoblotting. Whole cells were subjected to SDS-PAGE followed by immunoblot analysis with antibodies to two periplasmic proteins (DegP and Skp) and three outer membrane proteins (OmpA, OmpF, and PhoE). The bars in the diagram indicate the percentage of the precursor and mature form of the proteins detected in the SecE depleted cells as compared to the mature form detected in the control cells. The quantification is based on three independent samples. D. Quantification of the levels of IbpA/B, SecA, Ffh, SecB, and PspA in whole cells. SecE depleted and control cells were subjected to SDS-PAGE followed by immunoblot analysis with antibodies to the components listed above. The bar graph shows the fold- changes calculated as the average band intensities detected in SecE depleted cells/average band intensities detected in control cells. The quantification is based on three independent samples.

Figure 4. Characterization of aggregates isolated from SecE depleted cells. Aggregates isolated from SecE depleted and control cells were analyzed by SDS-PAGE. Proteins were stained with colloidal Coomassie, and subsequently identified by MALDI- TOF MS PMF (Table 2, Supplemental table 2). If several proteins were identified in the same band, the first gene name listed corresponds to the protein with the highest Mascot MOWSE score.

26 Figure 5. 2DE analysis of the outer membrane proteome from SecE depleted and control cells. Cells were labeled with [35S]methionine for 1 minute followed by a chase of 10 minutes with cold methionine. The outer membrane fractions were isolated by density centrifugation from a mixture of labeled and non-labeled cells as out-lined in Supplemental figure 1 (see ‘Experimental procedures’ for details). The outer membrane fractions were used for separation by 2DE. Proteins were identified by MALDI-TOF MS PMF from spots excised from with Coomassie stained gels (Table 3, Supplemental table 3). The outer membrane fraction of SecE depleted cells was contaminated with aggregates that co-sediment with the outer membrane during density centrifugation (Supplemental figure 2). The spots corresponding to the proteins in these aggregates were identified and removed from the analysis set as described in the ‘Image analysis’ section of the ‘Experimental procedures’. Differences in the outer membrane proteomes of the SecE depleted and control cells were analyzed using PDQuest. Significantly affected (P<0.05) proteins are indicated by bold numbers in Table 3 and Supplemental table 3. A. Representative 2D gels showing proteins in the outer membrane fraction stained with colloidal Coomassie (protein steady-state levels) B. Representative 2D gels showing proteins in the outer membrane fraction detected by phosphorimaging (protein insertion). C. Bar graph showing the fold-changes (average spot intensity from SecE depleted samples/average spot intensity of control sample) of proteins visualized by Coomassie (black) and phosphorimaging (grey). A fold-change of 100 indicates that a spot was only detected in SecE depleted samples, a fold-change of 0.01 indicates that a spot was only detected in the control samples. Numbers refer to spot positions on the gels in Figure 5A and B.

Figure 6. 2D BN-PAGE analysis of the inner membrane proteome from SecE depleted and control cells. Cells were labeled with [35S]methionine for 1 minute followed by a chase of 10 minutes with cold methionine. The inner membrane fractions were isolated by density centrifugation from a mixture of labeled and non-labeled cells as outlined in Supplemental figure 1 (see ‘Experimental procedures’ for details). The inner membrane fractions were analysed by 2D BN-PAGE. Proteins were identified by MALDI-TOF MS PMF (Table 4, Supplemental table 4) from spots excised from Coomassie stained gels. If several proteins were identified in one spot, the first gene name listed corresponds to the protein with the highest Mascot MOWSE score. Primary gene names were taken from SwissProt (www.expasy.org). Differences in the inner membrane

27 proteomes of SecE depleted and control cells were analyzed using PDQuest. Significantly affected (P<0.05) proteins are indicated by bold numbers in Table 4 and Supplemental table 4. A. Representative 2D BN-PAGE gels with proteins detected by staining with colloidal Coomassie (protein steady-state levels). B. Representative 2D BN-PAGE gels with proteins detected by phosphorimaging (protein insertion). C. Bar graph showing the fold changes (average spot intensities from SecE depleted samples/average spot intensities of control samples) of proteins detected by Coomassie staining (black) and phosphorimaging (grey). A fold-change of 100 indicates a spot that was only detected in SecE depleted samples, a fold-change of 0.01 indicates that it was only detected in the control samples. Numbers refer to spot positions on the gels in Figure 6A and B.

Figure 7. Correlations between properties of inner membrane proteins and the effect on their steady-state levels and insertion upon SecE depletion. A. Fold-changes of steady-state levels (Coomassie) and insertion (phosporimaging) plotted against the number of amino acids in the largest translocated domain of each protein (Supplemental table 5). Almost all proteins that were positively affected by SecE depletion do not contain any large periplasmic domains. Exceptions to this trend are indicated in the plots with their gene names and number of predicted transmembrane segments. B. Proteins were divided into two groups; proteins with large translocated domains (≥60 amino acids) and proteins with small translocated domains (≤60 amino acids). The fold-changes upon SecE depletion on these two groups were plotted against the number of transmembrane segments (Figure 7B, Supplemental table 5).

28 Table 1. Comparative 2D-gel analysis and mass spectrometry identification of proteins from total lysates of SecE depleted and control cells.

Spots visualized by silver staining (Figure 3A) were quantified and compared using PDQuest. Spot quantities were normalized using the 'total quantity of valid spot' method. Values of 0.01 and 100 correspond to spots that are missing or turned on in the SecE depletion, respectively. All spots that were significantly (P<0.05) changed upon SecE depletion were excised from gels stained with colloidal Coomassie and proteins were identified by MALDI-TOF MS PMF.

predicted predicted pI fold change MW (kDa) observed spot Nr. gene name protein name local. (precursor observed pI (SecE depl. (precursor MW (kDa) /mature) /control) /mature) (a) (b) (c) (d) (e) (f) (g) (h) (i)

1 ybbN Protein ybbN c 31.8 4.5 31.8 4.5 2.23 2 groL 60 kDa chaperonin c 57.3 4.85 55.3 4.68 100 3 dnaK Chaperone protein dnaK c 69.1 4.83 68.9 4.83 1.74 4 groL 60 kDa chaperonin c 57.3 4.85 60.2 4.83 1.96 5 potD Spermidine/putrescine-binding periplasmic protein p 38.9/36.5 5.24/4.86 35.87 4.76 0.01 6 metQ D-methionine-binding lipoprotein metQ im/om lp 38.9/36.5 5.24/4.86 27.25 4.73 2.96 6 rpsB 30S ribosomal protein S2 c 26.7 6.61 27.25 4.73 2.96 6 grpE Protein grpE c 21.8 4.68 27.25 4.73 2.96 7 metQ D-methionine-binding lipoprotein metQ im/om lp 38.9/36.5 5.24/4.86 27.4 4.88 4.32 8 no id. 27.98 5.11 100 9 fliY Cystine-binding periplasmic protein p 29.3/26.1 6.22/5.29 25.58 5.18 0.13 10 livJ Leu/Ile/Val-binding protein p 39.1/36.8 5.54/5.28 38.7 5.22 0.20 11 hdhA 7-alpha-hydroxysteroid dehydrogenase c 26.8 5.22 23.39 5.21 0.37 12 clpB Chaperone clpB c 95.6 5.37 79.12 5.29 2.96 13 ushA Protein ushA p 60.8/58.2 5.47/5.4 59.65 5.31 0.20 14 cysK Cysteine synthase A amb 34.5 5.83 35.47 5.34 0.01 14 ompT Protease 7 om 35.6/33.5 5.76/5.38 35.47 5.34 0.01 14 trxB Thioredoxin reductase c 34.6 5.3 35.47 5.34 0.01 15 no id. 26.48 5.35 7.68 16 bla Beta-lactamase TEM p 31.5/28.9 5.69/5.46 29.8 5.39 0.46 17 no id. 24.4 5.39 12.90 18 znuA High-affinity zinc uptake system protein znuA p 33.8/31.1 5.61/5.44 33.29 5.52 0.07 19 yehZ Hypothetical protein yehZ p 32.6/30.2 5.82/5.56 31.19 5.48 0.01 20 ybiS Protein ybiS p 33.42/30.86 5.99/5.6 31.19 5.57 0.34 20 yggE Hypothetical protein yggE p 26.6/24.5 6.1/5.60 31.19 5.57 0.34 21 yodA Metal-binding protein yodA p 24.8/22.3 5.91/5.66 23.15 5.6 0.01 22 dppA Periplasmic dipeptide transport protein p 60.3/57.4 6.21/5.75 55.21 5.74 0.36 23 ompC Outer membrane protein C om 40.4/38.3 4.58/4.48 32.71 5.76 100 24 oppA Periplasmic oligopeptide-binding protein p 60.97/58.5 6.05/5.85 54.94 6.02 0.13 25 tolB Protein tolB p 46.0/43.6 6.98/6.14 41.31 6.09 0.23 26 ompA Outer membrane protein A om 37.2/35.2 5.99/5.60 37.3 6.07 100 27 rbsB D-ribose-binding periplasmic protein p 31.0/28.5 6.85/5.99 30.67 6.03 100 28 rbsB D-ribose-binding periplasmic protein p 31.0/28.5 6.85/5.99 28.46 6.02 0.17

(a) The numbering corresponds to the spots in the 2D-gel images in Figure 3A. (b) Gene name from the Swiss Prot database for E. coli . 'no id.' indicates that no protein was identified in the spot. (c) Protein name from the Swiss Prot database for E. coli . 'no id.' indicates that no protein was identified in the spot. (d) Localization based on the information given in the Swiss Prot for E. coli . Unknown localizations were predicted by PSORT. For integral membrane proteins, the number of transmembrane segments are indicated. Abbreviations: amb., ambiguous localization; c., cytoplasmic; im lp., inner membrane lipoprotein; local., localization; om., outer membrane; om lp., outer membrane lipoprotein; p., periplasmic. (e) Protein sizes (in kDa) predicted from amino acid sequences. Two sizes are given for secretory proteins, the first size corresponds to the precursor form, and the second size corresponds to the mature form of the protein. (f) pI predicted from amino acid sequence. Two values are given for secretory proteins, the first value corresponds to the precursor form, and the second value corresponds to the mature form of the protein. (g) Size of proteins calculated from the spot position on the 2D-gels used for the analysis. (h) pI of proteins calculated from the spot position on the 2D-gels used for the analysis. (i) Fold-change, i.e., the ratio of the average intensity of significantly (P<0.05) affected spots in the gels of the SecE depletion to the average intensity of matched spots in the control gels.

29 Table 2. Mass spectrometry identification of proteins in aggregates isolated from lysates of SecE depleted and control cells. Protein aggregates were extracted from lysates of SecE depleted and control cells. The protein content of the aggregates was analysed by 1DE followed by MALDI-TOF MS PMF (Figure 4, Supplemental table 2) or nanoLC-MS/MS analysis of solubilized aggregates digested with trypsin (Supplemental table 2).

predicted MW band nr. hit rank gene name protein name local./ TMs (kDa) (precursor /mature)

(a) (b) (c) (d) (e) (f)

2 adhE Aldehyde-alcohol dehydrogenase c 96.1 17 atpA ATP synthase subunit alpha c, im ass 55.4 22, 30, 33 1 bla Beta-lactamase TEM p 31.6/28.9 6 clpB Chaperone clpB c 95.7 38 25 crp Catabolite gene activator c 23.6 28 cysK Cysteine synthase A amb 34.4 14 cysN Sulfate adenylyltransferase subunit 1 c 52.6 25 dacA Penicillin-binding protein 5 1 * 40.3/38.3 23 dacC Penicillin-binding protein 6 1 * 43.6/40.8 19 degP Protease do p 49.4/46.8 23 dnaJ Chaperone protein dnaJ c 41.1 9 dnaK Chaperone protein dnaK c 69.1 42 13 dps DNA protection during starvation protein c 18.6 14 elaB Protein elaB 1 11.3 4, 5 6 fusA Elongation factor G c 77.6 18 12 glgA Glycogen synthase c 53.0 7, 36 17 glgB 1,4-alpha-glucan branching enzyme c 84.3 11 glmS Glucosamine--fructose-6-phosphate aminotransferase c 66.9 16 guaB Inosine-5'-monophosphate dehydrogenase c 52.0 11, 36 htpG Chaperone protein htpG c 66.0 21 hupA DNA-binding protein HU-alpha c 9.5 46, 47 16 ibpA Small heat shock protein ibpA c 15.8 23 iscS Cysteine desulfurase c 45.2 49 9 lpp Major outer membrane lipoprotein om lp 8.3/6.4 12 lysS Lysyl-tRNA synthetase c 57.6 5-methyltetrahydropteroyltriglutamate--homocysteine 4, 5 15 metE c 84.7 methyltransferase 26 mreB Rod shape-determining protein mreB c 37.1 29 nlpD Lipoprotein nlpD im lp 40.2/37.5 27, 33, 34 3 ompA Outer membrane protein A om 37.2/35.2 24, 25 5 ompC Outer membrane protein C om 40.3/38.3 23, 24 18 ompF Outer membrane protein F om 39.3/37.1 24 ompT Protease 7 om 35.5/33.5 41 8 ompX Outer membrane protein X om 18.7/16.2 13 ptsI Phosphoenolpyruvate-protein phosphotransferase c 63.6 35 rplC 50S ribosomal protein L3 c 22.2 37 rplD 50S ribosomal protein L4 c 22.1 40 24 rplE 50S ribosomal protein L5 c 20.5 39 rplF 50S ribosomal protein L6 c 18.9 44 rplI 50S ribosomal protein L9 c 15.8 1 rpoC DNA-directed RNA polymerase beta' chain c 155.1 10, 31 rpsA 30S ribosomal protein S1 c 61.2 32 rpsB 30S ribosomal protein S2 c 26.7 19, 21 rpsC 30S ribosomal protein S3 c 25.8 43 rpsE 30S ribosomal protein S5 c 17.5 40 rpsG 30S ribosomal protein S7 c 20.0 30 48 26 rpsJ 30S ribosomal protein S10 c 11.7 3 secA Preprotein translocase subunit secA c 102.0 43, 44 10 skp Chaperone protein skp p 17.7/15.7 45 4 slyB Outer membrane lipoprotein slyB om lp 15.6/13.8 25 spb Sulfate-binding protein p 18.7/16.2 20 7 tolB Protein tolB p 36.6/34.7 15 tolC Outer membrane protein tolC om 54.0/51.5 2 tufA Elongation factor Tu c 43.3 8 typA GTP-binding protein typA/BipA c 67.4 20 yajG Uncharacterized lipoprotein yajG om lp 21.0/19.0 35 yfiO Lipoprotein yfiO om lp 23.9/21.7 11 ygiW Protein ygiW p 14.0/12.0 27 yhjK Protein yhjK 2 73.1 25 28 yncE Hypothetical protein yncE sec 38.6/35.3 22 yqjD Hypothetical protein yqjD 1 11.1 35 yrbC Protein yrbC sec 23.9/21.7

(a) The numbering corresponds to the bands in the 1D-gel shown in Figure 4. (b) The ranking is based on the Mascot MOWSE score of proteins identified by in-solution digestion/ nanoLC-ESI-MS/MS. (c) Protein name extracted from the Swiss Prot database for E. coli . (d) Gene name extracted from the Swiss Prot database for E. coli . (e) Localization based on the information given in the Swiss Prot database for E. coli . Unknown localizations were predicted by PSORT. For integral membrane proteins, the number of transmembrane segments are indicated. '*' indicates that the membrane inserted segment may work as an anchor rather than a true transmembrane segment. Abbreviations: amb., ambiguous localization; c., cytoplasmic; im ass., inner membrane associated; im lp., inner membrane lipoprotein; om., outer membrane; om lp., outer membrane lipoprotein; p., periplasmic; sec., secretory; TMs., trans-membrane segments. (f) Protein sizes (in kDa) predicted from amino acid sequence. Two sizes are given for secretory proteins, the first size corresponds to the precursor form, and the second size corresponds to the mature form of the protein.

31 Table 3. Comparative 2D-gel analysis and mass spectrometry identification of proteins in the outer membrane fraction of SecE depleted and control cells. The spots in 2D-gels of the outer membrane fraction from SecE depleted and control cells were excised from gels stained with colloidal Coomassie (Figure 5A). Proteins were identified by MALDI-TOF MS PMF. Spots visualized by Coomassie staining and phosphorimaging (Figure 5A and B) were quantified and compared using PDQuest. Quantities of Coomassie stained spots were normalized using the 'total quantity of valid spot' tool, excluding spots detected in the 2D gels of protein aggregates extracted from the outer membrane fraction (Supplemental figure 1). Quantities of spots visualized by phosphorimaging were normalized using the normalization factor calculated for the corresponding Coomassie stained gel. Values of 0.01 and 100 correspond to spots that are missing or turned on in the SecE depletion, respectively. Significant fold-changes (P<0.05) are indicated with bold numbers. Graphs of fold-changes are shown in Figure 5C.

coomassie phosphor. predicted fold fold predicted pI observed gene local. MW (kDa) observed change change spot nr. protein name (precursor MW name /TMs (precursor pI (SecE (SecE /mature) (kDa) /mature) depl. depl. /control) /control) (a) (b) (c) (d) (e) (f) (g) (h) (i) (i) 1 mdtE Multidrug resistance protein mdtE im lp 41.3/38.9 5.73/5.12 53.22 4.14 1.06 n.d. 2 dcrB Protein dcrB amb 19.8/17.8 5.09/4.91 17.7 4.29 0.22 0.70 3 yfgL Lipoprotein YfgL om lp 41.9/39.9 4.72/4.61 37.75 4.35 0.83 0.76 4 fhuE FhuE receptor om 81.2/77.4 4.75/4.72 77.17 4.65 0.25 0.56 5 hemX Putative uroporphyrinogen-III C-methyltransferase 1 42.9 4.68 45.08 4.59 1.93 1.93 7 imp LPS-assembly protein om 89.7/87.1 4.94/4.85 91.39 4.87 0.50 0.01 8 yaeT Outer membrane protein assembly factor yaeT om 90.6/88.4 4.93/4.87 88.4 4.87 0.33 0.62 12 fadL Long-chain fatty acid transport protein om 48.5/45.9 5.09/4.91 44.86 4.84 1.03 1.17 14 nlpB Lipoprotein 34 om lp 36.9/34.4 5.34/4.96 34.68 4.74 0.52 1.01 15 metQ D-methionine-binding lipoprotein metQ om/im lp 29.5/27.2 5.13/4.93 26.94 4.76 0.07 n.d. 16 tsx Nucleoside-specific channel-forming protein tsx om 33.6/31.4 5.07/4.87 27.47 4.86 0.61 1.25 17 ybaY Hypothetical lipoprotein ybaY om/im lp 19.5/17.7 7.88/6.31 23.21 4.78 0.44 n.d. 19 mdtE Multidrug resistance protein mdtE im lp 41.2/38.9 5.73/5.12 40.1 4.97 0.31 n.d. 20 tsx Nucleoside-specific channel-forming protein tsx om 33.6/31.4 5.07/4.87 27.54 4.96 0.47 0.50 21 mipA MltA-interacting protein om 27.8/25.7 5.50/5.03 25.74 4.95 0.26 0.64 23 ompX Outer membrane protein X om 18.6/16.4 6.56/5.3 16.12 4.91 0.49 1.13 25 cirA Colicin I receptor om 74.1/71.2 5.11/5.03 75.78 5.05 0.11 0.77 26 yiaF Hypothetical protein yiaF amb 30.43 9.35 23.57 5.01 0.21 0.91 27 fhuA Ferrichrome-iron receptor om 82.4/78.7 5.47/5.13 79.29 5.16 0.32 0.77 28 btuB Vitamin B12 transporter btuB om 68.4/66.3 5.23/5.10 66.13 5.13 0.48 1.02 30 nlpA Lipoprotein 28 im lp 29.4/27.1 5.77/5.29 26.64 5.12 0.23 0.01 31 mipA MltA-interacting protein om 27.8/25.7 5.50/5.03 25.52 5.11 0.49 1.02 33 pal Peptidoglycan-associated lipoprotein om lp 16.9/16.6 6.29/5.59 17.65 5.08 1.02 1.14 36 fepA Ferrienterobactin receptor om 82.1/79.8 5.39/5.23 81.97 5.23 0.07 0.43 37 yfiO Lipoprotein yfiO om lp 27.9/25.8 6.16/5.48 24.87 5.2 0.59 0.62 38 ompX Outer membrane protein X om 18.6/16.4 6.56/5.3 15.7 5.25 0.64 0.72 39 pal Peptidoglycan-associated lipoprotein om lp 16.9/16.6 6.29/5.59 17.43 5.38 1.19 1.12 44 yfiO Lipoprotein yfiO om lp 27.9/25.8 6.16/5.48 25.39 5.53 0.69 n.d. 45 ompW Outer membrane protein W om 22.9/20.9 6.03/5.58 21.5 5.54 0.31 0.10 48 fusA Elongation factor G c 77.6 5.24 83.22 5.69 0.09 n.d. 49 yicH Hypothetical protein yicH sec 62.3/58.7 5.67/5.38 63.8 5.85 0.15 n.d. 50 ompA Outer membrane protein A om 37.2/35.2 5.99/5.60 31.74 6 1.22 1.12 53 ompX Outer membrane protein X om 18.6/16.4 6.56/5.3 16.49 6.19 0.37 0.46

(a) The numbers corresponds to the spot numbers in the 2D-gels shown in Figure 5A and B. (b) Gene name extracted from the Swiss Prot database for E. coli . (c) Protein name from the Swiss Prot database for E. coli . 'no id.' indicates that no protein was identified in the spot. (d) Localization based on the information given in the Swiss Prot data base for E.coli . Unknown localizations were predicted by PSORT (http://psort.hgc.jp/form.html). Abbreviations: amb., ambiguous localization; c., cytoplasmic; im lp., inner membrane lipoprotein; local., localization; om., outer membrane; om lp., outer membrane lipoprotein; sec., secretory; TMs., transmembrane segments.

32 (e) Protein sizes (in kDa) predicted from amino acid sequences. Two sizes are given for secretory proteins, the first size corresponds to the precursor form including the signal sequence, and the second size corresponds to the mature form of the protein after signal sequence processing. (f) pI predicted from amino acid sequence. Two values are given for secretory proteins, the first value corresponds to the precursor form, and the second value corresponds to the mature form of the protein. (g) Size of proteins calculated from the spot position on the 2D-gels shown in Figure 5. (h) pI of proteins calculated from the spot position on the 2D-gels shown in Figure 5. (i) Fold-change, i.e., the ratio of the average spot intensity. Fold-change of significantly (P<0.05) affected spots are indicated by bold numbers. 'n.d.' indicates that the spot was not detected. Abbreviations: phosphor., phosphorimaging

33 Table 4. 2D Blue Native PAGE analysis of the inner membrane proteome of SecE depleted and control cells. Spots in 2D BN PAGE gels of inner membranes (Figure 6A) were excised from gels stained with colloidal Coomassie. Proteins were identified by MALDI-TOF MS PMF. Proteins belonging to the same complex have the same gray fill color. Spots visualized by Coomassie staining and/or phosphorimaging (Figure 6A and B) were quantified and compared using PDQuest. 0.01 and 100 correspond to spots that are missing or turned on in the SecE depletion, respectively. Significant fold- changes (P<0.05) are indicated with bold numbers.

predicted Coomassie phosphor. observed gene MW (kDa) observed fold change fold change spot nr. protein name local. /TMs native MW name (precursor MW (kDa) (SecE dep (SecE depl (kDa) /mature) /control) /control)

(a) (b) (c) (d) (e) (f) (g) (h) (h)

1 yjeP Hypothetical mscS family protein yjeP 10 124.0 103.7 1000 0.52 n.d.

2 kefA Potassium efflux system kefA 11 127.2 100.7 1000 0.23 n.d.

3 ftsH Cell division protease ftsH 2 70.7 69.3 1000 0.24 0.56

4 hflK Protein hflK 1 45.5 45.4 1000 0.31 n.d.

5 hflC Protein hflC 1 37.7 37.4 1000 0.43 n.d.

6 ybbK Hypothetical protein ybbK 1 33.7 32.2 1000 1.36 n.d.

6 corA Magnesium transport protein corA 2 36.6 32.2 1000 1.36 n.d.

7 nuoC NADH-quinone oxidoreductase subunit C/D c, im ass 68.7 67.1 966 0.70 n.d.

8 creD Inner membrane protein creD 6 49.8 42.7 733 0.20 n.d.

8 hemY Protein hemY 2 45.2 42.7 733 0.20 n.d.

9 groL 60 kDa chaperonin c 57.2 64.0 828 3.72 1.30

10 atpA ATP synthase subunit alpha c, im ass 55.2 55.4 638 1.36 0.91

11 atpD ATP synthase subunit beta c, im ass 50.2 53.0 637 1.50 1.44

12 atpG ATP synthase gamma chain c, im ass 31.6 31.4 614 1.19 1.49

15 no id. 22.5 609 2.02 0.93

13 atpH ATP synthase delta chain c, im ass 19.3 20.8 606 1.11 n.d.

14 atpF ATP synthase B chain 1 17.3 18.9 599 1.47 n.d.

16 fadE Acyl-coenzyme A dehydrogenase 2 89.2 85.1 600 1.28 n.d.

16 plsB Glycerol-3-phosphate acyltransferase c, im ass 91.3 85.1 600 1.28 n.d.

17 wzzE Lipopolysaccharide biosynthesis protein wzzE 2 39.6 38.6 603 0.90 0.72

17 ybdG Hypothetical protein ybdG 1 36.4 38.6 603 0.90 0.72

18 nuoC NADH-quinone oxidoreductase subunit C/D c, im ass 68.7 68.1 543 0.86 n.d.

19 nuoB NADH-quinone oxidoreductase subunit B c, im ass 25.1 25.5 510 0.59 n.d.

20 nuoI NADH-quinone oxidoreductase subunit I c, im ass 20.5 23.5 509 0.84 n.d.

21 wzzB Chain length determinant protein 2 36.5 35.1 531 0.55 0.07

22 narG Respiratory nitrate reductase 1 alpha chain c, im ass 140.4 113.1 463 1.92 n.d.

23 narI # Respiratory nitrate reductase 1 gamma chain 5 25.5 21.7 437 2.58 0.79

24 acrB lavine resistance protein B 12 113.6 92.7 478 0.23 0.01

25 sdhA Succinate dehydrogenase flavoprotein subunit c, im ass 64.4 68.6 440 0.92 n.d. Succinate dehydrogenase hydrophobic membrane anchor 26 sdhD 4 12.9 12.0 427 0.94 n.d. subunit 27 sdhB Succinate dehydrogenase iron-sulfur subunit c, im ass 26.8 27.2 438 1.44 1.14

27 manZ Mannose permease IID component 1 31.3 27.2 438 1.44 1.14

28 manX PTS system mannose-specific EIIAB component c, im ass 34.9 36.8 439 2.40 1.26

29 no id. 30.5 436 1.02 0.78

34 30 no id. 13.7 434 1.00 n.d

31 yrbD Hypothetical protein yrbD sec 19.6/16.5 87.5 426 0.36 n.d.

32 no id. 30.5 406 0.52 0.47

33 nuoC NADH-quinone oxidoreductase subunit C/D c, im ass 68.7 69.3 363 0.81 n.d.

34 atpA ATP synthase subunit alpha c, im ass 55.2 56.4 366 1.02 0.83

35 atpD ATP synthase subunit beta c, im ass 50.2 54.0 364 1.33 1.26

36 atpG ATP synthase gamma chain c, im ass 31.6 32.2 348 1.68 1.73

37 atpF ATP synthase B chain 1 17.3 19.5 301 1.27 n.d.

38 no id. 88.8 327 1.10 n.d.

39 mscS Small-conductance mechanosensitive channel 3 30.9 24.4 302 0.77 0.94

40 no id. 93.9 228 0.61 n.d.

41 aas AAS bifunctional protein 2 80.7 78.7 248 4.38 1.59

41 yhjG Hypothetical protein yhjG 2 75.1 78.7 248 4.38 1.59

42 ppiD Peptidyl-prolyl cis-trans isomerase D 1 68.2 71.6 264 0.22 0.01

43 msbA Lipid A export ATP-binding/permease protein msbA 5 64.5 59.7 238 0.64 0.68

44 hemX Putative uroporphyrinogen-III C-methyltransferase 1 43.0 49.0 265 0.84 0.83

45 ydjN Hypothetical symporter ydjN 9 48.7 35.1 272 0.98 0.69

46 exbB Biopolymer transport exbB protein 3 26.3 24.7 243 0.31 0.46

47 no id. 17.4 231 0.01 0.01

48 secA Preprotein translocase subunit secA c, im ass 102.0 108.1 209 50.77 100.00

49 cyoB Ubiquinol oxidase subunit 1 15 74.4 50.7 203 0.15 0.81

50 cyoA Ubiquinol oxidase subunit 2 2 34.9 32.6 201 0.25 0.79

51 yhbG Probable ABC transporter ATP-binding protein yhbG c, im ass 26.7 27.5 201 0.35 n.d.

52 ygiM Hypothetical protein ygiM 1 23.1 22.6 218 0.31 n.d.

52 tolQ Protein tolQ 3 25.6 22.6 218 0.31 n.d.

53 atpF ATP synthase B chain 1 17.3 20.1 200 0.11 0.88

54 ppiD Peptidyl-prolyl cis-trans isomerase D 1 68.2 71.4 176 0.51 0.59

55 no id. 49.4 159 100.00 1.84

56 no id. 49.2 160 0.50 1.70

57 no id. 44.3 159 0.84 0.83

58 dacC Penicillin-binding protein 6 1 43.6/40.8 42.6 158 0.87 0.81

58 mdtE Multidrug resistance protein mdtE im, lp 41.2 42.6 158 0.87 0.81

58 cysA Sulfate/thiosulfate import ATP-binding protein cysA c, im ass 41.1 42.6 158 0.87 0.81

59 cysA Sulfate/thiosulfate import ATP-binding protein cysA c, im ass 41.1 43.1 177 2.27 1.37

59 dacC Penicillin-binding protein 6 c, im ass 41.1 43.1 177 2.27 1.37

59 hemY Protein hemY 2 45.2 43.1 177 2.27 1.37 60 dacA Penicillin-binding protein 5 1 44.4/41.3 41.4 1.40 n.d. 60 dppF Dipeptide transport ATP-binding protein dppF c, im ass 37.6 41.4 160 1.40 n.d.

61 dppD Dipeptide transport ATP-binding protein dppD c, im ass 35.8 37.3 160 1.70 n.d.

62 no id. 36.0 193 4.71 n.d.

63 metQ D-methionine-binding lipoprotein metQ im/om lp 29.4/27.2 29.0 166 0.27 0.48

63 nlpA Lipoprotein 28 im/om lp 29.4/27.1 29.0 166 0.27 0.48

64 no id. 25.2 160 14.15 1.96

65 yfgM Protein yfgM 1 22.2 23.7 165 0.41 0.66

66 nuoG NADH-quinone oxidoreductase subunit F c, im ass 100.2 102.5 159 1.30 n.d.

67 nuoF NADH-quinone oxidoreductase subunit G c, im ass 49.3 51.1 150 0.95 n.d.

35 68 mgtA Magnesium-transporting ATPase, P-type 1 10 99.5 98.6 149 1.34 n.d.

69 copA Copper-transporting P-type ATPase 8 87.7 93.4 158 0.90 1.44

69 mrcA Penicillin-binding protein 1A 1 93.6 93.4 158 0.90 1.44

70 frdA Fumarate reductase flavoprotein subunit c, im ass 65.8 74.2 145 1.24 n.d.

71 secD Protein-export membrane protein secD 6 66.6 65.5 151 0.47 0.60

72 cydC Transport ATP-binding protein cydC 6 62.9 57.5 150 1.86 1.79

72 cydd Transport ATP-binding protein cydD 6 65.1 57.5 150 1.86 1.79

73 cydA Cytochrome d ubiquinol oxidase subunit 1 7 58.2 48.1 137 0.19 n.d.

74 metN Methionine import ATP-binding protein metN c 37.8 40.4 147 1.41 n.d.

75 degS Protease degS p 37.6/34.6 37.7 151 0.45 n.d.

76 metQ D-methionine-binding lipoprotein metQ im/om lp 29.4/27.2 29.1 135 0.61 0.67

77 rnfG Electron transport complex protein rnfG 1 21.9 22.6 143 0.10 0.01

78 no id. 21.4 135 1.03 1.14

79 atpF ATP synthase B chain 1 17.3 20.1 149 2.07 1.25

80 glnP Glutamine transport system permease protein glnP 5 24.4 19.0 128 0.83 0.92

81 yhcB Putative cytochrome d ubiquinol oxidase subunit III 1 15.2 14.5 144 2.41 1.52

82 yhcB Putative cytochrome d ubiquinol oxidase subunit III 1 15.2 14.6 133 1.05 0.90

83 ydiJ Hypothetical protein ydiJ amb 113.2 103.2 115 2.27 n.d.

84 plsB Glycerol-3-phosphate acyltransferase c, im ass 91.3 94.4 116 0.91 1.31

85 gcd Quinoprotein glucose dehydrogenase 5 86.7 89.7 117 0.41 0.66

86 ppiD Peptidyl-prolyl cis-trans isomerase D 1 68.2 73.1 126 0.68 0.99

87 nuoC NADH-quinone oxidoreductase subunit C/D c, im ass 68.7 72.2 119 0.96 n.d.

88 oxaA Inner membrane protein oxaA 6 61.5 60.6 125 0.31 0.42

89 ybhG Membrane protein ybhG 1 36.4/34.4 37.9 130 0.56 1.17

90 nuoB NADH-quinone oxidoreductase subunit B c, im ass 25.1 26.5 119 1.17 n.d.

91 yajC Membrane protein yajC 1 11.9 11.7 111 0.65 0.51

92 secD Protein-export membrane protein secD 6 66.6 67.5 106 0.21 0.33

92 yicH Hypothetical protein yicH amb 62.3 67.5 106 0.21 0.33

93 no id. 52.8 115 n.d 1.42

94 dadA D-amino acid dehydrogenase small subunit c, im ass 47.6 52.2 105 1.04 1.07

94 zipA Cell division protein zipA 1 36.5 52.2 105 1.04 1.07

95 ndh NADH dehydrogenase amb 47.2 49.3 111 1.63 1.36

96 proP Proline/betaine transporter 12 54.8 41.7 111 0.62 0.85

97 metQ D-methionine-binding lipoprotein metQ im/om lp 29.4/27.2 29.1 110 0.96 0.94

98 no id. 96.4 92 0.62 n.d.

99 yijP Membrane protein yijP 5 66.6 62.9 96 0.45 n.d.

100 ydgA Protein ydgA p, im ass 54.7/52.8 59.3 96 0.42 0.66

101 emrA Multidrug resistance protein A 1 42.7 45.5 89 0.86 0.72

102 dacA Penicillin-binding protein 5 1 44.4/41.3 43.6 92 0.53 0.49

102 dacC Penicillin-binding protein 6 1 43.6/40.8 43.6 92 0.53 0.49

103 pyrD Dihydroorotate dehydrogenase c, im ass 36.8 39.4 83 0.37 0.10

104 gadC Probable glutamate/gamma-aminobutyrate antiporter 12 55.1 38.3 98 0.27 0.70

105 glpT Glycerol-3-phosphate transporter 12 47.2 50.3 87 0.54 0.42

106 ompA Outer membrane protein A om 37.2 31.8 87 0.93 0.87

106 metQ D-methionine-binding lipoprotein metQ im/om lp 29.4/27.2 31.8 87 0.93 0.87

36 CDP-diacylglycerol--glycerol-3-phosphate 3- 107 pgsA 4 20.6 17.2 86 1.83 1.73 phosphatidyltransferase 108 dnaK Chaperone protein dnaK c 69.0 83.7 77 5.89 1.72

109 dld D-lactate dehydrogenase c, im ass 64.5 73.3 71 0.72 1.15

110 cysK Cysteine synthase A c 34.4 38.1 63 1.10 0.86

111 lepB Signal peptidase I 2 36.0 35.2 69 0.50 0.95

112 metQ D-methionine-binding lipoprotein metQ im/om lp 29.4/27.2 29.5 70 0.39 n.d.

112 rpsB 30S ribosomal protein S2 c 26.6 29.5 70 0.39 n.d.

113 no id. 27.2 57 12.16 1.73

114 no id. 23.0 63 2.34 0.96

115 no id. 21.5 58 0.30 0.42

116 no id. 19.1 72 0.61 n.d.

117 yajC Membrane protein yajC 1 11.9 11.7 71 1.44 1.43

(a) The numbering corresponds to the spots in the 2D BN PAGE gels shown in Figure 6A. (b) Gene name from the Swiss Prot database for E. coli . '#' indicates that the criteria for MS identification was not fulfilled, but annotation fits with expected size of complex and monomer. (c) Protein name from the Swiss Prot database for E. coli . (d) Localization based on the information given in the Swiss Prot database for E. coli. Unknown localizations were predicted by PSORT. For integral membrane proteins, the number of transmembrane segments is indicated. '*' indicates that the membrane inserted segment may work as an anchor rather than a true transmembrane segment. Abbreviations: amb., ambiguous localization; c., cytoplasmic; im ass., inner membrane associated; im lp., inner membrane lipoprotein; local., localization; om., outer membrane; om lp., outer membrane lipoprotein; sec., secretory; TMs., transmembrane segments. (e) Protein sizes (in kDa) predicted from amino acid sequences. Two sizes are given for secretory proteins, the first size corresponds to the precursor form, and the second size corresponds to the mature form of the protein. (f) Size of protein calculated from the spot position on the 2D BN PAGE gels used for the analysis. (g) Native mass (in kDa) based on the position in the 2D BN PAGE gel in Figure 6A. (h) The fold-change, i.e., the ratio of the average intensity of spots in gels of the SecE depleted cells to the average intensity of matched spots in the control gels. 'n.d.' indicates that the spot was not detected. Abbreviations: phosphor., phosphorimaging.

37 Figure 1

A. C. α SecE α SecY α SecG Native MW 2.5 control (kDa) 2 SecE depletion 160 1.5 600

OD 1 < 0.5

0 02468101214 time (h) << B. 66

control SecE depl control SecE depl control SecE depl α SecE

α SecY D. α SecG 200 precursor α SecA mature 150 α SecD 100 α SecF

α YidC % of control 50

α FtsQ 0 0310chase (min) α Lep

α Fo b pre-OmpA OmpA α Fo c control Sec 0 0.5 1.0 1.5 2.0 control SecE depl control SecE depl control SecE depl E depl fold-change (SecE depl/control)

38 Figure 2

A. B. control SecE depl 4 4 10 10 2.13 μm 100 control 3 3 SecE depl 10 10 2.86 μm 80

60 2 2 10 10 40 Side Scatter Side Scatter 1 1 10 10 % of Maximum 20

0 0 10 10 0 0 1 2 3 4 0 1 2 3 4 0 1 2 3 4 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 Forward Scatter Forward Scatter FM4-64

39 Figure 3

A. B.

control Mw (kDa) SecE depletion Mw (kDa) 12 100 12 100 2 groL 3 75 3 75 4 13 4 13 8 no id. 22 24 22 24 2 23 ompC 50 50 26 ompA 25 25 10 26 26 27 rbsB 5 14 37 37 17 no id. 18 18 23 1 19 20 27 1 20 27 16 16 15 no id. 6 7 28 6 7 8 28 9 9 15 17 17 7 metQ 11 21 25 11 21 25 6 metQ, rpsB, grpE 20 20 12 clpB 15 15 1 ybbN 4 groL 3 dnaK 10 10 16 bla 11 hdhA pI 4.6 5.0 5.4 6.0 pI 4.6 5.0 5.4 6.0 22 dppA 20 ybiS, yggE 25 tolB C. D. 13 ushA 10 livJ p precursor αDegP 28 rbsB m mature αIbpB 9 fliY p αPhoE m αSecA 24 oppA 18 znuA p αSkp m αSecB 5 potD p 14 cysK, ompT, trxB αOmpA m αFfh 19 yehZ 21 yodA αOmpF p m αPspA control SecE depl 0 50 100 150 200 250 control SecE depl 0.1 1 10 0.0010.01 0.11 10100 1000 % of control fold-change fold-change (SecE depl/control) (SecE depl/control)

40 Figure 4

Mw (kDa)

150 1

2 100 3 4 5 6 7 75 8 9 10 11 12 13 14 15 16 50 17 18 19 20 21 22 23 24 37 25 26 27 28 29 30 31 32 33 25 34 35 36 37 20 38 39 40

41 42 43 44 15 45 46 47 48

10 49

control SecE depl

41 Figure 5

A. Coomassie

control Mw (kDa) SecE depletion Mw (kDa)

42 7 40 100 78 100 8 27 36 48 36 48 4 9 25 4 25 27 75 9 75 10 28 49 10 28 11 11 49 51 51 1 18 29 1 18 29 50 50 5 12 43 5 12 19 6 13 19 37 3 3 52 54 14 37 14 50 50 15 16 20 20 30 15 16 30 21 37 41 44 21 37 44 31 25 31 25 17 26 17 26 32 32 45 45 20 22 20 22 2 33 39 2 33 39 53 53 23 38 23 38 15 24 34 46 24 34 46 15 35 47 35 47

pI4.2 5.0 5.8 6.6 4.2 5.05.8 6.6

B. phosphorimaging

control Mw (kDa) SecE depletion Mw (kDa)

7 8 100 42 4 25 27 36 48 7 8 100 9 36 48 10 28 75 4 9 25 27 10 75 51 28 18 29 51 5 18 29 12 50 5 50 12 3 6 13 14 37 3 50 14 52 54 37 50 16 20 30 16 20 21 31 37 25 30 21 31 37 41 25 26 45 26 22 20 45 22 20 2 33 39 33 39 23 38 53 2 15 23 38 53 46 15 24 34 46

pI4.2 5.0 5.8 6.6 4.2 5.05.8 6.6

C. 100 Coomassie phosphorimaging 10 3 yfgL 44 yfiO 38 ompX 16 tsx 37 yfiO 14 nlpB 29 tolC 7 ostA 23 ompX 31 mipA 28 btuB 20 tsx 17 ybaY 53 ompX 8 yaeT 27 fhuA 19 mdtE 45 ompW 21 mipA 4 fhuE 30 nlpA 2 dcrB 26 yiaF 49 yicH 25 cirA 48 fusA 36 fepA 15 metQ 1 39 pal 33 pal 1 mdtE 0.1 12 fadL 50 ompA

0.01 fold-change (SecE depl/control)

0.001

42 Figure 6

A. Coomassie

control Mw (kDa) SecE depletion Mw (kDa) 22 1 1 22 48 66 2 24 66 83 2 24 68 83 98 100 68 98 100 38 40 69 84 69 84 16 31 38 85 108 16 31 85 41 41 108 70 42 54 86 109 75 70 3 18 25 33 87 3 42 54 86 75 7 18 25 33 87 109 71 92 7 92 9 9 71 43 88 99 43 99 72 100 88 34 34 72 100 10 10 11 35 49 55 67 11 35 44 56 73 94 49 67 94 95 50 4 56 95 50 4 44 73 8 59 57 8 57 101 58 101 59 58 102 102 60 74 5 60 74 96 17 89 96 103 17 89 103 110 5 28 61 28 61 75 104 21 75 104 110 21 62 105 37 45 62 105 37 45 111 111 6 6 12 36 50 12 29 36 50 29 106 32 106 32 76 112 63 63 76 97 112 27 97 113 27 113 51 51 90 19 39 64 90 19 64 46 25 39 46 25 20 65 20 15 52 77 114 15 65 114 23 115 52 77 13 78 23 115 20 13 78 14 53 79 20 37 80 116 37 53 47 14 79 80 116 107 107 81 15 15 30 81 82 30 82 26 91 117 26 10 91 117 10

Native mass 880 440 160 66 Native mass 880 440 160 66 (kDa) (kDa)

B. phosphorimaging

control Mw (kDa) SecE depletion Mw (kDa)

24 48 69 84 100 100 85 108 69 84 85 41 41 108 86 109 75 3 42 54 3 54 86 109 75 9 71 92 9 71 92 88 43 100 43 88 10 34 72 72 100 11 35 93 10 34 49 55 94 93 44 11 35 49 55 94 56 95 50 44 56 101 95 50 59 57 57 101 58 96 102 59 17 58 102 80 104103 96 28 110 17 89 104 103 21 105 37 28 110 45 111 45 105 37 21 111 12 36 50 29 106 12 36 50 32 29 106 63 76 97 112 32 76 97 112 27 113 27 63 113 39 46 64 25 39 46 64 25 65 114 15 77 65 77 114 23 115 15 78 20 23 78 115 53 20 79 53 79 47 80 80 107 107 81 15 81 82 15 82

91 117 10 91 117 10

Native mass 880 440 160 66 Native mass 880 440 160 66 (kDa) (kDa)

43 Figure 6

45 ydjN 87 nuoC 97 metQ 67 nuoF 106 ompA, metQ Coomassie 25 sdhA 84 plsB phosphorimaging 17 wzzE, ybdG 69 mrcA 58 dacC, mdtE, cysA 101 emrA 48 secA 18 nuoC 108 dnaK 44 hemX 41 aas, yhjG 20 nuoI 9 groL 80 glnP 23 narI # 33 nuoC 81 yhcB 39 mscS 28 manX 109 dld 7 nuoC 83 ydij 86 ppiD 59 cysA, dacC, hemY 91 yajC 79 atpF 43 msbA 22 narG 96 proP 72 cydC, cydD 76 metQ 107 pgsA 19 nuoB 61 dppD 89 ybhG 36 atpG 21 wzzB 95 ndh 105 glpT 102 dacA, dacC 11 atpD 1 yjeP 14 atpF 54 ppiD 27 sdhB, manZ 111 lepB 117 yajC 71 secD 74 metN 99 yijP 60 dacA, dppF 75 degS 10 atpA 5 hflC 100 ydgA 6 ybbK, corA 65 yfgM 68 mgtA, copA 85 gcd 35 atpD 112 metQ, rpsB 66 nuoG 103 pyrD 16 fadE, plsB 31 yrbD 37 atpF 51 yhbG 70 frdA 88 oxaA 12 atpG 46 exbB 90 nuoB 52 ygim, tolQ 13 atpH 4 hflK 104 gadC 110 cysK 63 metQ, nlpA 82 yhcB 50 cyoA 94 dadA, zipA 3 ftsH 34 atpA 24 acrB 26 sdhD 2 kefA 42 ppiD 0.1 1 10 100 1000 92 secD 92 yicH fold-change (SecE depletion/control) 8 creD, hemY 73 cydA 49 cyoB 53 atpF 77 rnfG

0.00 0.01 0.10 1 10 fold-change (SecE depletion/control)

44 Figure 7

Coomassie phosphorimaging 5 3 41; aas, 2TMs 41; yhjG, 2TMs

2 3 41; aas, 2TMs 41; yhjG, 2TMs 69; mrcA, 1TM 81; yhcB, 1TM 6; ybbK, 1TM 2 81; yhcB, 1TM 1 6; ybbK, 1TM 16; fadE, 2TMs 1 fold-change (SecE depl/control) fold-change (SecE depl/control)

0 0 0 100 200 300 400 500 600 700 800 900 0 100 200 300 400 500 600 700 800 900 translocated domain translocated domain (nr. of amino acids) (nr. of amino acids) B.

all periplasmic domains shorter than 60 aa at least one periplasmic domain longer than 60 aa

Coomassie phosphorimaging

5 3

2 3

2 1

1 fold-change (SecE depl/control) fold-change (SecE fold-change (SecE depl/control) fold-change (SecE

0 0 024 6810121416 02 46 810121416 nr. of TMs nr. of TMs