Structural and functional studies of the Escherichia coli YidC
DISSERTATION
Presented in Partial Fulfillment of the Requirements for the Degree Doctor of Philosophy in the Graduate School of The Ohio State University
By
Seth William Hennon
Ohio State Biochemistry Program
The Ohio State University
2015
Dissertation Committee:
Professor Ross E. Dalbey, Advisor
Professor Thomas Magliery
Professor Dehua Pei
Professor Natividad Ruiz
Copyright by
Seth William Hennon
2015
Abstract
This dissertation examines the structure and dynamics of the E. coli YidC and explores the membrane insertion of TolQ, which requires YidC and SecYEG. Chapter one of the dissertation is a review of membrane protein targeting and insertion in mainly prokaryotic organisms. There are two bacterial targeting pathways: one pathway is for post- translational translocation and involves a wide variety of components including the chaperones trigger factor, SecA, and SecB; the second pathway is for co-translational targeting and occurs by direct binding of the signal peptide of the ribosome nascent chain complex to the signal recognition particle (SRP) followed by targeting to the membrane and SecYEG. The Sec machinery is conserved in all domains of life and is the major translocase for both co- and post-translational insertion. In its membrane protein insertion function, SecYEG can form a complex with a wide variety of proteins including SecA, the ribosome, SecDF(YajC), FtsY, and the YidC insertase. The structure and function of these proteins are discussed in the first chapter. In addition to exporting proteins by the SecYEG translocase, proteins can be exported by the twin arginine translocation (Tat) pathway. In contrast to the Sec pathway, the Tat pathway acts to transport substrates that exhibit fast folding kinetics, require co-factors, or oligomerize in the cytoplasm before export.
Substrates of this pathway have a twin arginine signal peptide which allows targeting to
ii the Tat machinery. The Tat complex is composed of two types of proteins: TatC and a combination of TatA like proteins (TatA or TatB). Many characteristics of this pathway are still under debate including: the targeting factors involved, oligomeric state of the Tat complex, and mechanism of pore formation. However, it is known that TatC is the main recognition site for the signal sequence.
Chapter two reviews the YidC family of proteins in bacteria, chloroplasts (called Alb3), and the mitochondria (termed Oxa1). All of the homologs contain a core group of five conserved transmembrane (TM) segments while Gram-negative bacteria also contain an extra N-terminal transmembrane segment and a large periplasmic domain. In eukaryotes and Gram-positive bacteria, there are two paralogs: one typically binds directly to the ribosome and functions in co-translational insertion and the other functions in post- translational insertion. A big development in the field during the last year was the determination of two crystal structures of this family of proteins. The structures revealed that YidC contains a hydrophilic cavity that spans the cytoplasmic leaflet of the inner membrane and is exposed to water and a lipid environment. An evolutionarily conserved cytoplasmic hairpin domain was also observed and shown to be dynamic in both structures.
Additionally, mechanistic studies have begun to determine the features of a substrate that allow it to be inserted by YidC or by the YidC/Sec complex.
In chapter three, in vivo cysteine cross-linking studies were utilized to probe the proximity relationship and dynamics of the five conserved transmembrane domains of the E. coli
YidC. Thio-specific homo bi-functional reagents of varying spanner lengths and disulfide
iii bond formation catalyzed by iodine were used to probe cysteine pairs located in the membrane border regions or embedded in the membrane. We observed that all of the transmembrane segments probed (TM3, TM4, TM5, and TM6) had a face oriented toward
TM2 and that they all came in close contact in the membrane interior. TM2 and TM3 appeared to be in close contact with each other and formed the most cross-links out of all of the cysteine pairs that were tested. The dimeric state of YidC has long been debated but no dimers were observed in our crosslinking studies even with cysteine mutants in both
TM2 and TM3 which were proposed to be the interface of a YidC dimer. Our studies also revealed that YidC is a very dynamic protein in the membrane vesicles utilized in our experiments. Both rigid and flexible reagents with a wide range of spanner lengths were able to crosslink most cysteine pairs tested in the cytoplasmic border regions of YidC. This flexibility continued into the membrane interior but disappeared as the cysteine mutants were moved toward the periplasmic half of the membrane. Additionally, the loop that links the large periplasmic P1 domain to TM2 was efficiently crosslinked to the periplasmic portions of TM3 and TM4. This region was previously shown to be important for the function of YidC and we propose that it could act to maintain the permeability barrier when
YidC is inserting substrates. These in vivo results support and confirm flexibility that has also been observed in the crystallographic B-factors of the two YidC structures as well as molecular dynamics simulations. They also provide a framework for beginning to tease out the structural dynamics and conformational changes that occur during the catalytic cycle of YidC. However, many of the proximity relations between the TM segments of
YidC determined by crosslinking did not fit with the reported crystal structures. The large
iv discrepancies between the crosslinking and x-ray structures could be due to the fact that
YidC is a very dynamic protein, interacts with many proteins in vivo, or possibly due to some misfolded YidC in our studies.
Finally, chapter 4 investigates the membrane insertion of the E. coli TolQ protein (part of the Tol-Pal complex) which was previously thought to insert into the membrane in a Sec independent manner. This protein, containing a short (19 residue) periplasmic N-tail and a short (18 residue) periplasmic loop, was chosen as a potential substrate of YidC because it has features of known YidC substrates and also because it was proposed to insert by a
Sec-independent mechanism, which suggest that YidC may be involved. By studying the wild-type TolQ protein, we determined that both YidC and the Sec pathway were able to facilitate insertion of the TolQ periplasmic loop and that SecA and the proton motive force
(pmf) were not required. We also made mutations in the periplasmic loop of TolQ to determine which structural properties make a protein dependent of YidC and the Sec translocase. Interestingly, mutations of the loop, which normally contains one positively charged glutamic acid and one negatively charged lysine, altered the requirements.
Increasing the number of charged residues and thus the overall hydrophilicity of the loop caused an increased dependence on YidC and SecYEG and adding two negatively charged residues caused insertion to be completely dependent on both insertases. Based on these results, we hypothesize that the P1 loop of TolQ inserts at the interface of YidC and
SecYEG in the holo-translocon but further studies are needed in order to provide a comprehensive understanding of the structural features that dictate the translocase requirements for membrane insertion. v
Dedication
This document is dedicated to my family.
vi
Acknowledgments
First, I would like to thank my advisor, Dr. Ross E. Dalbey, for his guidance and critical analysis of my data as well as his encouragement and support throughout the course of my graduate studies.
Additionally, I am also grateful for the constructive criticism and helpful suggestions that were provided by Dr. Stephen White and Dr. Andreas Kuhn regarding my cross-linking project. I would like to thank all of my committee members Dr. Thomas Magliery, Dr.
Natividad Ruiz, and Dr. Dehua Pei for their time and guidance.
I would also like to thank former and current lab members Dr. Peng Wang, Dr. Jijun
Yuan, Dr. Lu Zhu, Dr. Raunak Soman, Yuanyuan Chen, Bala Subramani, Karthika
Shanmugam, and Haoze He for their advice and friendship.
Most importantly, I would like to thank my parents, Pam and Dave Hennon; my sister,
Amanda Hennon; and my family members for their love and support. Lastly, I would like to thank my friends for their encouragement and support during this journey.
vii
Vita
October 6, 1984 ...... Born in Warren, OH
2003...... Joseph Badger High School
2007...... B.S. Biology and Chemistry,
Mount Union College
2007 - Present ...... Graduate Teaching and Research Associate,
Department of Chemistry and Biochemistry,
The Ohio State University
Publications
Hennon, S.W., Soman, R., Zhu, L., and Dalbey, R.E. (2015) YidC/Alb3/Oxa1 family of insertases. Journal of Biological Chemistry. Accepted; waiting to be published as a group with other articles.
Hennon, S.W., and Dalbey, R.E. (2014) Cross-linking-based flexibility and proximity relationships between the TM segments of the Escherichia coli YidC. Biochemistry. 53,
3278-3286, doi: 10.1021/bi500257u.
viii
Fields of Study
Major Field: Ohio State Biochemistry Program
ix
Table of Contents
Contents Abstract ...... ii
Dedication ...... vi
Acknowledgments ...... vii
Vita ...... viii
Publications ...... viii
Fields of Study ...... ix
Table of Contents ...... x
List of Tables ...... xv
List of Figures ...... xvi
CHAPTER 1 ...... 1
INTRODUCTION ...... 1
1.1 Overview of bacterial membrane protein translocation ...... 1
1.2 Targeting and recognition of substrates...... 5
SecB ...... 7
SRP ...... 9
x
1.3 The SecYEG pathway ...... 11
SecA ...... 15
SecDF(YajC)...... 17
Mechanism for insertion into the inner membrane ...... 18
1.4 The Tat pathway ...... 19
Translocase composition ...... 20
Targeting to Tat...... 21
Oligomeric State and mechanism of the Tat complex ...... 22
1.5 The YidC pathway ...... 25
1.6 Figures ...... 26
CHAPTER 2 ...... 36
YidC/Alb3/Oxa1 Family of Insertases ...... 36
2.0 Contributions ...... 36
2.1 Introduction ...... 36
2.2 Distribution, topology and function of YidC/Oxa1/Alb3 proteins ...... 37
2.3 Oxa1 Family ...... 39
2.4 Alb3 Family ...... 42
2.5 YidC Family ...... 44
2.6 Recent Insights ...... 46
xi
2.7 Conclusion ...... 52
2.8 Tables...... 54
2.9 Figures ...... 56
CHAPTER 3 ...... 59
Cross-linking-based flexibility and proximity relationships between the TM
segments of the Escherichia coli YidC ...... 59
3.1 Introduction ...... 59
3.2 Results ...... 61
Topology and crosslinking approach ...... 61
Crosslinking of paired Cys residues in helices 2 and 3 ...... 63
Crosslinking of paired Cys residues in helices 2 and 4 ...... 65
Crosslinking of paired Cys residues in helices 2 and 5 ...... 65
Crosslinking of paired Cys residues in helices 2 and 6 ...... 66
3.3 Discussion ...... 67
3.4 Materials and Methods ...... 74
Materials ...... 74
Plasmids and strains ...... 74
Construction of YidC with double Cys residues ...... 74
xii
Expression of YidC Cys pairs and membrane preparation for crosslinking studies
...... 75
Site-directed cross-linking ...... 75
Complementation assay ...... 76
3.5 Tables...... 78
3.6 Figures ...... 81
CHAPTER 4 ...... 92
YidC and Sec translocase facilitate membrane insertion of TolQ, a three-spanning
membrane protein ...... 92
4.1 Introduction ...... 92
4.2 Results ...... 95
TolQ membrane topology and protease assay ...... 95
Insertion of the P1 loop of wild-type TolQ is partially YidC and Sec-dependent
but independent of the pmf ...... 97
The effects of mutations of the charged residues in the P1 loop on translocase
requirements ...... 99
The effects of adding charged residues to the P1 loop on translocation ...... 101
Inhibition of the membrane insertion of the positively charged E173R is due to
the pmf ...... 102
4.3 Discussion ...... 103 xiii
4.4 Materials and Methods ...... 107
Materials ...... 107
Strains, Plasmids, and Growth conditions ...... 107
Protease accessibility studies ...... 108
Mutagenesis of TolQ...... 109
4.5 Figures ...... 110
CHAPTER 5 ...... 118
Conclusions ...... 118
5.1 Summary of work performed ...... 118
5.2 Future directions ...... 120
5.3 Tables...... 123
References ...... 124
xiv
List of Tables
Table 2.1 – Substrates of the YidC homologs ...... 54
Table 3.1 – Complementation efficiency of YidC double cysteine mutants in JS7131 .. 78
Table 3.2 – Crosslinking efficiency of YidC double cysteine mutants ...... 80
Table 5.1 – Comparison of distances between the E. coli YidC structure and the crosslinking results...... 123
xv
List of Figures
Figure 1.1 – Structure of the SRP-FtsY complex ...... 26
Figure 1.2 – Structure of trigger factor ...... 27
Figure 1.3 – Structure of SecB ...... 28
Figure 1.4 – Structure of the Sec complex ...... 29
Figure 1.5 – Structure of SecA...... 30
Figure 1.6 – Structure of SecDF ...... 31
Figure 1.7 – Structure of SecA bound to SecYEG ...... 32
Figure 1.8 – Structure of the holo-translocon ...... 33
Figure 1.9 – Schematic of protein targeting in E. coli (continued) ...... 34
Figure 2.1 – Crystal structures of YidC ...... 56
Figure 2.2 – Substrate contacts of YidC ...... 57
Figure 2.3 – Model for the insertion of a single-span membrane protein ...... 58
Figure 3.1 – Cys-based alkylation method mapping the membrane border regions, and complementation assay (continued) ...... 81
Figure 3.2 – Crosslinking of paired Cys residues in TM2 and TM3 (continued) ...... 83
Figure 3.3 – Crosslinking of paired Cys residues in TM2 and TM4 ...... 85
Figure 3.4 – Crosslinking of paired Cys residues in TM2 and TM5 ...... 86
xvi
Figure 3.5 – Crosslinking of paired Cys residues in TM2 and TM6 ...... 87
Figure 3.6 – Cartoon showing one possible arrangement of TM2, TM3, TM4, TM5 and
TM6 of YidC based on the crosslinking analysis ...... 88
Figure 3.7 – The amount of YidC detergent solubilized from membranes prepared from
BL21 or C41 bearing pEH1YidC ...... 89
Figure 3.8 – Crosslinking of specific Cys pairs expressed in C41 ...... 90
Figure 3.9 – Comparison of reagent crosslinking efficiency ...... 91
Figure 4.1 – Membrane topology of TolQ and expression controls (continued) ...... 110
Figure 4.2 – Expression of the wild-type TolQ in the pLZ1 plasmid under various conditions (continued) ...... 112
Figure 4.3 – Insertion requirements of positively charged mutants ...... 114
Figure 4.4 – Insertion requirements of negatively charged mutants ...... 115
Figure 4.5 – Insertion requirements of charged and more hydrophilic mutants ...... 116
Figure 4.6 – Dependence on the pmf for membrane insertion ...... 117
xvii
CHAPTER 1
INTRODUCTION
1.1 Overview of bacterial membrane protein translocation
Membranes are found in Prokaryotic, Eukaryotic, and Archaea kingdoms of life where they control the flow of molecules and ions between a cell and its surroundings. The membranes of organelles also function to sequester metabolic reactions in eukaryotic cells. Membranes are composed of a variety of phospholipids and contain proteins which perform a variety of crucial biological functions that are embedded in the phospholipids or peripherally bound to them [1]. Depending on the organism, there can be a single membrane or multiple membranes such as with eukaryotes, which have a variety of organelles. There are many different eukaryotic organelles each with its own membrane: the nuclear envelop, the plasma membrane, mitochondria have inner and outer membranes, endosome, golgi complex, lysozome, peroxisome, endoplasmic reticulum, chloroplast envelope, and thylakoid membrane. Cellular membranes are not static structures; proteins and
1 phospholipids can diffuse laterally throughout the membrane in order to perform their specific tasks.
Bacterial cells are less complex than eukaryotic cells and lack organelles and a membrane bound nucleus. They are defined as prokaryotes, which also encompasses a separate kingdom, Archaea. Bacteria can be divided into two distinct and fundamentally different groups based upon their ability to bind Gram’s stain. Gram-negative and Gram-positive bacteria vary based on the number and composition of their membranes. Gram-positive bacteria have the least complicated cell envelope and contain only a cytoplasmic membrane that is surrounded by many layers of peptidoglycan. Gram-negative bacteria have two membranes which are separated by a periplasm containing layers of peptidoglycan; an inner cytoplasmic or plasma membrane, which is similar to the Gram-positive cytoplasmic membrane, and an outer membrane enriched with lipopolysaccharides. There is a subset of bacteria which have an atypical membrane such as the thick hydrophobic mycolic acid layer surrounding the cytoplasmic membrane in mycobacteria.
The phospholipids that compose the lipid bilayer of membranes are amphipathic molecules with a hydrophilic head region and a hydrophobic tail. The tails typically contain two fatty acid hydrocarbon chains of varying length and form the hydrophobic core of the bilayer.
Polar groups and phosphate make up the hydrophilic head which is located on the exterior of the bilayer and is in an aqueous environment. Membrane proteins also bear the same general characteristics of the bilayer; the transmembrane segments are generally enriched
2 in hydrophobic amino acids while the peripheral and aqueous exposed regions typically are enriched in polar residues.
On average, about 30% of the proteome encodes membrane proteins and 20% of the E. coli proteome is composed of cytoplasmic membrane proteins [2]. Membrane proteins perform a wide variety of functions and can form complex machineries composed of multiple subunits; they can vary greatly in size, the number of transmembrane segments, and topology. There are membrane embedded sensors encoded by every cell that enable the cells to recognize environmental stimuli and act to transmit the stimuli across the membrane into the interior of the cell. Additionally, a majority of known drug targets are membrane proteins which illustrates their importance for cell function. Membrane proteins are able to perform a wide variety of functions including: transport across membranes or between cellular compartments, nutrient and protein transport, signaling, mediating cell division, electron transfer, respiration, photosynthesis, and ATP synthesis in the mitochondria and chloroplast [3, 4].
Ribosomes function to synthesize almost all proteins in the cytosol which poses a problem for any protein whose destination lies beyond the cytosol. This raises the question of how proteins are transported from the cytosol to their final destination in the cell. In prokaryotes, proteins are transported from the cytosol to the cell surface or even secreted from the cell. In eukaryotes, proteins are inserted into or across the ER membrane and, depending on the protein, can move through the secretion pathway in order to be secreted
3 out of the cell. Other eukaryotic proteins can be imported from the cytosol into the peroxisome, mitochondria, or nucleus.
Specialized proteins, which are conserved across all domains of life, are needed in order to insert proteins into the endoplasmic reticulum of eukaryotes or the cytoplasmic membrane of a bacterial cell. There are two distinct targeting pathways to accomplish this task; one for co-translational targeting and one for post-translational, both of which are discussed in more detail in Section 1.2. After membrane proteins reach their appropriate membrane and are inserted, they still need to be assembled and folded into the correct tertiary structure in order become active.
The membrane biogenesis of proteins generally occurs in three stages: targeting, insertion, and folding. Occurring in the cytosol, the first stage involves the recognition of the nascent chain’s signal peptide by the Signal Recognition Particle (SRP) followed by binding of the ribosomal nascent chain-SRP complex to the SRP receptor at the membrane surface. The transmembrane segments of the nascent chain are then inserted into the membrane and some of the hydrophilic loops between transmembrane segments are translocated across the membrane. Once in the membrane, the protein is folded in the final step followed by assembly into quaternary structures, if necessary. Should one of these steps fail, there are quality control pathways in the cell that are able to recognize and degrade misfolded or aggregated proteins [5, 6].
Many intriguing questions remain in the field of membrane protein biogenesis in bacterial systems. How many pathways are involved for membrane protein insertion? In addition
4 to the established pathways, are there others that have not been discovered? Do the pathways act individually or can they function cooperatively? Within each pathway, how are the insertases able to recognize the substrates that they insert? Is there redundancy in these pathways in order to compensate for a deficiency? In this chapter, we will focus on protein targeting to the membrane and the insertion pathways followed by a description of the main players which are the Sec, Tat, and YidC pathways.
1.2 Targeting and recognition of substrates
Because proteins destined for transport are synthesized in the cytosol, they need to be recognized as such during translation and then targeted to the proper machinery, which is usually SecYEG in bacteria. Recognition of the protein to be exported is typically accomplished by the Signal Recognition Particle (SRP) recognizing an N-terminal signal peptide of the membrane protein as it exits the ribosomal tunnel; SRP will be discussed in more detail later in this chapter. The signal peptide, used for targeting secretory proteins to the membrane, is an N-terminal extension of the protein that has a tripartite structure and is typically removed after translocation.
Signal peptides act to target nascent chains to the membrane and are generally twenty to thirty amino acid extensions of a protein. The structure of signal peptides is conserved and contains: an N domain with one to three positively charged amino acids at the N-terminus; a central hydrophobic core composed of ten to fifteen residues (H domain); and a polar, hydrophilic domain at the C-terminus of the signal peptide (C domain) containing the
5 signal peptidase cleavage site. There are very few conserved residues in the signal peptide but they can be predicted with accuracy based upon their conserved tripartite structure [7].
Most signal peptides attached to the exported proteins are believe to span the membrane prior to being removed by Leader Peptidase (signal peptidase). Leader peptidase is a novel membrane bound serine protease whose C-terminal catalytic domain is located in the periplasm and contains a Ser-Lys dyad for catalysis [8]. The function of leader peptidase is to release the mature exported protein from its signal peptide which acts as a membrane anchor sequences; e.g., unprocessed exported proteins accumulate in the cytoplasmic membrane. This ability of signal peptides to act as a transmembrane segment is further supported by nuclear magnetic resonance (NMR) studies which showed that signal peptides have a stable -helical domain at the N-terminus and a flexible C domain in a membrane environment [9, 10]. The leader peptidase cleavage site is located in the C domain and contains small, neutral amino acids at the -1 and -3 positions with the consensus sequence being Ala-X-Ala. In many bacteria, integral membrane proteins do not have a cleavable signal sequence because their hydrophobic transmembrane segments act as a signal for targeting and insertion as well as membrane anchors [11].
There are two major targeting pathways in E. coli and both target proteins to the SecYEG translocation machinery. The cytosolic chaperone SecB targets pre-proteins containing signal peptides and acts in a post-translational mechanism while co-translational targeting of inner membrane proteins occurs through the action of ribosome bound nascent chains
(RNC) which are targeted by SRP. The competition between these two pathways occurs
6 as soon as the nascent chain emerges from the ribosomal exit tunnel [12] when SRP and trigger factor (a peptidyl-prolyl cis-trans isomerase) attempt to bind to the N-terminal region of the protein [13]. SRP seems to be able to distinguish its substrate membrane proteins from cytosolic proteins based on the hydrophobicity of the transmembrane segments [14, 15]. Hoffman et al. have shown that trigger factor is able to bind to the ribosome which allows it to be near nascent chains being synthesized and exiting the ribosome [16]. The binding of trigger factor to the signal peptide acts to block the interaction of SRP which allows targeting to the SecB pathway [13]. During extension of the post-translationally targeted nascent chain by the ribosome, chaperones are needed in order to keep the protein from aggregating and in an unfolded state until it can be targeted to SecYEG [17]. Additionally, Calloni et al. used a proteomics approach to identify substrates of trigger factor; most were found to be outer membrane proteins which further supports the role of trigger factor as a post-translational targeting factor [18].
SecB
Protein synthesis by the ribosome in E. coli and translocation of secretory pre-proteins containing signal sequences does not occur concurrently so chaperones are needed in order to keep proteins in an unfolded state because translocation via the Sec pathway occurs in an unfolded manner. Additionally, unfolded proteins of any type are susceptible to degradation in the cytosol and have a tendency to aggregate because of their hydrophobic groups which further increases the need for a chaperone to be present. SecB is a chaperone
7 specifically for proteins destined for secretion and is mainly found in the proteobacteria
[19]. Its native quaternary fold and functional state is as a homo-tetrameric protein which is a dimer of dimers [20].
In the x-ray crystallography structure of SecB by Xu et al. from Haemophilus influenza, a channel was observed on either side of the tetramer with a length of approximately 70Å and was thought to act as the binding site for unfolded proteins. Two sub-sites were also observed: a deep cleft lined with aromatic residues which the authors proposed could bind to hydrophobic regions of the polypeptide and a shallow hydrophobic groove believed to be involved with the binding of β-sheet regions of nascent polypeptides [20]. It seems as though SecB binds to regions that are typically buried in the tertiary structure [21] which makes sense based on the composition of the two binding sub-sites. A general binding motif was determined based upon binding studies with peptides [22]; the consensus was a sequence of about nine amino acids enriched in aromatic and basic residues. However, sequences of this type are equally distributed in both soluble and secreted proteins which does not explain how SecB is able to selectively bind substrates based upon this consensus sequence.
Some bacteria lack SecB and it is thought that other general folding chaperones (such as
DnaK and DnaJ) are able to take its place [23]. However, SecB is a unique folding chaperone because it preferentially binds with high affinity to SecA when it is in a complex with SecYEG [24]. The C-terminus of SecA contains the so-called zinc binding domain which is a conserved sequence of twenty-two amino acids that act as a SecB binding
8 domain [25]. This SecA domain, utilizing three cysteine residues and one histidine residue to coordinate a zinc ion, is essential for the SecA/SecB binding interaction to occur [26].
Binding occurs via electrostatic interactions and not directly to the zinc; however, the zinc is required because it acts to stabilize the fold of the binding domain in SecA [27]. High affinity binding of SecB to the dimeric form of SecA was observed while a low affinity was seen for the monomer; this led Randall et al. to propose that SecB is released from the membrane when SecA transitions from the dimeric to monomeric form [28].
SRP
Co-translational targeting is accomplished through the action of SRP and its receptor, FtsY, and is the major pathway for the insertion of most cytoplasmic membrane proteins. The eukaryotic SRP contains six protein subunits and one RNA component; while the E. coli
SRP is very similar, it is much less complex than the eukaryotic homolog [29]. It contains a 48 kDa GTPase component (Ffh or the fifty-four homolog of eukaryotic SRP54) and a
4.5S RNA component which has been shown to be important for stability of SRP [30, 31] as well as large structural changes in the holo-enzyme because of its role in coordinating the communication between substrate and the GTPase domains of SRP and FtsY. SRP is able to recognize and bind the signal sequence of a pre-protein based upon its increased hydrophobicity [32, 33]. Targeting of the SRP and RNC complex to the membrane is accomplished by the receptor FtsY which can bind peripherally to the polar head groups of the lipids that compose the membrane. FtsY is different from the eukaryotic homolog
9 because it lacks the membrane integrated β-subunit which causes it to be distributed between the cytoplasm and the membrane [34]; however, the function of the cytoplasmic portion is still unclear [35].
When FtsY associates with the membrane, it does so by interacting with anionic phospholipids [36] but it has also been shown to interact with SecYEG [14] which the authors propose enables the transfer of the nascent chain to the Sec machinery. The GTP binding affinity of both proteins is increased when Ffh binds to FtsY and allows GTP to bind to sites on both proteins [37, 38]. Transfer of the RNC to SecYEG occurs when GTP is hydrolyzed by Ffh and FtsY; the entire Ffh/FtsY/ribosome complex dissociates and the ribosome is then able to bind to SecYEG so that translocation can occur [39].
The fifty-four homolog of SRP has three major domains: the amino terminal N domain which interacts with the ribosome, the GTPase (G) domain contains the GTP binding site and the methionine rich C-terminal M domain responsible for recognition of the signal sequence as well as binding to the RNA component of SRP. FtsY also contains an N-G domain whose acidic N domain is involved in membrane targeting [40]. Because the N-G domains of both proteins are closely associated and very similar, the binding of the two proteins occurs through the interaction of these two domains [41, 42]. The interaction of
SRP and FtsY is sensed by the RNA component of SRP which then facilitates the recruitment of the ribosome followed by the transfer of the RNC to the translocase [43].
An x-ray crystallography study of SRP showed that the M domain contains a deep groove that is lined with hydrophobic residues including many methionine residues, a large
10 number of these hydrophobic residues are conserved [44]. This groove is believed to be the signal sequence binding site and is closely associated with the RNA domain; this caused
Batey et al. to propose that the binding site is composed of both RNA and protein [45].
A model for how SRP targets ribosome nascent chains to the membrane has been proposed in which this targeting process is directed by the concerted action of the SRP and FtsY proteins. During the targeting process, both proteins have discrete conformational changes that occur in a sequential manner. Initially, SRP binds to the ribosome and remains bound with high affinity if a nascent chain is exposed from the exit tunnel of the ribosome. This
SRP complex with the RNC allows the M domain to bind to the signal peptide which is followed by a conformational change in the G domain and thus increased affinity for GTP.
Once the complex is bound to FtsY, the closed state of the complex occurs such that the
N-G domains form an interface surrounding the GTP molecules. GTP hydrolysis causes the complex to dissociate and allows the transfer of the RNC to SecYEG for translocation
[46, 47]. SRP can also be used to target Sec independent proteins to YidC in a manner that is not completely understood. This function is utilized for MscL and the non-native, procoat-lep construct; SRP targeting to YidC is further supported by crosslinking data to the C-terminus of YidC [48].
1.3 The SecYEG pathway
The SecYEG complex is the primary insertase and translocase in E. coli and has homologs in all kingdoms of life [49]. The main components of this translocase machinery (SecY,
11
E, G, and A) were identified through genetic screens; mutations that caused secretion defects were called sec alleles while suppressors that allowed protein secretion of a defective signal peptide were prl alleles and were based upon the gene names [50, 51].
SecYEG is able to perform two similar functions: secretion across the membrane as a
“translocase” and insertion into the inner membrane as an “insertase.” The Sec complex is also capable of transporting substrates in two directions; it can laterally insert them into the membrane and transport them vertically across the membrane because of its unique structural features which will be discussed later in this chapter.
In E. coli, the main protein channel is composed of three integral protein subunits: SecY,
SecE, and SecG which interact in order to form a stable quaternary structure [52]. As previously stated, this complex is evolutionarily conserved and is called Secαγβ in eukaryotes. The bacterial SecYEG machinery is able to contact and work cooperatively with a variety of proteins in order to carry out its co- and post-translational insertion duties including: SecA, the ribosome, FtsY, SecDF(YajC), and YidC. Solved in the year 2004 from the archaeal Methanococcus jannaschii, the first x-ray crystallography structure revealed the SecYEβ complex in the resting state and was a breakthrough in the membrane protein field [53].
SecY contains ten transmembrane (TM) segments arranged into a clamshell shape containing two halves; one half is composed of TMs 1-5 while the second half is TMs 6-
10 which are connected by the loop between TM 5&6. The structure also revealed that cytoplasmic loops 4-6 in the C-terminal half of the protein are exposed so that they may
12 bind to targeting factors and the ribosome [54-56]. SecE acts as a molecular clamp to hold the two domains of SecY together on the “back” of SecY while the non-essential SecG makes very little contact with SecY and is located at the periphery of the structure.
SecY forms a unique channel structure with an hour-glass shape on both sides of the membrane containing a constricted hydrophobic pore in the center of the membrane composed of six isoleucine residues. These isoleucine residues point toward the center of the pore and are believed to maintain the permeability barrier by forming a seal around the polypeptide during translocation [57]. Each side of the channel acts as a funnel which goes from a diameter of approximately 20-25Å at the periphery to about 4Å at the pore ring.
Because the structure was solved in the closed state, there is an α-helical plug domain, which is an extension of TM2a, blocking the periplasmic side of the pore ring. The plug is displaced during translocation which allows for an open state and an aqueous path from one side of the membrane to the other so that translocation of the unfolded polypeptide can occur [58]. Deletion of the plug does not significantly affect protein translocation [59]; however, these mutants have been shown to fluctuate between the open and closed state
[60] even in the absence of signal peptide binding. Li et al. proposed that neighboring loops can replace a deleted plug domain [61] while permanently displacing the plug, instead of removing it, is toxic to the cell [58] which most likely occurs because of the loss of the permeability barrier.
SecYEG is also capable of performing lateral integration of TM segments into the membrane which occurs through a lateral gate to the lipid bilayer in its structure. Based
13 upon the M. jannaschii structure, the lateral gate is located in SecY between TMs 2b&3 and TMs 7&8 which correspond to the “front” part of the translocon. Lipids do not enter the Sec channel when the lateral gate is open during translocation and Gumbart et al. propose this does not occur because the exiting TM segments block lipids from entering the lateral gate region [62]. The binding of the signal peptide to the lateral gate activates the channel and opens the lateral gate [63].
Historically, the oligomeric state of the functional SecYEG has been controversial; the monomeric state is sufficient for translocation and insertion in vivo and in vitro but di- and multi-meric states have been observed through native PAGE, cross-linking, and electron microscopy [64-68]. Two different models have been proposed for the interaction of two copies of SecYEG in the dimeric state: back to back and front to front. The back to back orientation occurs through the interaction of two SecE TM segments and was observed in crosslinking and cryo-EM studies [69, 70]. The front to front interaction occurs by the lateral gate regions of two SecY proteins interacting to form a large pore and has been proposed based upon a low resolution electron density map and in vivo crosslinking studies
[68, 71]. A recent study by Park et al. showed that both dimeric states can exist in a resting
SecYEG translocon based on in vivo cysteine crosslinking experiments [65]. It is likely that both states transiently exist in the cellular membrane and they could be due to interaction with various binding partners and/or substrate dependent changes.
14
SecA
Because SecYEG does not consume energy in order to translocate substrates, some sort of driving force is required. This force can be provided by the proton motive force, ribosomal protein synthesis during co-translational insertion, or from SecA for post-translational insertion. SecA is an ATP dependent motor protein which associates with almost all of the translocation components including: SecB, the ribosome, and SecYEG. It is able to exist in soluble and membrane bound forms much like FtsY and its membrane association involves negatively charged phospholipids and the SecYEG cytoplasmic loops [71-73]. It has been shown that the ATPase activity of SecA is regulated by unfolded pre-proteins, the
Sec complex, phospholipids, and SecB [72, 74]. SecA actively drives insertion by utilizing
ATP hydrolysis and a two helix-finger mechanism; the proton motive force (PMF) is also a driving force to translocate the polypeptide chain and acts in some cases to determine the orientation of the signal peptide [75].
SecA contains eight distinctive domains in its tertiary structure: the N-terminal nucleotide binding domain 1 (NBD1), NBD2, peptide binding domain (PBD), helical scaffold domain
(HSD), helical wing domain (HWD), intramolecular regulator of ATP hydrolysis 1 (IRA1),
IRA2, and the C-terminal domain (CTD). It contains a DEAD motor like DNA and RNA helicases which is composed of NBD1&2 and the PBD. The motor function comes from
NBD1&2 interacting to form a single ATP binding domain and ATP hydrolysis causes conformational changes within the motor domain and PBD. The PBD not only binds the substrate and translocon but it also forms the substrate binding clamp in cooperation with
NBD2 and the HSD. The CTD contains the zinc-binding region and is used to bind SecB 15 and phospholipids; it is not required for catalysis but acts to inhibit extraneous ATP hydrolysis when the translocon is not bound. The native state remains under debate but
SecA appears to form an anti-parallel dimer in most structural and functional studies [76-
78].
Recently, SecA has been shown to bind the ribosome at the L23 protein which is near the exit tunnel and has been hypothesized to be involved with targeting [79]. Huber et al. also showed that two conserved positively charged residues in the -helical linker domain C- terminal to NBD2 are required for binding to a region of acidic and polar residues on L23.
This study reported an increased affinity of SecA for RNCs based on co-sedimentation studies. The authors proposed that SecA is able to sample the nascent chains emerging from the exit tunnel in order to look for less hydrophobic (post-translational targeting) signal peptides. If it is unable to find one, it can dissociate and allow the ribosome to bind to SecYEG for co-translational insertion. Otherwise, it can remain bound to the nascent chain and allow SecB or some other chaperone to bind for post-translational targeting.
Wu et al. showed that ribosome binding to SecYEG was increased when nascent chains had a hydrophobic sequence exposed from the exit tunnel [80]. Additionally, SecA was able to compete with SecYEG for ribosome binding of both translating and non-translating ribosomes. Two ribosomal SecA binding sites were identified by Singh et al. via cryo-EM studies; one site was located on L23 and the other site was in proximity to the L22 and L24 proteins which could allow two copies of SecA to bind at the same time [81]. Their cryo-
EM structure of SecA bound to the ribosome was compared with the cryo-EM structure of
16 trigger factor bound to L23 [82] and no steric clash was observed between the binding location of SecA and that of TF. The authors proposed that both SecA and trigger factor could bind simultaneously to the ribosome and that targeting could occur based upon whichever is able to recognize its substrate. However, steric clash was observed when the structure of SecA bound to the ribosome and the SRP/ribosome [83, 84] structure were superimposed which led to the conclusion that the two proteins must compete for binding to the ribosome.
SecDF(YajC)
Another heterotrimeric complex is able to interact with SecYEG in order to enhance its translocase function and is composed of SecD, SecF, and YajC [85]. A crystal structure of the SecD&F proteins, each containing six TM segments, in a complex revealed that they both have large periplasmic domains [86]. SecDF is not strictly required for protein translocation but SecDF knockouts have cold sensitive phenotypes and are defective in translocation [87]. The complex is thought to act via PMF driven conformational changes in the periplasmic domains which act to enhance translocation by preventing the translocated region from sliding backward through the Sec channel.
YajC has a single TM segment with a large cytoplasmic domain and is not required for functionality of the complex. Other than being associated with SecDF, there is no known function of YajC [85]. Another function of the SecDF(YajC) complex is that it is thought to recruit YidC to SecYEG by acting as a bridge between the two complexes [88, 89].
17
Additionally, Tsukazaki et al. propose that SecDF pulls substrates on the periplasmic side of the SecYEG channel through conformational changes associated with the proton transfer
[90]. In this process, SecDF conducts protons to the cytosol from the periplasm.
Mechanism for insertion into the inner membrane
After targeting of RNCs to the inner membrane by SRP and FtsY, the ribosome is bound to SecYEG and their respective channels are aligned via the cytoplasmic loops on the C- terminal half of SecY and ribosomal proteins (L23, L24, L29) [91-93]. Frauenfield et al. propose that the contact of the ribosome with lipids acts to alter the membrane environment near the lateral gate such that it is poised for protein insertion [91]. Binding of the ribosome switches SecYEG to the pre-open state which involves partial opening of the lateral gate while the central pore is still plugged by the helix [92] and the widening of the central pore
[94]. The signal anchor sequence of the membrane protein is then able to insert into the channel and migrates toward TMs 2b and 7 of the lateral gate. This binding interaction is believed to then allow the plug domain to be displaced and the lateral gate to open which switches SecYEG to the open state for insertion to occur. Inserted TM segments are able to leave the lateral gate as other TMs are being translated and inserted through the channel of the Sec complex.
18
1.4 The Tat pathway
Generally, Twin arginine translocation (Tat) substrates are folded before translocation, oligomerized, or have cofactors which were inserted in the cytosol prior to being exported.
A large number of the substrates in bacteria are redox proteins required for anaerobic respiration. Other Tat substrate functions include cell envelope biogenesis and remodeling, cell separation, energy metabolism, iron and phosphate nutrition, nitrogen fixation, and virulence factors in some pathogenic bacteria [95]. Some substrates can be oligomeric complexes with only one subunit of the complex containing a signal sequence with the other subunits piggy-backing on it during transport. Another type of substrate exported by the Tat complex can be a secreted protein without a signal sequence that is exported by Tat because it folds very rapidly [96, 97].
Signal sequences of the Tat pathway contain the same general features and tripartite structure as with Sec signal peptides and are cleaved by signal peptidase I (SP1). They have a lower hydrophobicity than Sec signal peptides and contain positive charges on either side of the SP1 cleavage site. A distinct consensus sequence S-R-R-x-F-L-K is contained in the signal peptides of Tat substrates and the twin arginine residues are the reason why this pathway is called the twin arginine translocation pathway. This conserved motif is located at the C-terminal end of the N-region and the arginine pair is almost completely conserved. Mutation of the twin arginines to lysines abolishes translocation and there are only a few exceptions where one of the residues can be replaced by a Lys, Asn, or Gln.
19
Interestingly, recent studies by van der Ploeg et al. have shown that Tat substrates can be redirected to the Sec translocase by high extra-cellular salt concentrations [98, 99].
Translocase composition
There are two types of Tat subunits: the TatC protein, which is the large central core of the complex containing 6 TM segments, and some combination of TatA like proteins (TatA and TatB) which have a single TM segment. Gram positive bacteria have very simple Tat machinery containing one copy each of TatA and TatC. Gram negative bacteria and chloroplast complexes are composed of TatA, TatB, and TatC. Yen et al. have proposed that TatB came about because of gene duplication of TatA since they have the same composition [100]. TatB proteins have different functions from TatA and are thus believed to be an early duplication.
TatA contains an N-terminal transmembrane segment which is connected by a hinge to an amphipathic helix followed by a C-terminal tail as predicted by sequence analysis. Several
NMR structures have confirmed that TatA adopts the predicted conformation and further studies have revealed that it adopts a Nout-Cin orientation within the membrane [101, 102].
TatB has a structure very similar to TatA but generally has a longer C-terminal tail. In E. coli, transport of Tat substrates is abolished without TatB and can be compensated very slightly by TatA for synthetic substrates with the twin arginine signal sequence attached to reporter proteins [103, 104]. X-ray crystallography studies by Rollauer et al. revealed that
TatC adopts a ”curved wall” structure which is covered by a periplasmic cap [105]. This
20 cap is formed from the first two periplasmic loops of TatC and acts to cover the periplasmic side of the groove that travels along the concave face of the curved wall.
Targeting to Tat
There are no known targeting factors for the Tat pathway but substrate signal sequences have been shown to bind to DnaK, Trigger Factor, SlyD, and FK506 binding proteins [106-
108]. Another example of chaperone binding involves some bacterial redox proteins which require specific chaperones called redox enzyme maturation proteins (REMPs). These chaperones bind to substrate signal sequences and are thought to prevent targeting to Tat until folding is completed at which point the REMPs are able to dissociate [109].
Substrates have been shown to bind to the lipid bilayer [110] but further studies suggested that lipid binding is not a critical step for proper targeting to the membrane [107, 111].
The main recognition site for twin arginine signal peptides has been identified to be on
TatC. Cross-linking studies utilizing photoprobes introduced into the signal sequence revealed that TatC is able to crosslink when the signal peptide binds to the TatBC complex
[107, 112]. TatB was also shown to be near the signal sequence through cross-linking studies but it was determined that the observed cross-linking only occurred when TatC was present [112]. Suppressor mutations were used to discover that the signal sequence recognition site lies on the N-tail and first cytoplasmic loop of TatC [113-115]; this was then confirmed by the crystal structure which showed that the R-R binding site contained two negatively charged residues to bind the twin arginines [105].
21
Oligomeric State and mechanism of the Tat complex
Blue Native PAGE of E. coli and pea thylakoids estimate that the composition of the TatBC complex is somewhere between four to eight copies in a 1:1 ratio of TatB to TatC [116,
117]. The arrangement of the protein components in these complexes appears to be a core made up of TatC proteins with TatB proteins bound to the core [118-120]. Substrate binding has been shown to stabilize the complex and thus it is believed that the TatBC complex undergoes large conformational changes during the binding and insertion steps of catalysis [121-123]. Celedon et al. showed that all TatC subunits in the TatBC complex are able to bind to substrate [117] but it was unclear if they are able to coordinate and act in concert or if they work on an individual basis. Based upon further studies, it appears as though each TatC in the complex is functionally active and that they cooperatively build a translocon when TatA binds and oligomerizes.
Once the signal peptide binds to the TatBC complex, TatA is recruited and oligomerizes to form the translocase [117]. However, there is some disagreement on how the oligomer is arranged in order to form the translocase as well as on the mechanism for how the translocation of substrate is actually accomplished. The rest of this section will describe the various structural data that has been obtained and the two main models for how the Tat complex is able to translocate folded substrates through the assembled machinery.
The structure of the TatABC oligomer is unknown because of the transient nature of the complex; but, EM studies have shown ring like structures when TatA molecules were
22 solubilized in detergent [116, 124-127]. A variety of sizes of the TatA oligomer were observed in these studies so solution nuclear magnetic resonance spectroscopy (NMR) was utilized to determine the structure of a small TatA oligomer created by varying the protein to detergent ratio. It was determined in this study that the TatA molecules form an interaction using their transmembrane segment [124]. Another NMR study utilized a crosslinking strategy to link the flexible TatA N-termini of multiple subunits in a high detergent concentration solution so that the formation of large oligomers was limited in order to facilitate the structural studies by NMR [128]. In this study, Zhang et al. saw the same TM interaction that was observed previously but also saw additional transient interactions between the amphipathic helix domains as well. Other studies have revealed alternative orientations of the amphipathic helix [101, 129, 130] including a transition from peripherally associated to membrane embedded in the cytoplasmic leaflet [131].
Transport via the Tat pathway occurs in a two-step mechanism after the signal peptide binds to the TatBC complex; the first step is the oligomerization of TatA where the speed of association depends upon the concentration of TatA in the membrane [117, 132, 133].
The second step is the actual transport of the substrate through the TatABC complex across the inner membrane. The two step mechanism and the order are corroborated by data showing that TatA is recruited to the TatBC complex before substrate can be translocated
[134]. Multiple studies have shown that the PMF is required for the oligomerization of
TatA but the requirement and role of the PMF after oligomerization and/or for transport continues to remain unclear [118, 129, 134, 135]. As mentioned previously, two
23 translocation mechanisms have been proposed; the positives and negatives of both are discussed below.
The pore model proposes that TatA oligomers are able to form an aqueous pore or channel much like the Sec system but with a larger size. This theory is supported by the EM studies already mentioned which observed TatA oligomers formed into particles that resembled pores [125]. EPR studies support this observation as they revealed that the TM segment of multiple TatA subunits can form a ring like structure [136]. However, the TM segments do not appear to be able to form an aqueous channel in the membrane. An aqueous channel lined with the amphipathic helices has been proposed but further studies are needed to confirm the presence of a transmembrane helical hairpin which would be required to form a pore in this manner [125, 137, 138]. Furthermore, forming a large pore would cause ion leakage if not properly sealed which would be difficult to achieve since Tat is able to export a diverse group of proteins with different folds and sizes.
The membrane destabilizing model proposes that TatA molecules can somehow act to weaken the local membrane environment in order to promote translocation. This could occur due to the local accumulation of the amphipathic helix (APH) and the interaction that it has with the cytoplasmic leaflet of the membrane [139, 140] or the fact that the TatA TM segment has been shown to be shorter than the width of the bilayer [124]. Problems with this hypothesis include the fact that a translocation intermediate has been shown to accumulate near TatA [111] and that distant residues on a folded substrate were able to contact TatA which leads to the logical conclusion that there must be a pore arrangement
24
[141]. Clearly, more work is needed in order to provide insight into the orientation of the
TatA oligomer and the mechanism of translocation.
1.5 The YidC pathway
Most proteins are inserted into the inner membrane via the SecYEG translocase but another pathway exists for a subset of proteins which are able to be inserted by YidC. Small phage proteins were the first to be recognized as substrates [142] but many other native E. coli proteins have been found to be YidC substrates [143]. Additionally, Welte et al. showed that some polytopic SRP targeted Sec dependent proteins can be inserted in vitro by utilizing YidC as an insertase [48]. Consistent with this observation, targeting of many proteins to YidC has been shown to occur by the SRP/FtsY pathway [144, 145].
Intriguingly, substrates that are only YidC dependent do not have a known targeting pathway; it is possible that targeting might occur based on the charge/polarity of the substrates as recent studies have proposed but this will be discussed in greater detail in
Chapter 2. YidC can also function as a chaperone for some Sec dependent substrates and acts to help partition substrates into the membrane, facilitate release from Sec, or act as an assembly site for oligomerization once substrates have been inserted into the membrane
[89, 146-151]. The YidC family of proteins is the topic of Chapter 2 where the history and function of the protein family will be discussed as well as recent structure and functional studies in addition to a proposed mechanism for membrane insertion by YidC based on two recent x-ray crystallography structures of the protein.
25
1.6 Figures
Figure 1.1 – Structure of the SRP-FtsY complex
Structure of the Escherichia coli SRP-FtsY complex with GMP-PCP bound. Ffh is teal and is shown in complex with the 4.5S RNA (tan and blue) and FtsY (orange). The N and
G domains of Ffh and FtsY interact in order to form the complex. The positions of the flexible linker and the C-terminal M domain are also indicated.
26
Figure 1.2 – Structure of trigger factor
Structure of the Escherichia coli Trigger Factor. The substrate binding domain is composed of the Head (purple), Arm 1 (green), and Arm 2 (yellow). The tail (cyan) is the ribosome binding domain.
27
Figure 1.3 – Structure of SecB
The tetrameric structure of SecB from E.coli; the four subunits in the dimer of dimers are shown in scarlet, grey, maize, and blue. (A) The front view of the dimeric interface with the peptide binding cleft (pink arrow) visible spanning from the top of the blue subunit to the bottom grey subunit and (B) Space filling model overlaid on the ribbon diagram of the same orientation. (C) Side view after rotation 90 degrees clockwise and (D) is the space filling model overlaid on the ribbon view.
28
Figure 1.4 – Structure of the Sec complex
(A) Front view of Secγβ from M. jannaschii viewed in the plane of the membrane (B)
Side view rotated 90 degrees counter clock-wise (C) View from the cytoplasm (D) View from the periplasm. The lateral gate consists of two helices (TM2b and TM7) shown in red. Secβ is shown in orange and Secγ (SecE homolog) is in yellow. The periplasmic plug helix formed by TM2a is shown in green. The two pseudo-symmetrical halves of the
Sec61a complex can be observed in C&D. Except for the plug and lateral gate regions,
TMs 1-5 are in blue and 6-10 are in cyan.
29
Figure 1.5 – Structure of SecA
Structure of E. coli SecA bound to ATP. The nucleotide binding domain (NBD) is shown in green bound to ATP. The intermolecular regulator of ATP hydrolysis (IRA) has two parts; IRA1 is shown in orange and IRA2 is in yellow. The helical scaffold domain (HSD; cyan) connects IRA2 to the helical wheel domain (HWD; red). The protein binding domain
(PBD) was disordered but the location is shown in blue.
30
Figure 1.6 – Structure of SecDF
Structure of SecDF from Thermus thermophilus viewed in the plane of the membrane. A pseudo-symmetrical structure can be observed consisting of TMs 1-6 (SecD region) colored in blue and TMs 7-12 (SecF region) colored in orange. There are two major periplasmic domains: P4 is shown in green while the P1 base is magenta and the P1 head is in red. Tsukazaki et al. observed a different orientation of P1 in higher resolution studies where the head was oriented above the base and toward P4 region. A catalytic mechanism was proposed based upon this observation [90]. 31
Figure 1.7 – Structure of SecA bound to SecYEG
Structure of SecA bound to SecYEG from Thermotoga Maritima. SecYEG is colored like
Figure 1.2 such that the lateral gate consists of two helices (TM2b and TM7) shown in red.
SecG is shown in orange and SecE is in yellow. For simplicity, the majority of SecA is shown in green. The IRA1 two-finger helix domain (pink) is shown intercalated in the cytoplasmic cavity of SecY. The PBD was not present in Figure 1.5 because it was disordered and is shown here in purple. 32
Figure 1.8 – Structure of the holo-translocon
Hypothetical structure of the holo-translocon containing SecYEG (blue with grey surface),
YidC (brown with green surface), and SecDF (purple) viewed in the plane of the membrane. SecDF has previously been crosslinked to the face of SecY that is made by
TMs 1-5 and is shown positioned on that face as well as in contact with YidC, which is located at the lateral gate of SecY. The P1 head of SecDF is in position to bind to the luminal part of substrates on the periplasmic side after being translocated across the membrane via SecYEG. 33
Figure 1.9 – Schematic of protein targeting in E. coli (continued)
34
Figure 1.9 Continued
Schematic of protein targeting in E. coli. (A) Ribosome nascent chains can directly bind to YidC during co-translational insertion. (B) Most proteins are targeted for co- translational insertion by the binding of SRP (pink and blue) to the signal peptide after it emerges from the ribosome exit tunnel followed by targeting to the membrane by FtsY
(green) and insertion via YidC or SecYEG. (C) In post-translational translocation, trigger factor (cyan) or SecA (red) can bind to the signal peptide and target the protein to chaperones to prevent folding and aggregation (SecB shown in yellow). The proteins are then targeted to the holo-translocon (SecYEG in blue with a grey surface, YidC in brown with a green surface, SecDF in purple, and the SecA dimer in red and orange) for translocation. After translocation, the signal peptide is removed by signal peptidase
(black).
35
CHAPTER 2
YidC/Alb3/Oxa1 Family of Insertases
2.0 Contributions
The manuscript for the review presented in this chapter was prepared by Seth Hennon.
Raunak Soman and Lu Zhu will also be listed on the journal article because of their critical reading of the manuscript and helpful suggestions.
2.1 Introduction
In all cells, membrane proteins play crucial roles in energy production, substrate transport, signaling, and metabolite exchange. They function as ATPases, photosynthetic complexes, chemosensors, and permeases. Membrane proteins make up 25 to 30 percent of the proteins in a cell and comprise over 50 percent of known drug targets. In order to assemble
36 proteins into the membrane lipid bilayer, translocation and insertion machineries are present in cells. The Sec translocase is the major translocase that inserts proteins into and across the endoplasmic reticulum (ER) of eukaryotic cells, the cytoplasmic membrane of bacterial and archaeal cells, and the thylakoid membrane of plants. When translocating proteins across the membrane, the Sec machinery does so in an unfolded state [53, 152].
Also operating within the eukaryotic cells, the Guided Entry of Tail-Anchored (GET) machinery delivers tail-anchored proteins to the ER membrane for membrane insertion by a post-translational mechanism [153]. A very different translocase called the twin arginine translocation (Tat) machinery functions to translocate folded proteins across the membrane
[95, 154]. The Tat pathway can translocate proteins across the cytoplasmic membrane in bacterial and archaeal cells and across the thylakoid membrane of chloroplasts.
This review will focus on the YidC/Oxa1/Alb3 family of proteins that operates in bacteria and certain eukaryotic organelles in order to facilitate membrane protein insertion. We will discuss the distribution and function of these insertases and highlight the recent structural work on these novel proteins which provides insight into how these proteins catalyze membrane protein insertion and protein assembly at a molecular level.
2.2 Distribution, topology and function of YidC/Oxa1/Alb3 proteins
In eukaryotes, there are two members of the YidC/Oxa1/Alb3 family that span the membrane five times [155, 156]: Oxa1 and Oxa2 (Cox18) in the mitochondrial inner membrane [157] and Alb3 and Alb4 in the chloroplast thylakoid membrane [158]. The
37 two paralogs in each system differ in their C-terminal region. The mitochondrial Oxa1 contains a long positively charged C-terminal region that constitutes the ribosomal binding domain, whereas the Oxa2 lacks this domain [157]. Oxa1 functions in co-translational insertion while Oxa2 functions in post-translational insertion [159-161]. In the chloroplast,
Alb3 also possesses a long C-terminal region but it recognizes a special targeting protein called SRP43, while Alb4 lacks this C-terminal domain [162]. In bacteria, the number of
YidC paralogs and the number of membrane spanning regions possessed by YidC can vary.
In Gram-positive bacteria, two paralogs (YidC1 and YidC2) can be found and they both span the membrane five times [163, 164]. YidC2 proteins possess a long C-terminal region that can directly bind to the ribosome while YidC1 lacks such a domain [165]. In Gram- negative bacteria, only one YidC is found and it contains an extra TM segment at the N- terminus of the protein and a large periplasmic domain between TM1 and TM2 [166]. In some organisms, there can be multiple paralogs. For example, plants can have four or more
YidC members: 2 paralogs in mitochondria and 2 or more in chloroplasts.
Interestingly, in B. subtilis, the gene expression of a YidC paralog is regulated by a novel post-translational mechanism. Chiba et al. [167] discovered that when SpoIIIJ (YidC1) is functional, the expression of YqiG (YidC2) is repressed, while the expression of YqiG is upregulated when SpoIIIJ is inactivated. Using a genetic approach, Chiba and coworkers found that SpoIIIJ inserts the MifM sensor protein that is encoded by the mifM gene which precedes the yqiG gene. When SpoIIIJ inserts MifM, the RNA hairpin loop following mifM masks the Shine Dalgarno sequence required for expression of yqiG. However, when
SpoIIIJ does not insert MifM, the membrane protein substrate undergoes translational 38 arrest. This arrest causes changes in the RNA structure leading to unfolding of the hairpin loop and expression of YqiG. Arrest of protein synthesis occurs by a multisite ribosomal stalling mechanism [168] that is different from the stalling mechanism that is observed with SecM, which controls the expression of the secA gene in E. coli [169].
The cellular function of the YidC family members is important for the assembly of energy transducing complexes [170]. The bacterial YidC and mitochondrial Oxa1 primarily function to insert and assemble the protein complexes involved with respiration [170-174] while the chloroplast Alb3 paralogs are necessary for photosynthesis and thylakoid biogenesis [175-178]. The enzyme activity of these insertases is remarkably conserved
[160, 179-182](Table 2.1). For example, the mitochondrial Oxa1 and bacterial YidC (with the C-terminal ribosome binding domain appended) can functionally substitute for each other [160, 182] and the chloroplast Arabidopsis thaliana Alb3 and Alb4 can replace the
E. coli YidC and function to insert proteins into the bacterial cytoplasmic membrane [180,
181]. Table 2.1 shows the substrates that have been identified for these insertases in mitochondria, chloroplast, and bacteria.
2.3 Oxa1 Family
The YidC/Oxa1/Alb3 family of proteins was discovered in the year 1994 by scientists in the mitochondrial field. The key findings were that a mutation in a new protein called Oxa1
(for oxidase assembly factor) affected cytochrome c oxidase biogenesis [171, 172] and the formation of the F1Fo ATP synthase [173]. Oxa1 was determined to be essential for
39 translocation of the N and C-terminal tails of Cox2 [183, 184]. Further studies showed
Oxa1 is a general machinery for the insertion of proteins from the matrix into the mitochondrial inner membrane. In Saccharomyces cerevisiae, Oxa1 mediates the insertion of all mitochondrially-encoded membrane proteins: Atp6, Atp9, Cox1, Cox2, Cox3 and
Cytb subunits [185].
In addition to inserting mitochondrial-encoded membrane proteins into the inner membrane, Oxa1 mediates the insertion of nuclear-encoded mitochondrial proteins. Hell et al. [186] showed that the nuclear encoded Oxa1 is involved in its own biogenesis. Oxa1 is synthesized in the cytosol and subsequently imported into the mitochondria where it assembles into the inner membrane. For this process, the Oxa1 machinery already residing in the inner membrane is critical for the translocation of the N-tail of the transported Oxa1, which is initially imported into the matrix [186]. Oxa1 also participates in the membrane biogenesis of the nuclear-encoded Mdl1, a 6 spanning protein of the mitochondrial inner membrane with its amino and carboxyl-termini in the matrix [187]. During Mdl1’s import into the mitochondria, the TIM22 complex and Oxa1 machineries cooperate to insert the protein into the membrane. During this process, the TIM22 translocase engages the Mdl1 membrane protein at the inner membrane and integrates the amino-terminal and carboxyl- terminal domains into the membrane by a stop transfer mechanism. Strikingly, the central region of Mdl1 is completely translocated into the mitochondrial matrix and then re- inserted into the inner membrane by Oxa1.
40
A distinctive feature of Oxa1 is that it contains a long positively charged carboxyl-terminal tail exposed to the mitochondrial matrix. This domain enables Oxa1 to be permanently bound to the matrix-localized ribosome [161, 188]. The ribosome-bound Oxa1 is in proximity to the large ribosomal proteins Mrp20 [188] and Mrp40 [189], which are homologous to bacterial ribosomal proteins L23 and L34 and known to be located at the polypeptide exit site within the large ribosome subunit. Interestingly, deletions in the mitochondrial Mrb20 and Mrp40 proteins, which do not affect ribosomal protein synthesis, have profound effects on the assembly of the oxidative phosphorylation complexes in the inner membrane [189, 190]. These and other studies highlight the importance of Oxa1- ribosome complexes for the assembly of respiratory chain complexes [189, 191].
Oxa2 (also called Cox18), the other Oxa1 paralog, is important for the assembly of cytochrome c oxidase and specifically for the biogenesis of Cox2, which is encoded in the mitochondrial genome and transfers the electron from cytochrome c to Cox1. Oxa2 facilitates the post-translational translocation of the C-terminal domain of Cox2 both in
Saccharomyces cerevisiae and Neurospora crassa [157, 192, 193]. In the inner membrane, Oxa2 forms a complex with Pnt1 and Mss2 in which they cooperate together in the biogenesis of Cox2 [192]. Interestingly, the translocation and assembly of Cox2 can occur in the absence of Oxa2; but, this requires Oxa1 to be overexpressed and the protein Yme1 (an ATP-dependent mitochondrial protease residing in the inner membrane that is also involved in protein folding) [194] to act as a chaperone to help fold and assemble Cox2, leading to a functional cytochrome c oxidase complex [195].
41
2.4 Alb3 Family
In chloroplasts, transposon tagging in Arabidopsis thaliana revealed that the ALB3 gene encodes a protein crucial for photosynthesis. Knockouts of the gene ALBINO3 (encoding
Alb3) led to a photosynthesis defective mutant with an Albino appearance [175], giving rise to its name Alb3.
Alb3 is critical for the post-translational insertion of the LHCP (Light-harvesting chlorophyll-binding protein) into thylakoids [196]. LHCP is a nuclear-encoded protein, which is synthesized in a precursor form containing a stroma transit peptide. It is imported from the cytosol into the chloroplast via the TOC/TIC complex located in the envelope.
After import into the stroma and removal of the stroma targeting peptide, the LHCP forms a transit complex with the chloroplast SRP (cpSRP) [197]. cpSRP is unusual among SRP’s in that it lacks RNA and contains a novel protein subunit cpSRP43 [197] in addition to cpSRP54 in Arabidopsis thaliana. It is also distinctive because it is involved in both co- and post-translational targeting. The LHCP-cpSRP54-cpSRP43 transit complex is then directed to the SRP receptor (cpFtsY) at the thylakoid membrane [198], which requires
GTP to be bound to both cpSRP54 and cpFtsY. Interestingly, the two GTPase proteins interact efficiently which enables the chloroplast SRP pathway to bypass the SRP RNA requirement seen in other SRP targeting systems [199]. At the membrane, Alb3 recognizes, by utilizing its C-terminal domain, the cpSRP43 protein within the transit complex [200].
Additionally, part of the Alb3 membrane-embedded domain is also involved in this cpSRP43 recognition step [201]. Following posttranslational targeting of LHCP to the
42 membrane by cpSRP, Alb3 inserts LHCP into the thylakoid membrane in a SecY independent manner [202]. We should note that some chloroplasts contain SRP-RNA like a majority of SRP complexes as well as cpSRP43 [203].
In addition to acting independently, Alb3 may function cooperatively with the thylakoid cpSecYE translocase to insert proteins co-translationally into the thylakoid membrane just as the bacterial YidC can function with the SecYEG translocase. Co-purification and cross-linking studies have established that Alb3 interacts with cpSecY [204]. Potential
Alb3 substrates that use the co-translation insertion pathway are D1, D2, CP43, PS1-A, and ATPase subunit CF0III. These proteins interact with Alb3 as demonstrated using the split-ubiquitin system [205]. However, so far, it has not been shown that these proteins or any other membrane protein strictly requires both Alb3 and cpSecY for insertion into the thylakoid membrane of plants.
The other chloroplast paralog, Alb4, has a weaker phenotype in comparison to Alb3 disruption when a knockout mutation is made [158]. Nevertheless, Alb4 is involved in thylakoid biogenesis since the ultrastructure of chloroplasts shows alterations in the knockout [158]. Characterization of the ALB4 knockout in Arabidopsis thaliana reveals that Alb4 is involved in the assembly of CF1CFo-ATP synthase, since the alb4 null mutant has reduced amounts of the ATPase subunits and Alb4 interacts with CF0II and CF1 subunits of CF1Fo-ATP synthase [180].
Besides playing important roles in plants, Alb3 homologs have also been shown to perform critical functions in Algae. In Chlamydomonas reinhardtii, both Alb3 homologs, Alb3.1 43 and Alb3.2, are crucial for photosynthesis [176, 206]. The Alb3.2 paralog is important for the assembly of Photosystem II (PSII) in the thylakoid membrane since depletion of Alb3.2 leads to a reduced amount of PSI and PSII [206]. Alb3.2 has been proposed to function as a chaperone during the assembly of the PSII and, as predicted, was found to interact with the core D1 subunit of PSII [206].
2.5 YidC Family
Six years after establishing that the yidC gene was present in the genome of both Gram- negative and Gram-positive bacteria [172], the role of YidC in membrane protein biogenesis was discovered. Scotti et al. [89] showed that YidC can be cross-linked to newly synthesized FtsQ during membrane protein insertion and that YidC co-purifies with
SecYEG and SecDF(YajC). Furthermore, YidC was discovered to be essential for the growth of E. coli and to promote the insertion of the Sec-independent M13 phage procoat protein [142] which was previously thought to insert by an unassisted mechanism.
The number of substrates that are inserted by the bacterial YidC is believed to be much larger than the number of substrates inserted by Oxa1 in mitochondria and Alb3 in chloroplast, partly because YidC can function both independently and cooperatively with the Sec machinery. As an independent insertase, YidC has been shown to insert Foc
(subunit c of F1Fo ATP synthase), MscL, the phage proteins M13 procoat and Pf3 coat, and
TssL, a tail anchored membrane protein [142, 144, 150, 151, 207-209]. Interestingly, YidC is able to insert the tail anchored protein TssL in bacteria on its own even though a
44 dedicated translocase that is not present in bacteria is employed in eukaryotes [153].
Together with the Sec translocase, YidC promotes insertion of the F1Fo ATP synthase subunits Foa and Fob, NuoK (NADH-quinone oxidoreductase subunit K), and CyoA
(subunit 2 of the cytochrome bo oxidase) [210-215]. Interestingly, in a recent study, both
SecY and YidC appear to be involved in the assembly of Type IV prepilin in cyanobacterium Synechocystis PCC 6803 [216].
YidC performs its Sec-dependent function as part of a holoenzyme; a super-complex containing YidC, the SecYEG channel, and SecDF(YajC) [89, 217]. In the holoenzyme,
YidC is in close proximity to the SecYEG translocation channel [218] where it can interact with membrane protein substrates as they exit the channel through the lateral gate. The
SecY lateral gate (TM2b, TM3, TM7 and TM8) region is where membrane protein substrates exit [53] and can be cross-linked to YidC when photoprobes are introduced into the lateral gate region [218]. Additional residues in this or other regions may also be important for the YidC and SecY interaction. Recently, Li et al. discovered Gly 355 in
TM2 and Met 471 in TM4 of the E. coli YidC to be important for YidC-SecY interaction using a synthetic lethal screen [219]. It has been hypothesized that the interaction of YidC with SecYEG is facilitated by SecDF(YajC) [88]. However, crosslinking between the
SecY lateral gate and YidC is seen even in the absence of SecDF(YajC) [218]. Moreover, there are populations of the Sec translocon in the membrane that contain only YidC-SecYE with no SecDF(YajC) present [220], showing that YidC can interact directly with the
SecYEG complex. In the holo-complex, YidC functions to promote the removal of transmembrane segments of inserting membrane proteins from the Sec channel [221], 45 facilitate their integration into the lipid bilayer [222], and acts as an assembly site for multi- spanning membrane proteins [223]. As part of its folding function, YidC plays a direct role in the helix-helix packing of membrane proteins [224, 225]. This explains why YidC is required for the folding, but not insertion, of LacY [224, 225] and MalF [149] and is required for the assembly of the maltose transporter MalFGK [149].
Structure-function studies revealed that the conserved 5 TM segment region of the E. coli
YidC is critical for its function in membrane protein insertion [226]. However, the large periplasmic region of the E. coli YidC is not essential for the insertase function. X-ray crystallographic studies showed that the periplasmic domain has a -super sandwich fold which is found in proteins that bind sugars and may play some role in the folding of newly inserted proteins [227, 228]. One cryo-electron microscopy study revealed that YidC may function as a channel since YidC was able to form a homodimer [229]. The YidC dimer was found to sit at the exit channel of the ribosome near the predicted L23 protein. Another cryo-EM study showed that YidC was bound to the ribosome as a monomer [230].
Although it is not entirely clear what the native state is in vivo, data suggest that the functional unit of YidC is a monomer [231].
2.6 Recent Insights
In the year 2014, Kumazaki et al. [232] reported the first crystal structure of YidC; a landmark contribution to the membrane biology field. The X-ray structure of YidC was solved from Bacillus halodurans at 2.4Å resolution. Remarkably, the structure showed that 46 the five membrane-embedded TM segments of YidC contain a hydrophilic cavity within the inner leaflet of the membrane, which is closed from the extracellular side of the membrane (Fig. 2.1A and B). Interestingly, the hydrophilic groove is open to the cytoplasm and the lipids of the membrane. At the entrance of the hydrophilic cavity on the cytoplasmic side of the membrane is a helical hairpin-like domain that may be involved in the initial recruitment of YidC substrates. Intriguingly, the hydrophilic groove contains the strictly conserved positively charged arginine residue R73, which is essential for YidC1 to promote growth in a Bacillus subtilis YidC1 depletion strain and for YidC1 to promote membrane insertion of the single-span MifM protein [232]. The arginine R73 is proposed to attract the negatively charged residues in the MifM translocated region because substitution of the three negatively charged MifM residues to neutral residues prevented translocation.
In addition to the novel hydrophilic groove, the structure also provided some intriguing information about the dynamics of YidC. Kumazaki et al. [232] isolated two distinct forms of YidC (3WO6 and 3WO7) that had varied positions of the C1 loop between TMs 1&2.
The cytoplasmic halves of the TM segments, which make up the groove, also showed positional variance when the two structures are compared. This was corroborated with high crystallographic B-factors in these regions as well as by molecular dynamics simulations. Two other recent studies also pointed to the dynamic and flexible nature of this region. Flexibility was observed in membrane vesicles by utilizing intramolecular chemical cross-linking of the E. coli YidC [233]. There was cross-linking between the
47 conserved core transmembrane segments with a wide range of different sized crosslinking agents, which suggests that the flexibility occurs in vivo.
Additionally, Wickles et al. performed protein evolutionary co-variation analysis, lipid- versus-protein-exposure, and molecular dynamics simulations with the E. coli YidC and were able to determine a model that closely matched the crystallographic structure [234].
The helical hairpin between TM2 and TM3 of the E. coli YidC (TM1 and TM2 of B. halodurans) was also predicted by this study along with the flexibility of this region.
Thinning of the membrane was observed during molecular dynamics simulations which could provide some insight into the translocation mechanism. Given the fact that the hydrophilic cavity spans only the inner leaflet of the membrane, thinning of the membrane could help decrease the energy required to finish translocation across the outer leaflet.
The crystal structure of the Gram-negative E.coli YidC was also reported in late 2014 to
3.2 Å by Kumazaki et al. [235]. Like the Gram-positive B. halodurans structure, the five conserved core TMs are tightly packed in the periplasmic half of the membrane and spread out in the cytoplasmic half (Fig. 2.1C and D). The E. coli YidC also contains a hydrophilic groove which is open to the cytoplasm and membrane. The C1 region forms a helical hairpin that is flexible based on crystallographic B-factors but the arrangement is rotated when compared to the B. halodurans structure. Previously crystallized, the P1 domain was also present and the large cleft was shown to be oriented away from the membrane which could allow it to bind substrate proteins or molecules as previously proposed [227, 228].
48
Interestingly, the previously reported substrate contacts [232, 236, 237] are located in the groove between TM3 and TM5 of the E. coli YidC and are found on the exterior region of the TM as well as in the hydrophilic groove [235] (Fig. 2.2). It is likely that residues facing the center of the hydrophilic cavity are involved with binding of the hydrophilic translocated region of the substrate while those facing the membrane help insert and laterally integrate the hydrophobic segment of the substrate into the membrane. In addition to reporting the structure, the authors also investigated the importance of the conserved arginine residue using a complementation assay. Hydrophilic groove mutations were able to rescue growth at 37°C except for T362A, which was previously shown to be inactive
[235]. However, two cold sensitive mutants (R366A and R366M) were lethal at 20°C which the authors propose illustrates an important role for the conserved positive charge under certain conditions.
The importance of the evolutionarily conserved positive charge for the function of the YidC family was also probed by Chen et al. in a recent publication [238]. The positive charge was determined to be essential for the Gram-positive S. mutans YidC2 much like the results observed in B. subtilis [232]. For E. coli and chloroplast homologs, the charge was not essential for function in an E. coli YidC depletion strain. Mitochondrial Oxa1 had been studied previously and the conserved charge was determined to be important for activity
[239]. Variation was observed for the importance of the charge when different substrates were analyzed which suggests that there are different insertion requirements based on the characteristics of the substrate [238]. Additionally, by making deletions and mutations, the
49
C-terminal half of the conserved helical hairpin was found to be important for the activity of E. coli YidC [238].
Insights into how YidC is able to perform multiple roles via interactions with a plethora of
Sec substrate proteins have been made recently on various fronts. By isolating a stable holo-translocon (HTL) from E. coli, Schulze et al. [217] were able to determine the components present, the ratios, and some interactions. They observed two versions of the
HTL: the first contained one copy of SecYEG, YidC, SecDF and YajC that functions in membrane protein insertion, while the second was a SecYEG dimer which functions in protein export. Interactions were also determined through DSP chemical cross-linking and were formed between SecD/YidC as well as between SecY, E, and G. The functional state of YidC not in the HTL is still unknown but it appears as though YidC is a monomer in membranes [231].
The crystal structures of YidC also provide insight into the mechanism by which YidC functions as a translocase [232] (Fig. 2.3). In the initial step, a YidC substrate binds to the membrane (Fig. 2.3A). Subsequently, the N-tail region of the substrate is recruited to the
YidC hydrophilic groove facilitated (Fig. 2.3B and C), in some cases, by the interaction of the negatively charged residues in the N-tail with the positively charged arginine in the aqueous cavity. The N-tail is then released from the groove and crosses the outer leaflet of the membrane. This latter step is believed to be catalyzed by the action of the electrical potential (positive side in the periplasmic space) acting on the negative charged residues in the translocated region of the substrate and by hydrophobic interactions between the
50 substrate TM segment and YidC (Fig. 2.3D). The TM segment of the substrate most likely moves along a greasy slide formed by TM3 and TM5 of the E. coli YidC in order to form a transmembrane configuration. Presumably, a similar mechanism would be used for the
YidC substrate TssL; the only difference would be that the C-terminus is translocated.
Two recent studies have started to lay the framework for determining the characteristics of substrates that govern the insertion machinery used for integration into the membrane. Zhu et al. [240] proposed that the charge composition of the translocated periplasmic domain or the transmembrane segment could determine which pathway model single-span membrane proteins used for insertion. Hydrophobic TM segments could insert independently and decreasing the hydrophobicity lead to a dependence on YidC, SecYEG or both. A positive charge in a translocated loop required YidC/SecYEG while a negative charge in the loop only required YidC, which corresponds nicely to the mechanism proposed by Kumazaki et al. [232] A more in depth study was able to predictably alter the requirements of a translocated loop in the M13 procoat protein by changing the polarity and charge of the loop [241]. Lowering the polarity and number of charges enabled the substrate to translocate via an independent mechanism while increasing the polarity caused
YidC and YidC/Sec to be required. By increasing the hydrophobicity of the transmembrane segments, Soman et al. [241] were able to lower the translocase requirement.
These two recent studies build upon the framework laid by earlier studies in the mitochondria as well as E. coli. In the year 2004, Herrmann et al. determined that Oxa1
51 was important for the insertion of proteins with highly charged domains and, more specifically, for proteins with negative charges in the translocation region [242]. Negative charges in the TMs of NuoK were shown to be determinants for YidC and mutating these residues to lysines caused the protein to be only Sec dependent for insertion [215]. The authors also proposed that YidC’s conserved role for respiratory proteins could be due to the fact that respiratory proteins often have negative charges that are essential for function.
By utilizing a genome scale approach, Gray et al. discovered that substrates with charge unbalanced TM segments were significantly more likely to depend on YidC for insertion
[243]. However, many of the YidC-dependent proteins identified did not have unbalanced
TM segments and it was proposed that other features also contribute to a protein’s requirement for YidC.
2.7 Conclusion
The YidC/Oxa1/Alb3 proteins are a novel group of insertases that function to insert, fold, and assemble proteins into the lipid bilayer. They are particularly important for the assembly of energy transducing complexes vital for cellular respiration and photosynthesis.
Different than the Sec translocases, the YidC/Oxa1/Alb3 insertases can translocate only short hydrophilic regions of membrane proteins across the membrane and seem to play a more prominent role in the folding of membrane protein substrates.
The emerging structural data suggest that the YidC/Oxa1/Alb3 family of proteins do not function as channels. Rather, they promote the transport of hydrophilic regions of
52 membrane proteins by possessing a hydrophilic groove within the inner leaflet of the membrane. This novel structural feature is not found to date in other translocases. Recent studies suggest that the hydrophilic cavity most likely recruits the hydrophilic region of the substrate, thus allowing it to transfer halfway across the membrane. It is then released from the groove in a manner not completely understood and crosses the outer leaflet of the membrane. In light of the recent advancements, this is an exciting time for the membrane protein field and we are poised to finally understand the intricacies of how proteins are inserted into the membrane and how YidC and Sec are able to work cooperatively in order to accomplish this complex process.
53
2.8 Tables
YidC family Independent Function Foc F1Fo ATP synthase subunit MscL Large-conductance mechanosensitive channel TssL Part of the type VI secretion system M13 procoat* M13 bacteriophage major coat protein Pf3 coat* Pf3 bacteriophage major coat protein
Cooperative with Sec Function Foa F1Fo ATP synthase subunit Fob F1Fo ATP synthase subunit NuoK NADH-quinone oxidoreductase subunit CyoA Cytochrome bo(3) ubiquinol oxidase subunit
Alb3 family Alb3 Function LHCP Organizes chlorophyll into chlorophyll protein complex CII
Alb4 Function CFoII CF1CFo ATP synthase subunit
CF1 CF1CFo ATP synthase subunit
Oxa1 family Mitochondrial Encoded Function Atp6 F1Fo ATP synthase subunit Atp9 F1Fo ATP synthase subunit Cox1 Cytochrome c oxidase subunit Cox2Ŧ Cytochrome c oxidase subunit Cox3 Cytochrome c oxidase subunit Cytb Cytochrome bc1 subunit
Nuclear Encoded Function Oxa1 Existing Oxa1 is required for new Oxa1 insertion Mdl1 ATP dependent permease
Table 2.1 – Substrates of the YidC homologs (continued)
54
Table 2.1 Continued
Known substrates of each YidC homolog. The YidC family is used mainly to insert proteins associated with respiration. *non-native protein Ŧaided by Oxa2
55
2.9 Figures
Figure 2.1 – Crystal structures of YidC
Ribbon representations of the B. subtilis (A and B) and E. coli (C and D) YidC viewed transversely through the membrane with the E. coli P1 domain omitted for simplicity. The conserved arginine residue is shown in purple. The structures are very similar except for the orientation of the C1 domain, shown in cyan.
56
Figure 2.2 – Substrate contacts of YidC
(A) The structure of the E. coli YidC with the previously determined Sec-independent substrate contacts shown in purple. (B) Close up view of the area boxed in A with the residues labeled. I478 and Y517 were determined in the B. halodurans YidC.
57
Figure 2.3 – Model for the insertion of a single-span membrane protein
(A) Substrate is peripherally associated with the membrane and YidC is in the resting state.
(B) The hydrophilic region to be translocated becomes associated with the hydrophilic groove of YidC, possibly via electrostatic interaction. (C) A variety of factors are believed to facilitate the release of substrate including: the proton motive force, hydrophobic interactions between YidC and substrate, and possibly thinning of the membrane by YidC.
(D) After successful translocation and insertion via a greasy slide, YidC returns to the resting state. This model was adapted from Kumazaki et al. [232].
58
CHAPTER 3
Cross-linking-based flexibility and proximity relationships between the
TM segments of the Escherichia coli YidC
3.1 Introduction
The YidC/Alb/Oxa1 family members promote the insertion, folding and assembly of proteins into the inner (cytoplasmic) membrane in bacteria, as well as the insertion of proteins into the thylakoid membrane of chloroplasts and the inner membrane of mitochondria, respectively [244]. Bacterial YidC can function independently, as well as cooperatively, with the SecYEG translocase to promote membrane insertion and folding
[143, 245, 246]. As an independent insertase, the E. coli YidC has been shown to promote the membrane insertion of the non-native M13 phage procoat [142, 247] and Pf3 coat proteins [207], along with the endogenous substrates subunit c of F1Fo ATPase [151, 208,
214, 248], MscL[144],TssL [209] and the N-terminal domain of CyoA [211, 212].
59
Working with the SecYEG machinery, YidC catalyzes the insertion of subunit a and b of
F1Fo ATPase [214], NuoK [215], and TatC [221].
In all YidC homologs, there is a conserved core region that contains 5 TM segments [155,
156], which forms the catalytic insertase domain facilitating membrane insertion and folding [226]. The most variable regions are the N-terminal domain and C-terminal cytoplasmic tail [155] possessing, in some cases, a long positively charged segment that functions in ribosome binding [161, 188]. In Gram-negative bacteria, YidC contains an extra TM segment at the very amino-terminus and a large translocated region [166] that is known to bind SecDF [249]. X-ray crystallography analysis revealed that the periplasmic domain possesses a super- sandwich fold [227, 228], however the function of this non- essential region in membrane protein biogenesis is not clear [226]. More recently, cryo- electron microscopy studies of YidC ribosome/nascent chain complexes revealed that
YidC binds its substrate either as a dimer [229] or monomer [230]. While YidC obviously can form a dimer under certain conditions, fluorescence correlation spectroscopy experiments show that the functional unit of YidC in the membrane is a monomer [231].
The YidC substrate contacts have been mapped using disulfide crosslinking to Pf3 coat.
YidC is able to bind substrates by recognizing the substrate TM region during membrane insertion. The YidC substrate contact region includes TM3 [236, 250], which may serve as the principal recognition site, as well as TM1, TM4, and TM5 [237]. Contacts are observed within the YidC/substrate hydrophobic regions, which span the membrane [237].
60
The data support a model where YidC functions mainly as a hydrophobic platform to bind the hydrophobic segments of the substrate during membrane protein insertion [226, 251].
In order to better understand the structural features of the membrane-embedded core region of YidC, we have used Cys-based crosslinking studies pioneered by the Kaback lab [252-
254] to determine the proximity relationships between TM segments. Our data show that
YidC is a dynamic membrane protein on the cytoplasmic side of the membrane and support a model where the YidC TM3, TM4, TM5, and TM6 each have a face in the vicinity of and oriented toward TM2 near the middle of the membrane.
3.2 Results
Topology and crosslinking approach
The E. coli YidC spans the inner membrane six times with its N- and C-termini in the cytoplasm [166] and contains a membrane-associated region that is predicted to link the large periplasmic folded region to TM2 (Fig. 3.1A). To help define the TM arrangement within the conserved C-terminal core of YidC, we used a Cys crosslinking method to determine the proximity of engineered Cys residues located in different TM regions. A Cys was introduced into TM2 or its flanking region and a second Cys was then introduced in
TM3, TM4, TM5 or TM6. To examine crosslinking between the two Cys residues, we introduced tandem Thrombin protease sites into the cytoplasmic loop C1 in between TM2 and TM3 (Fig. 3.1A). Without the addition of crosslinking reagents, the thrombin protease completely cleaved YidC. If crosslinking occurs, then the two segments of YidC remain
61 together as a full-length protein after Thrombin cleavage. In our crosslinking studies, we have defined four categories based on the amount of crosslinking observed during Western blot analysis: strong (above 50%), moderate (21-50%), weak (6-20%) or none (5% or less).
Table 3.2 summarizes all of our crosslinking results in this paper, which were performed in duplicate.
Crosslinking in the membrane was initiated by adding iodine or two homo-bifunctional crosslinking reagents: 1,2-MTS and 1,6-MTS (see Fig. 3.1B for structure of reagents).
Iodine catalyzes the formation of a disulfide bond (~2 Å) whereas 1,2-MTS has a flexible spanner length of ~5 Å and 1,6-MTS flexibly spans ~11 Å between the two MTS groups.
Maleimide crosslinkers are used along the cytoplasmic and periplasmic border regions because they are known to react in a water-exposed environment more so than in the non- polar membrane interior. The maleimide reagents were o-PDM and p-PDM, (see Fig. 3.1B for chemical structure), which had fixed 6 Å and 11 Å spanner lengths, respectively and
BMH, which has a flexible 16 Å spanner length.
Despite treatment with various crosslinking reagents, Figure 3.1C shows that a Cys-less control was entirely cleaved by thrombin. Formation of a disulfide bond with iodine or crosslinking with a bifunctional reagent would result in the two fragments remaining together. The protein runs at a similar position as the full-length parent YidC protein on a
SDS-PAGE gel, as seen with the T373C/E415C mutant (Fig. 3.1C). As a negative control, we examined crosslinking with Y370C/R447C, a mutant containing cysteines on opposite sides of the membrane.
62
We confirmed that the Cys-less YidC (and all pairs of Cys mutants) was functional by performing a complementation assay (Fig. 3.1D; and see Table 3.1). We assayed the mutant for activity using the YidC depletion strain JS7131 [226] in which YidC expression is controlled by the araBAD promoter [142]. While growth in the presence of glucose represses the endogenous YidC expression and leads to depletion of YidC in growing cells, growth in arabinose allows for YidC expression. The overnight culture of JS7131 bearing
YidC mutants was back-diluted 1:100 and grown in LB media for 2h. The culture was then serially diluted and spotted on the LB agar plate containing arabinose or glucose plus
IPTG. Figure 3.1D shows that the Cys-less, T373C/E415C, and Y370C/R447C YidC fully complement the YidC-depletion strain. As a negative control, we confirmed that JS7131 containing the pEH1 empty vector could not support cell growth on a glucose plus IPTG plate because YidC is required for cell viability.
Crosslinking of paired Cys residues in helices 2 and 3
We defined the proximity of Cys pairs located in TM2 and TM3 or their flanking regions.
An Y370C/E415C Cys pair was used to determine the distance between the cytoplasmic ends of TM 2&3 (Fig. 3.2A). The TM segments appear to be very dynamic because iodine
(forming a disulfide) and all of the maleimide crosslinkers ranging from 6 Å to 16 Å were able to crosslink at the cytoplasmic border. We switched to the MTS crosslinkers to analyze
Cys pairs in the membrane since they are considerably more reactive than maleimides and so they can react with lipid exposed Cys [[255] also see Fig. 3.9 where we compare the efficiency of crosslinking of all six reagents for a Cys pair in hydrophilic (Y370C/E415C) 63 and in hydrophobic (I361C/L427C) environments)]. Figure 3.2B shows that YidC becomes less flexible moving toward the membrane interior from the cytoplasmic border because 1,2-MTS and 1,6-MTS were still able to crosslink the R366C/P419C mutant but no crosslinking was observed with iodine. The trend was the same for the majority of the
361C mutants at various positions; 1,2 and 1,6-MTS were able to crosslink I361C/C423,
I361C/F424C, and I361C/L427C. Only 1,2-MTS crosslinking was seen for I361C/P419C and I361C/P425C, which suggests that the distances between the two helices are constricting toward the membrane interior. No crosslinking was observed moving toward the periplasmic side of the TMs, as evidenced by I358C/L427C. However, crosslinking, as well as flexibility, returned with the I358C/M430C mutant, which was evident in the 1,2 and 1,6-MTS crosslinks. Interestingly, crosslinking was also observed between the membrane associated periplasmic loop before TM2 and the periplasmic portion of TM3
(Fig. 3.2B). The F350C/M430C mutant was able to form a weak disulfide using iodine along with strong crosslinks using the 1,2-MTS reagent and moderate crosslinking with
1,6-MTS. A similar pattern was seen with F350C/M439C, except that the iodine crosslinking faded away. Interactions between the C1 loop and TM3 become more constricted at the periplasmic end of TM3; both iodine and 1,2-MTS were able to crosslink
W346C/A435C and W346C/S443C. For W346C/M439C, 1,2-MTS was the only reagent that was able to crosslink moderately. However, no crosslinking was observed for
W346C/R447C (Fig. 3.2B) and F350C/R447C (Fig. 3.2C) and weak crosslinking was observed with BMH for W346C/L446C (Fig. 3.2C).
64
Crosslinking of paired Cys residues in helices 2 and 4
Flexibility was also seen at the cytoplasmic border of TM 2&4, as iodine and the maleimide crosslinkers all showed crosslinking between the Y370C/S482C Cys pair (Fig. 3.3A).
However, no crosslinking is seen toward the interior of the membrane (Fig. 3.3B) with
R366C/F476C until the I361C/K480C mutant, which reacted with 1,2-MTS along with the
I361C/F476C mutant. This latter mutant, like three additional TM 2&4 mutants that were studied, was only partially active in complementation studies using JS7131 (Table 3.1).
The crosslinking then continues for the I361C/G472C and I358C/G472C mutants until
F350C/Y465C is reached and iodine is able to form a strong disulfide along with weak 1,2-
MTS crosslinking. Located beyond the TM borders and within the periplasmic region (Fig.
3.3C), the W346C/Q461C mutant was able to crosslink with p-PDM and BMH. This suggests that part of the membrane-associated region of P1 is within 11 Å of the periplasmic portion of TM4.
Crosslinking of paired Cys residues in helices 2 and 5
A slightly less dynamic cytoplasmic border (Fig. 3.4A) was observed using the
Y370C/K493C mutant; the maleimides were able to crosslink but a disulfide bond was not seen with iodine. Moderate-to-weak crosslinking with 1,2 MTS was seen throughout most of the tested TM Cys pairs between the TM2&5 segments (Fig. 3.4B). This pattern was seen with R366C/P499C, I361C/M495C, I361C/P499C, and weakly with I361/T503C in addition to weak 1,6-MTS crosslinking. Weak 1,2-MTS cross-linking was observed with
I358C/T503C; the I358C/507C mutant also showed crosslinking with the 1,2-MTS 65 reagent, which confirms the relationship between these two areas of the protein. Similarly, the N-terminal portion of the membrane-associated periplasmic segment before TM2 is predicted to be farther away from the periplasmic side of TM5, as we see crosslinking of the I336C/S511C mutant with BMH, which has the 16Å spanner length (Fig. 3.4C) and weakly with p-PDM. No other crosslinking was seen on the periplasmic side of the transmembrane segment, as identified with the F350C/L507C and W346C/P510C mutants.
Crosslinking of paired Cys residues in helices 2 and 6
Flexibility was once again seen at the cytoplasmic border of helices 2 and 6 using the Cys pair Y370C/E536C as cross-linking was observed for all of the maleimide crosslinkers
(Fig. 3.5A). Upon moving farther into the membrane, I361C/Q529C was able to crosslink using the 1,2-MTS reagent as was I361C/Q527C, but with less efficiency (Fig. 3.5B).
These two mutants would be on opposite sides of the helix, assuming the transmembrane segments adopt a helical conformation, which could explain the difference in reactivity.
Flexibility and strong crosslinking are again seen with I361C/I525C using the 1,2 and 1,6-
MTS crosslinkers. The remaining mutants located within the periplasmic TM region do not crosslink, as shown with the I358C/V523C, G355/S520C, and W354C/V519C mutants except for the F350C/L515C, which shows weak 1,2-MTS crosslinking. The periplasmic border of helix 6 shows some crosslinking, albeit weak crosslinking, to the membrane- associated helix before TM2, as evidenced by the weak crosslinking by iodine and moderate crosslinking by 1,2-MTS with the W346C/L515C mutant.
66
3.3 Discussion
In this work, we exploited Cys crosslinking to begin to define the proximity relationships and flexibility among the TM segments in the evolutionary conserved core domain of
YidC. Crosslinking is a very powerful approach to define the general structural features of a membrane protein, including interhelical distances and helix tilting [252-254, 256,
257]. In our studies, various crosslinking reagents with different spanner lengths were used to probe a wide range of distances between Cys pairs within YidC’s TM segments.
Crosslinking between Cys pairs can provide information regarding average distances, as well as distances that only occur transiently within the YidC protein due to natural protein dynamics and breathing motions. Moreover, crosslinking with multiple reagents can provide insight into the conformational changes that YidC undergoes. Another advantage of our study includes the utilization of membrane vesicles for our crosslinking experiments instead of protein solubilized in detergent. The membrane vesicles allow the protein to be probed in its natural membrane under conditions where protein synthesis and membrane protein insertion are not occurring. However, a complication in studying cellular membranes is that multiple states of the protein (i.e. YidC by itself or YidC bound to
SecYEG or substrate) may occur, which may make interpretation more challenging.
After crosslinking and isolating membrane pellets, we used the detergent DDM to extract
YidC from the membrane. The YidC solubilized in detergent (after the 18,000 x g spin) is considered to be a protein that is folded properly while the unextracted protein is aggregated [258, 259]. Under our experimental conditions using BL21 pEH1 YidC cells,
67 most of the YidC is extracted using DDM and found in the supernatant after centrifugation in order to remove the unsolubilized and aggregated protein (Fig. 3.7; ~90%, left panel, lane 1). No additional YidC is solubilized and found in the supernatant by repeating the
DDM extraction of the membranes (Fig. 3.7, lane 2); however, there is a small amount of
YidC that remains in the pellet (~10%, left panel, lane 3). Similarly, we saw excellent detergent extraction using the Walker strain C41 (DE3) that allows production of many membrane proteins to high levels [260]. In C41 pEH1YidC, all of the YidC was found in the first DDM extraction (Fig. 3.7, right panel, lane 1) with no visible YidC remaining in the membrane fraction (right panel, lane 3). Also, no detectable YidC is seen in the supernatant from the second DDM extraction with C41 (lane 2). For additional confirmation, a representative handful of double cysteine mutants were expressed in C41, allowing for optimal insertion and folding of YidC [261]. Fig. 3.8 shows that similar crosslinking results were observed in these studies when compared to BL21. This provides us increased confidence that the proximity relationships and dynamics we observed in our crosslinking studies using BL21 pEH1YidC reflect the properly folded YidC and not aggregated material.
Cys crosslinking between TM segments indicates that TM3, TM4, TM5 and TM6 come in close contact to TM2 near the membrane interior. For instance, we found that I361C in
TM2 can be crosslinked either to L427C in TM3, F476C in TM4, P499C in TM5, or I525C in TM6. Furthermore, this TM arrangement may show that there is a site between them where substrate binds, which is consistent with the previously reported substrate contacts
[237]. In our limited Cys crosslinking studies, TM2 and TM3 form the most contacts of 68 the Cys pairs tested and also seem to be closely associated. This close association of TM2 and TM3 is supported by the fact that a mutation in TM2 (T362E) suppresses the cold- sensitive phenotype of YidC in TM3 (C423R) [262]. The number of contacts between TM2 and TM3 may also arise because of the flexibility of the N-terminal half of TM3 due, in part, to the large number of helix-breaking residues (P419, G421, G422, P425, P431 and
G442). A GXXP motif (similar to the one present in the TM3 segment of YidC) has been shown to be important to increase the flexibility of a channel protein [263]. The membrane- associated periplasmic loop region preceding TM2 was able to crosslink efficiently with short spanner lengths to multiple upstream TM segments (TM3 and TM4 segments) (i.e.
W346C/M439C, F350/M439C, F350/M430C, W346C/Q461C, F350C/Y465C) close to the periplasmic side of the membrane (Figs 3.2 and 3.3). Either a portion or the entire surface of the P1 membrane associated loop (residues 325-352) interacting with TM3 and
TM4 could form the extracellular region of YidC and help maintain the permeability barrier when YidC is inserting substrate. This periplasmic region in between the large periplasmic domain and TM2 has been shown to be important for the function of YidC [226].
Our results are consistent with a model in which TM2 to TM6 face each other in the ground state and the substrate-binding site is toward the TM2-TM6 center [237]. Indeed, Klenner and Kuhn proposed, based on YidC/Pf3 coat disulfide crosslinking studies, a model where the substrate TM segment binds in between TM1, TM3, TM4 and TM5. They found the substrate TM segment makes contact to TM1, TM3, TM4 and TM5 across the entire length of the membrane [237]. Notably, Kohler et al [229] proposed that a YidC dimer is bound to the translating ribosome at the exit channel by cryo-electron microscopy (cryoEM). 69
They hypothesized that a pore is formed between TM2 and TM3 of one subunit and TM2 and TM3 of the other subunit. However, another cryo-EM study of a nascent chain complex bound to a YidC chimera showed the ribosome made contact with a YidC monomer [230] and spectroscopic studies showed that monomeric YidC in nanodiscs is sufficient to bind a substrate [231]. In our Cys crosslinking studies, we did not observe any dimers of YidC, despite placing double cysteines over the entire conserved core region.
Our experiments also reveal that YidC is a very dynamic protein within cellular membranes. The majority of the tested Cys pairs in the cytoplasmic border region between
TM2 and the downstream TM helices showed crosslinking between a range of both flexible and rigid crosslinkers with different spanner lengths. For example, Cys pairs between TM2 and TM3 showed crosslinking with the three bifunctional crosslinking maleimides with spanner lengths varying from 5 to 16 Å and iodine, which forms a disulfide bond. Also reflecting the dynamic nature of YidC is the fact that Cys 361 in TM2 can crosslink to a range of residues in TM3, TM4, TM5 and TM6. For instance, Cys 361 can be crosslinked to residues spanning Cys 419 to 427 in TM3, Cys 472 to 480 in TM4, and Cys 495 to 503 in TM5, and Cys 525 to 529 in TM6 all with the same 1,2-MTS bifunctional reagent that has a flexible spanner length of ~5 Å. Figure 3.6 shows a cartoon view of the flexibility observed in the cytoplasmic half of the YidC transmembrane segments. The residues crosslinked to the TM2 Cys 361 are indicated by circles and the residues crosslinked to
Cys 370 in the TM2 border region are indicated by squares. Residue 361 in TM2 is highlighted and the crosslinking to residues in other TM segments is indicated. We also show that the cytoplasmic region of YidC is flexible as indicated by the crosslinking with 70 multiple reagents of different spanner lengths. Clearly, YidC is in different protein states, such as YidC by itself or YidC associated with other proteins; e.g. SecYEG and/or substrate. This could be the reason for the varying distances in the cytoplasmic border regions as well as the flexibility of crosslinking observed with Cys 361 paired with Cys in other tested TM helices. One drawback of the crosslinking technique employed is that the residues only have to be in proximity for a fraction of a second for crosslinking to occur and this could also explain the flexibility that we observed.
Although flexibility of proteins in the membrane is not novel [264], a protein with multiple highly flexible TM segments is quite interesting. YidC has many conserved proline residues in the cytoplasmic half of the TM segments (425P and 431P in TM3, 468P in
TM4, and 499P in TM5), which could explain the observed flexibility in our study. This flexibility may explain why YidC is able to act as an insertase and assembly site for a wide range of membrane proteins despite it not requiring an energy source for activity.
During the course of this work, the structure of the Bacillus halodurans YidC2 containing five TM segments was solved by X-ray crystallography at 2.4Å resolution [265]. Briefly, the structure revealed that B. halodurans YidC2 has a novel protein fold with the five TM regions forming a hydrophilic groove that is open to both the cytoplasm and to the membrane interior. At the entrance of the hydrophilic groove, YidC2 has a hairpin structure in the cytoplasm, which is parallel to the membrane surface and might be involved in substrate binding. Intriguingly, Kumazaki et al (2014) proposed the C1 region containing the helical hairpin is very flexible; this is consistent with our data. First, the flexibility is
71 supported by the fact that, in the structure of the two B. halodurans YidC2 variants, the C1 has two very different orientations, with respect to the membrane-embedded region.
Second, the high crystallographic B factors of the C1 region, as well as the cytoplasmic side of the TM segments (in comparison to other regions of the protein), support the flexibility of these regions. Third, Molecular Dynamic simulations of the B. halodurans
YidC2 show that the hairpin structure within the cytoplasm is very dynamic and flexible.
Thus, the C1 region and TM cytoplasmic side are flexible both in the E. coli and B. halodurans YidC.
The B. halodurans structure reveals that TM1 is in proximity to TM2 and TM5, and TM5 is also in proximity to TM3 and TM4. Since the E. coli YidC has an extra TM segment at its amino-terminus, TM2, TM3, TM4, TM5 and TM6 of the E. coli YidC correspond to
TM1, TM2, TM3, TM4 and TM 5 of the B. halodurans YidC2. Therefore, at first glance, the structure is consistent with our E. coli YidC model where TM2 faces TM3 and TM6.
Although TM1 does not contact TM3 and TM4 in the case of the crystal structure, E. coli
TM2 could come close to the corresponding TM segments with breathing, such that it could be crosslinked with the 5Å MTS crosslinking agent. Indeed, our Cys 361 crosslinking data with Cys residues introduced into the TM3, TM4, TM5, or TM6 segment show that the
Cys 361 can come in close proximity to a wide variety of Cys residues even within a TM segment that is predicted to be far away from TM2 (Fig. 3.6).
The deviation of the observed distances between the structure and our crosslinking data could be due to: divergence of functional homologues, inherent flexibility and breathing of
72 the membrane embedded domain, structural variances between lipid and detergent solubilization, or conformational changes induced by substrate [266, 267] and/or partner protein binding [88]. The flexibility or different conformational states of YidC with substrate bound may account for the large proximity differences we observe from crosslinking versus those obtained from the crystal structure of the B. halodurans YidC2 without substrate.
In summary, this report on the E. coli YidC shows that the cytoplasmic border region and the cytoplasmic half of the YidC conserved TM segments are flexible. The flexibility may be important for changing conformational states of YidC upon capturing substrate. We also show that the loop that links the large P1 domain to TM2 can be crosslinked efficiently to the TM3 and TM4 segments. Our studies point to the TM2 segment being in close proximity to TM3, TM4, TM5 and TM6 in at least one, but possibly more conformational states.
73
3.4 Materials and Methods
Materials
N-N’-p-phenylenedimaleimide (p-PDM), N-N’-o-phenylenedimaleimide (o-PDM), lysozyme, thrombin and kanamycin were from Sigma-Aldrich. 1,2-Ethanediyl bismethanethiosulfonate (1,2 MTS) and 1,6 Hexanediyl bismethanethiosulfonate (1,6
MTS) cross-linkers were purchased from Santa Cruz Biotechnology. N-ethylmaleimide
(NEM), 1,6-bis(maleimido)hexane (BMH), and super signal west pico chemiluminescent substrate were from Pierce. Isopropyl 1-thio-β,D-galactopyranoside (IPTG), n-dodecyl-β-
D-maltopyranoside (DDM), and dithiothreitol were from Anatrace. Pfu Turbo DNA polymerase was obtained from Stratagene. The 6-His tag antibody was from Abcam.
Plasmids and strains
BL21(DE3), JS7131 (YidC depletion strain) and C41 were from our lab collection. The yidC gene was amplified from E.coli strain MC1060 and subcloned into the expression vector pEH1 using NdeI and HindIII. A 10x-His tag was introduced into the C-terminus of a Cys-less YidC (C423S) within pEH1-YidC using site-directed mutagenesis.
Construction of YidC with double Cys residues
The Quikchange mutagenesis method (Stratagene Inc) was utilized to make double Cys mutants for the crosslinking studies using the Cys-less YidC expression vector (pEH1-
74 yidC) with a tandem thrombin protease site (L-V-P-R-G-S)2 inserted in the C1 loop after
394R.
Expression of YidC Cys pairs and membrane preparation for crosslinking studies
Overnight cell cultures bearing the pEH1 plasmid expressing the YidC Cys pairs were back
o diluted 1:100, grown at 37 C until OD600 ~ 0.6, and induced with 1mM IPTG for 2h. The cell cultures were harvested, washed once with PBS buffer pH 7.2, and then resuspended in PBS buffer. Membrane vesicles were prepared by adding 0.1mM EDTA and 0.1mg/mL lysozyme to the culture followed by incubation on ice for 10 min. The membrane preparations were then sonicated and unbroken cells were pelleted by centrifugation at
5,000 x g for 5 min.
Site-directed cross-linking
Membrane samples with the YidC double Cys mutants possessing a tandem Thrombin protease site in the C1 loop were incubated with 0.5mM aqueous iodine for 10 min at room temperature to catalyze disulfide bond formation. To quench the disulfide reaction, NEM
(10 mM final concentration) was added to modify any unreacted Cys residues. For Cys pairs located in the cytoplasmic or periplasmic membrane border regions, crosslinking reactions were initiated by incubation with various homo-bifunctional cysteine specific crosslinkers, namely o-PDM, p-PDM, and BMH to a final concentration of 0.5mM at room temperature for 10 min. DTT was added at a final concentration of 10mM to quench the
75 reactions. Chemical crosslinking of Cys pairs in transmembrane regions was carried out for 10 min with two homo-bifunctional MTS reagents, namely 1,2 MTS and 1,6 MTS at a final concentration of 0.5mM as well as with iodine (0.5 mM final concentration) for 10 min. The MTS reactions were quenched by adding NEM at a final concentration of 10 mM to react with unmodified Cys residues. After crosslinking, the membranes were pelleted using ultra-centrifugation and the supernatant removed. The membrane pellet was resuspended in Thrombin reaction buffer [50mM Tris-HCl, 150mM NaCl, 2.5mM CaCl2,
1% DDM, pH 8.0] and the samples were then centrifuged at 18,000 x g for 5 min in order to remove unsolubilized or aggregated proteins and other debris. After centrifugation, 0.2 units of thrombin were added to the supernatant and the protease reaction was incubated overnight at room temperature. Gel sample buffer was added to the samples and the proteins were resolved on 15% (w/v) polyacrylamide gels. Following electroblotting onto nitrocellulose membranes, they were probed for YidC using the 6-His antibody against the
C-terminal His tag.
Complementation assay
The YidC depletion strain, JS7131 bearing the pEH1-YidC vector encoding the Cys pairs was grown in LB media containing 0.2% arabinose and 50μg/ml Kanamycin at 37°C. After washing twice with fresh LB media, the overnight culture was back-diluted 1:100 into LB media without arabinose. The culture was grown for 2h and then serially diluted (1:10,
1:100, 1:1,000 and 1:10,000) in LB medium. A portion of the diluted cells was then spotted on LB plates containing 50μg/ml Kanamycin, and 0.2% arabinose or 0.2% glucose plus 76
50μM IPTG to express the double cysteine mutants. The plates were then incubated at
37°C overnight.
77
3.5 Tables
Table 3.1 – Complementation efficiency of YidC double cysteine mutants in JS7131
(continued) 78
Table 3.1 Continued
Double cysteine mutants were tested for complementation using the pEH1 vector in JS7131 under wild type conditions (with arabinose) or under depletion conditions (in the presence of glucose) with 50 µM IPTG. Overnight cultures were diluted 1:100 and grown for 2h then serially diluted (1:10, 1:100, 1:1,000 and 1:10,000) in LB medium. The cells were then spotted onto LB agar plates and incubated overnight at 37ºC. The efficiency of complementation is expressed as a comparison to wild type levels with 1 being equal to wild type. If the mutant requires 10 times as many cells added to the plate for complementation to occur when compared to wild-type, it would have an efficiency of 0.1;
100 times would be 0.01; 1000 times the cells would be 0.001; and at 10,000 times the cell would be 0.0001, which fails to rescue growth.
79
Table 3.2 – Crosslinking efficiency of YidC double cysteine mutants
Crosslinking experiments were performed in duplicate and the westerns were quantitated using ImageJ in order to determine the percentage of the sample that was cross-linked. The average of the two values for each mutant are reported in the table and color coded according to the following categories: 5% or below = no cross-linking; 6-20% = weak cross-linking; 21- 50% = moderate cross-linking; 51-100% = strong cross-linking.
80
3.6 Figures
Figure 3.1 – Cys-based alkylation method mapping the membrane border regions, and complementation assay (continued) 81
Figure 3.1 Continued
(A) Approximate TM location of YidC cysteine mutants used in this study with the tandem
Thrombin site indicated. (B) Structure and spanner length of the bifunctional crosslinking reagents. (C) No crosslinking is observed with the Cys-less YidC or Y370C/R447C control and complete thrombin cleavage. Membranes were prepared from BL21 cells expressing the Cys-less YidC construct and were treated with various bifunctional crosslinking reagents or iodine catalyzing disulfide bond formation. After treatment with the chemical reagents, membranes containing YidC were solubilized with DDM and incubated with Thrombin, as described in the Material and Methods. Proteins were resolved on a 15% (w/V) polyacrylamide gel and electroblotted onto a nitrocellulose membrane and then analyzed with an anti-6His antibody. (D) Complementation assay for monitoring the functionality of the YidC mutants. JS7131 bearing the pEH1YidC Cys- less, the T373C/E415C, or the Y370C/R447C mutant was grown under YidC expression
(in the presence of arabinose) or YidC depletion (in the presence of glucose) condition.
Cultures were serially diluted and spotted onto a LB agar plate with glucose and 50 µM
IPTG in order to express the cysteine mutants. Plates were incubated overnight at 37 oC.
82
Figure 3.2 – Crosslinking of paired Cys residues in TM2 and TM3 (continued)
Membranes were prepared from BL21 cells expressing various Cys pairs and treated with various bifunctional crosslinking reagents or iodine, which catalyzes disulfide bond
83
Figure 3.2 Continued formation. The membranes were treated with Thrombin, as described in the Material and
Methods. Samples were analyzed by SDS-PAGE, followed by western blotting with anti-
6His antibody. The crosslinked YidC runs at a position close to the native YidC. (A)
Crosslinking between residues at or near the cytoplasmic border of the TM segments was performed using maleimides and iodine. (B) Residues thought to be located within the membrane were crosslinked using two MTS reagents and iodine. (C) For residues exposed to the periplasm, crosslinking was performed using the maleimides and iodine.
84
Figure 3.3 – Crosslinking of paired Cys residues in TM2 and TM4
Membranes prepared from BL21 cells expressing the double Cys mutants were prepared, treated with crosslinking reagents, and Thrombin digestion as described in Materials and
Methods. Samples were subjected to SDS-PAGE, followed by western blotting with anti-
His antibody. A. Crosslinking between residues near the cytoplasmic border of the TM segments was performed using maleimides and iodine. B. Residues thought to be located within the membrane were crosslinked using two MTS reagents and iodine. C. For residues exposed to the periplasm, crosslinking was performed using the maleimides and iodine.
85
Figure 3.4 – Crosslinking of paired Cys residues in TM2 and TM5
Conditions for membrane preparation and crosslinking are described in Materials and
Methods. Samples were analyzed by SDS-PAGE and western blotting with anti-His antibody. (A) Crosslinking between residues near the cytoplasmic border of the TM segments was performed using maleimides and iodine. (B) Residues thought to be located within the membrane were crosslinked using two MTS reagents and iodine. (C) For residues exposed to the periplasm, crosslinking was performed using the maleimides and iodine.
86
Figure 3.5 – Crosslinking of paired Cys residues in TM2 and TM6
Conditions for crosslinking and western blotting are described in the Materials and
Methods. (A) Crosslinking between residues near the cytoplasmic border of the TM segments was performed using maleimides and iodine. (B) Residues thought to be located within the membrane were crosslinked using two MTS reagents and iodine.
87
Figure 3.6 – Cartoon showing one possible arrangement of TM2, TM3, TM4, TM5 and TM6 of YidC based on the crosslinking analysis
In this schematic of the flexibility observed, we assume that there is only one population of YidC. For simplicity, the N-terminal portion of YidC was not shown and is represented by the hash marks. The orange circles depict residues that crosslink to 361C (membrane) with only 1,2 MTS; the blue circles are mutants that crosslink with 1,2 and 1,6 MTS. The tan squares depict residues that crosslink to 370C (cytoplasm) with all of the maleimides; the green squares are mutants that crosslink with iodine and all of the maleimides. The size of the symbol represents an approximation of the total crosslinking observed for that pair of residues.
88
Figure 3.7 – The amount of YidC detergent solubilized from membranes prepared from BL21 or C41 bearing pEH1YidC
o Cultures of BL21 or C41 bearing the pEH1 plasmid were grown at 37 C until OD600 ~ 0.6.
The cultures were induced with 1mM IPTG (final concentration) in BL21 for 2 hours and with 0.4 mM IPTG for C41 for 2 hours. The membrane vesicles were prepared by lysozyme treatment and sonication (see Materials and Methods). Membranes were then pelleted using ultra-centrifugation and the supernatant removed. The membrane pellet was resuspended in Thrombin reaction buffer containing detergent [50mM Tris-HCl, 150mM
NaCl, 2.5mM CaCl2, 1% DDM, pH 8.0] and the samples were then centrifuged at 18,000 x g for 5 min in order to remove unsolubilized or aggregated proteins and other debris.
After the 18,000 x g low speed spin, the supernatant was removed (sample 1) and the membranes were resuspended in buffer containing 1% DDM. The samples were then ultra- centrifuged to pellet any unextracted protein or aggregates and the supernatant (sample 2) and pellet (sample 3) were analyzed.
89
Figure 3.8 – Crosslinking of specific Cys pairs expressed in C41
To confirm the crosslinking results with BL21, membranes were prepared from cultures of
C41 bearing the pEH1-YidC double Cys mutants. The crosslinking and western blot analysis was done exactly as described for the crosslinking performed with BL21 except that 0.4mM IPTG was used for induction of the YidC mutants (see Materials and
Methods). Mutants at the membrane interface (A) and within the membrane (B) were used to confirm the results.
90
Figure 3.9 – Comparison of reagent crosslinking efficiency
The reactivity of the MTS and the maleimide reagents was determined using a membrane bound (I361C/L427C) or aqueous exposed (Y370C/E415C) mutant; both mutants showed a wide variety of strong crosslinking in our studies (see Figure 3.2). The MTS pattern is the same as previously described for the I361C/L427C mutant but none of the maleimides were able to crosslink because of the lack of cysteine thiolate ions within the membrane.
All of the maleimides were able to react in the aqueous environment of the Y370C/E415C mutant and MTS crosslinking was also observed.
91
CHAPTER 4
YidC and Sec translocase facilitate membrane insertion of TolQ, a three-
spanning membrane protein
4.1 Introduction
In bacteria, multiple pathways exist for insertion of proteins into the cytoplasmic membrane [151, 152, 268]. The main pathway employs the essential SecYEG as a translocation channel [53] and the integral subunits SecDF(YajC) [87, 269, 270], YidC
[271], and the motor ATPase, SecA [272-274]. These proteins together form a holoenzyme
[217, 271], which is usually powered by the proton motive force and by nucleotide triphosphate hydrolysis [275, 276] or by ribosomal protein synthesis [11]. Proteins that require both YidC and SecYEG to insert into the membrane are subunits a and b of the
F1Fo ATP synthase [214] and NuoK [215]. The second pathway employs only the YidC insertase to integrate proteins into the membrane. Substrates for the YidC pathway include
M13 procoat [142], Pf3 coat [207, 277], subunit c of the F1Fo ATP synthase [208, 214, 92
248], MscL [144, 278], and TssL [209]; all of these proteins share a common feature in that they have short translocated regions. The third pathway is a Sec/YidC independent pathway, which may use an unidentified protein component for insertion or proteins may insert spontaneously into the membrane [240].
Recently, the determinants of a protein that dictate whether it uses the SecYEG/YidC holoenzyme or the YidC insertase for membrane insertion have been investigated. Zhu et al (2013) showed that the introduction of a negatively charged residue into the periplasmic region of a model YidC/Sec independent single spanning protein changes its insertion requirements to YidC-only while addition of a positively charged residue to the periplasmic region changes the insertion requirement to YidC and Sec-dependent [240]. More recently,
Soman et al (2014) studied the Sec and YidC requirements starting with the YidC dependent M13 procoat-Lep protein [241]. Interestingly, the results showed that increasing the hydrophilicity of the periplasmic loop of the M13 procoat-Lep protein above a certain threshold makes insertion strictly YidC and Sec-dependent. At a lower hydrophilicity level, insertion requires only YidC and insertion was YidC and Sec-independent below this level. These studies build on earlier studies which suggested that a negatively charged residue in a TM segment could be a YidC insertase determinant [215] and that YidC is required for proteins that had a very weak positive inside rule [243].
The Tol-Pal complex is found in most Gram-negative bacteria and consists of five proteins:
TolQ, TolR, TolA, TolB, and Pal. These proteins form a complex with three components in the inner membrane (TolA, TolQ, and TolR), one component in the periplasm (TolB),
93 and a lipoprotein in the inner leaflet of the outer membrane which is the peptidoglycan- associated lipoprotein (Pal). Studies have shown that certain components of the Tol-Pal complex have structural features similar to the flagellar motor components of MotA and
MotB and the TonB system which is composed of TonB, ExbB, and ExbD [279, 280]. All three of these systems are able to use the proton motive force (pmf) to facilitate cellular processes that require energy [281, 282]. The normal function of the Tol-Pal complex is to maintain the integrity of the cell envelope but phage particles and colicins can highjack the system in order to enter the cell [283, 284].
Mutation of the Tol-Pal genes can cause periplasmic proteins to be released from the cell, abnormal cell division, and decreased motility [285-287]. An increase in the number of released outer membrane vesicles outside of the cell was also observed during mutational analysis of the E. coli Tol-Pal proteins which supports the hypothesis that they are involved in cell envelope integrity [288, 289]. Additionally, the Tol-Pal system has been proposed to participate in the surface expression of specific lipopolysaccharides [290, 291]. An additional function of the complex was proposed by Gerding et al. when they observed that
Tol-Pal proteins accumulate where cells constrict before forming daughter cells and proposed that the complex is involved in the invagination of the outer membrane [292].
In this paper, we examine the translocation requirements of TolQ, which is a candidate
YidC protein because of its short periplasmic domains, and test whether mutations in the periplasmic loop can alter its insertion requirements. TolQ spans the membrane three times and is oriented in the membrane with its N-terminus in the periplasmic space and its C-
94 terminus in the cytoplasm [293, 294]. It is a 230 amino acid protein that contains two short periplasmic loops (N-tail is 19 residues and P1 is 18 residues) and two large, positively charged cytoplasmic loops. Evidence that the N-terminus is exposed to the periplasmic space comes from the finding that its N-terminal initiating methionine is formylated. The reason why the formyl group is not removed is because the N-terminal segment is inserted quickly across the membrane before the processing by deformylase can occur [294]. In a previous study using fusion proteins, TolQ was proposed to insert into the cytoplasmic membrane by a Sec-independent mechanism [295].
We show here that both YidC and SecYEG facilitate the insertion of the periplasmic loop of TolQ across the membrane. In addition, we find that a mutation of a charged residue in the P1 loop of TolQ decreases its YidC/SecYEG dependence of insertion. Conversely, by adding two charged residues to the P1 loop, insertion of the protein becomes strictly dependent on both translocases. Furthermore, while insertion of the wild-type TolQ occurs independent of the pmf, insertion of a positively charged mutant was inhibited by the pmf.
These studies show that TolQ inserts at the YidC/SecYEG interface and that mutations in the periplasmic loop can alter its insertion requirements.
4.2 Results
TolQ membrane topology and protease assay
The topology of TolQ is predicted to span the membrane three times and is oriented in the membrane with its N-terminus in the periplasmic space and its C-terminus facing the
95 cytoplasm [293, 294] (Fig. 4.1A). Previous studies on the membrane insertion of TolQ and its derivatives suggested that it inserts by a Sec-independent mechanism [295]. Since
YidC was not discovered at the time, we wanted to test whether TolQ inserts using the
YidC insertase for membrane integration. Moreover, we wanted to test what are the features of a three spanning protein that dictate whether it uses the YidC/Sec holoenzyme for membrane insertion and compare the determinants to what has previously been determined with other membrane proteins [240, 241].
To study the membrane biogenesis of TolQ, we added a bungarotoxin tag
(WRYYESSLLPYPD) to the C-terminus in order to precipitate the protein using a Biotin conjugated toxin. Proteinase K accessibility was employed to assay the membrane insertion of the protein [240]. BL21 cells expressing TolQ were labeled with [35S]- methionine for 20 seconds, converted to spheroplasts to allow access to the inner membrane, and then treated with proteinase K (PK) for 60 min. The full length TolQ as well as the N-terminal fragment and C-terminal fragment (detected in our immunoprecipitation studies) can be observed in Figure 4.1B. This study employed rifampicin to suppress endogenous protein synthesis in BL21 such that TolQ was the only protein being expressed. TolQ was expressed in pLZ1, which is under the control of the T7 promoter. Rifampicin does not inhibit T7 polymerase and therefore, synthesis of TolQ is not suppressed. Figure 4.1B&C show that TolQ is digested by PK to give rise to an approximately 10 kDa molecular weight fragment that can be precipitated. This fragment is produced by digestion of the P1 loop of TolQ, as previously described [293]. In our current study, no obvious digestion of the N-tail of TolQ was detected by treatment with 96 proteinase K. Therefore, in this work, we confined our membrane insertion studies to examine the requirements for translocation of the P1 loop. For all of our experiments reported here, we used the protease degradation of OmpA as a positive control for efficiency of spheroplast formation because OmpA is not digestible in intact cells but digested from the periplasmic side of the membrane in spheroplasts.
Insertion of the P1 loop of wild-type TolQ is partially YidC and Sec-dependent but independent of the pmf
The YidC dependence of TolQ membrane insertion was assayed using JS7131, the YidC depletion strain [142]. JS7131 has the yidC gene under the control of the araBAD promoter. JS7131 cells expressing TolQ were labeled with [35S]-methionine for 20 seconds under YidC depletion conditions (0.2% glucose) and protease mapping was performed as described in Fig. 4.1C. Figure 4.2A shows that the translocation of the P1 loop of TolQ was inhibited under YidC depletion conditions but not blocked. As a control, we confirmed that -5 Procoat-Lep is strongly inhibited under YidC-depletion condition and that the export of OmpA was unaffected by YidC depletion (Figure 4.2A lower panel). The results of this YidC depletion study show that YidC is able to facilitate the insertion of
TolQ.
To determine the Sec dependence of translocation of the TolQ P1 domain, we used the
SecE depletion strain CM124, which has the chromosomal secE gene knockout and a complementing plasmid with SecE under the control of the araBAD promoter [296]. It is
97 known that when SecE is depleted, proteins that require the SecYEG machinery are inhibited because SecY is degraded due to its instability without SecE present. Membrane insertion was examined under SecE depletion conditions by growing cells in M9 medium containing glucose. CM124 expressing TolQ was labeled with [35S]-methionine for 20 seconds under SecE expression (0.2% arabinose + 0.4% glucose) and SecE depletion conditions (0.4% glucose). After radiolabeling, the cells were converted to spheroplasts to allow the protease to have access to the inner membrane and then treated with PK for 60 min as above. Figure 4.2B shows that upon SecE depletion there is a partial, but measurable, defect in membrane insertion of the P1 loop compared to SecE expression conditions. As a positive control, we confirmed that the Sec-dependent OmpA accumulated in the precursor form in the cytoplasm under SecE-depleted condition and was PK resistant. These studies show that SecYEG is not strictly required for translocation of the TolQ P1 domain, although it is able to facilitate membrane insertion.
The inner membrane proton motive force (pmf) has been shown to be important for the amount of TolQ II/III localized to the membrane [295]. To examine the pmf dependence of full length TolQ, cells expressing TolQ were induced with IPTG for 3 min and then treated with the proton uncoupler CCCP for 45 s to dissipate the pmf prior to labeling the protein for 20 seconds. Figure 4.2C shows that translocation of the P1 loop of TolQ is efficient in the presence or absence of a pmf, showing that its translocation occurs independent of the pmf. As a control, we confirmed that the pmf-dependent OmpA was blocked in CCCP-treated cells (Fig. 4.2C, lower panel).
98
In previous studies [295], SecA was shown to be unrequired for the insertion of both TM1 and the TM2/3 hairpin of TolQ. To examine the SecA dependence of TolQ, MC1060 cells expressing TolQ were treated with 3mM final concentration of sodium azide to inhibit the function of SecA [297] and then induced with IPTG for 3 min prior to labeling the protein for 20 seconds. Figure 4.2D shows that translocation of the P1 loop of TolQ is efficient in the presence or absence of SecA. Notably, the SecA-dependent OmpA was inhibited in its export in azide-treated cells (Fig. 4.2D, lower panel).
The effects of mutations of the charged residues in the P1 loop on translocase requirements
Our studies show that translocation of the P1 loop of TolQ is stimulated by both YidC and the SecYEG machinery. However, neither translocase was absolutely required for P1 translocation. Therefore, we wanted to test whether mutations in the P1 loop could either decrease or increase the translocase requirement for membrane insertion. Inspection of the P1 loop shows it is predicted to be 18 residues long with two charged residues; a lysine at position 160 and a glutamic acid residue at position 173. We initially made the following single mutations to change the overall charge of the loop: E173R, E173Q, K160N, and
K160E. These mutations are predicted to change the overall charge of the loop to +2, +1,
-1, and -2, respectively.
Does the E173R TolQ have an increased or decreased translocase requirement? Figure
4.3A (left panel) shows that membrane insertion of E173R TolQ, which has two positively
99 charged residues in the P1 loop, is impaired even under YidC expression conditions.
However, membrane insertion was further inhibited when YidC was depleted, showing
YidC stimulates insertion. Similarly, the SecYEG translocase is required for membrane insertion of E173R TolQ (Fig. 4.3A, right panel). Thus, although changing the negatively charged glutamic acid at position 173 to a positively charged arginine residue lowers the extent of membrane insertion, membrane insertion becomes more dependent on YidC and
SecYEG than seen for the wild-type TolQ.
Similar translocase requirements were found with the E173Q mutant of TolQ containing one positively charged residue in the P1 loop, although membrane insertion is more efficient than the E173R mutant under conditions where YidC and SecYEG are present.
Like the E173R mutant, membrane insertion of E173Q (Fig. 4.3B) was strongly inhibited under YidC depletion conditions (Glc, left panel) and by SecYEG depletion conditions
(Glc, right panel). This shows that changing the charged residues from a net charge of 0
(wild type conditions) to +2 (E173R) or to +1 (E173Q) is sufficient to increase the translocase requirements for translocation of the P1 loop of the E173Q TolQ mutant compared to the wild-type TolQ.
In contrast, membrane insertion of the K160E mutant, containing two negatively charged residues in P1, has a reduced YidC translocase requirement. The left panel of figure 4.4A shows that YidC depletion leads to only a small reduction in the translocation of the P1 loop. SecE depletion leads to a small effect on membrane translocation of the P1 loop which shows that SecE is able to facilitate insertion much like with the wild type protein.
100
Thus, we were able to reduce the translocase requirements compared to wild-type TolQ by changing the lysine at the 160 position to a glutamic acid.
Insertion of the K160N mutant, containing one negatively charged residue in P1, has an increased SecYEG translocase requirement compared to wild-type TolQ. The left panel of figure 4.4C shows that YidC depletion leads to only a small reduction in the translocation of the P1 loop, like the wild-type TolQ. SecE depletion leads to a greater effect on translocation of the P1 loop, which shows that TolQ became dependent on SecYEG for insertion.
The effects of adding charged residues to the P1 loop on translocation
In order to test whether the addition of charges in the P1 loop could either decrease or increase the translocase requirement for membrane insertion, hydrophobic amino acids were mutated in order to add one or two extra charged residues. The L164E and I171K mutants added a single negative or positive charge and the L164E/I171E double mutant added two negatively charged residues. These mutations are predicted to change the overall charge of the loop to -1, +1, and -2, respectively as well as increase the hydrophilicity.
Increased translocase requirements compared to wild-type were found with the L164E mutant of TolQ, which contains an overall charge of -1 but a total of three charges in the loop. Like the E173Q mutant, membrane insertion of L164E (Fig. 4.5B) was inhibited under YidC depletion conditions (Glc, left panel) and by SecYEG depletion conditions
101
(Glc, right panel). This shows that increasing the number of charged residues from a net charge of 0 to -1 and simultaneously increasing the hydrophilic nature is sufficient to cause the P1 loop to be dependent on YidC and SecYEG for insertion.
By keeping the hydrophilic nature of the loop the same as the L164E mutant but switching the net charge, the increased Sec dependence was removed for the I171K mutant.
However, this mutant still remained YidC dependent much like the L164E mutation.
Though this mutant is facilitated weakly by the Sec translocase like the wild-type TolQ, it is more dependent on YidC with the added positively charged residue.
The mutation of two hydrophobic residues to two negatively charged residues was able to significantly alter the insertion requirement of the TolQ periplasmic loop. The insertion switched from YidC facilitated to YidC dependent as observed in the left panel of Figure
4.5E. Insertion of this more hydrophilic mutant with a net charge of -2 became strictly dependent on Sec (Figure 4.5F) for proper membrane insertion. Thus, adding two charged residues made TolQ more dependent on YidC and strictly required the Sec translocase.
Inhibition of the membrane insertion of the positively charged E173R is due to the pmf
It is well established that translocation of positively charged residues can be inhibited by the pmf (positive periplasm; negative cytoplasm) [240, 298]. Therefore, we tested whether the E173R mutant was inhibited due to the increase in the positive charge of the P1 loop from 0 to +2. To examine the pmf dependence of TolQ, MC1061 cells expressing E173R
102
TolQ were induced with IPTG for 3 min and then treated with the proton uncoupler CCCP for 45 s prior to pulse labeling for 20 seconds, as described in Figure 4.2C. As expected, in the presence of the pmf, the translocation of the P1 loop of TolQ is inefficient (Fig.
4.6A). However, when CCCP was added, the translocation of P1 was more efficient; see the increased production of the protease resistant fragment (Fig. 4.6A). These studies show that it is the pmf that impedes the membrane insertion of the P1 loop of the positively charged mutant.
To examine whether other charge changes in the P1 loop can alter the role of the pmf in membrane insertion, we examined the K173Q, K160N and the K160E mutants which changed the predicted charges in the loop from +1, -1, and -2, respectively. As can be seen, membrane insertion occurred to the same extent for each of these mutants in the presence or absence of the pmf (Fig. 4.6B-4D). Thus, sequentially changing the charge from +1 to -2 in the P1 loop still results in pmf-independent insertion, as seen in the wild- type TolQ.
4.3 Discussion
The Previously, Lewin and Webster provided experimental evidence that suggested that
TolQ inserts by a Sec-independent mechanism [295]. First, they took the TolQ region containing TM2, the PI loop, and TM3 and inserted it within the EcoR1 endonuclease.
This construct called TolQ II/III was capable of inserting into the membrane since it was digested by trypsin in permeabilized cells, suggesting the P1 loop was getting translocated.
103
Second, they showed that the amount of TolQ II/III in the membrane was not affected by the addition of the SecA inhibitor azide, indicating that this TolQ construct inserts independent of SecA. Third, they determined that that overexpression of TolQ did not affect the export of the Sec-dependent maltose binding protein. Their hypothesis was that
TolQ uses a Sec-independent mechanism for membrane insertion because overexpression of a Sec-dependent protein often perturbs the export of another Sec-dependent protein
[299]. Lastly, they discovered that the amount of membrane localized TolQ II/III was strictly dependent on the pmf, suggesting that membrane insertion is pmf-dependent.
In this paper, we show that YidC and, to a smaller extent, SecYEG does facilitate the translocation of the TolQ P1 loop but that neither is strictly required. The weak Sec- dependence may be the reason why no defect was observed in the export of maltose binding protein, which uses the Sec machinery, when TolQ was overexpressed [295]. As observed by Lewin and Webster [295] for TolQ II/III, translocation of the P1 loop of the intact TolQ is unaffected by inhibition of the SecA function. However, with the full-length TolQ, we find that translocation of the P1 loop is not perturbed by collapsing the pmf. In all cases, we confirmed that the pmf-dependent OmpA is completely blocked in export when the uncoupler CCCP is added.
In addition to examining the insertion mechanism of the three spanning wild-type TolQ protein, we investigated the translocase requirements to insert the P1 loop of TolQ, since the addition of charged residues can drastically change the insertion mechanism of a single span Pf3-lep model protein [240] and the two-spanning M13 procoat-P2 protein [241].
104
With the wild-type TolQ, the P1 loop has a lysine residue at position 160 and a glutamic acid residue at 173. We made four mutants, namely E173R, E173Q, K160N and K160E, where one of the charged residues was changed. Interestingly, the E173R and E173Q mutants became more dependent on both YidC and SecYEG. The K160E mutant is the only mutation studied in this work that had a decreased insertase requirement, which was for YidC. The K160N variant showed an increased dependence on SecYEG but the YidC requirement remained the same. Mutations where charges were added (L164E, I171K,
L16E/I171E) all exhibited an increased insertase requirement; the insertion of two negatively charged residues caused the P1 loop to strictly require the Sec translocase.
In contrast to the other mutants studied here, the TolQ E173R mutant, which has two positively charged residues in the PI loop, was impaired in membrane insertion even with
YidC and SecYEG present. We tested whether the pmf, with its positive side on the periplasmic side of the membrane, was responsible for this inhibition since this mutant contains a loop with a predicted positive charge near +2. Upon collapsing the pmf with
CCCP, we found that membrane insertion was more efficient, although still below the level seen with the wild-type protein and other mutants studied here. This is consistent with the membrane electrical potential (positive potential in the periplasmic space) impeding the transfer of the positively charged residues in the P1 loop to the periplasmic side of the membrane. Similar results have been found in other studies [240, 298].
Our hypothesis is that TolQ is inserted at the interface of YidC and SecYEG, which explains the fact that both translocases can contribute to insertion, i.e. YidC facilitates
105 translocation of the P1 loop while SecYEG is only weakly needed. The remarkable finding is that roughly 30 to 40% of the protein requires YidC while only ~15 to 20% of the protein requires SecYEG for membrane insertion. This suggests two possible scenarios: that at least 60% of the protein can insert independent of the translocase components or that insertion is promiscuous [48, 300]. In the later model, the protein can be targeted to and use either SecYEG or YidC un-preferentially for insertion. Thus, when YidC is depleted, a portion of the proteins can still insert by using SecYEG and vice versa.
Future studies need to investigate how many negatively charged or polar residues must be added to the P1 loop of TolQ to make its insertion completely YidC dependent and Sec- dependent. This will allow us to test the charge/polarity hypothesis for translocation of the
P1 region of TolQ. This hypothesis works well for predicting the translocase requirement for the M13 procoat-Lep protein. In addition to these studies, we will probe how the N- terminus of the TolQ protein translocates across the membrane. To do so, we need to develop an assay to detect N-tail translocation. This could possibly be accomplished by using an indirect cysteine specific gel-shift modification assay, by extending the tail to promote protease cleavage, by disrupting possible secondary structure of the tail, or by altering the features of the tail to allow cleavage by a protease other than PK.
106
4.4 Materials and Methods
Materials
Sodium azide, carbonyl cyanide m-chlorophenylhydrazone (CCCP), and lysozyme were from Sigma-Aldrich. Isopropyl 1-thio-β-d-galactopyranoside was from Research Products
International Corp. PMSF was purchased from Affymetrix. Proteinase K (PK) was purchased from Qiagen. Tran35S-label, a mixture of 85% [35S]-methionine and 15% [35S]- cysteine, 1000 Ci/mmol, was from PerkinElmer Life Sciences. -bungarotoxin, biotin conjugate was from Life Technologies. Immobilized streptavidin was from Fisher
Scientific. Antisera to outer membrane protein A (-OmpA) and leader peptidase (-Lep) were from our own laboratory collection.
Strains, Plasmids, and Growth conditions
JS7131, the E.coli YidC depletion strain [142], BL21, and MC1060 are from our laboratory collection. The SecE depletion strain, CM124, was a generous gift from Beth Traxler and was previously described in [296]. In both JS7131 and CM124, the expression of the YidC and SecE proteins is under the control of the araBAD promoter with arabinose expressing the proteins and glucose inhibiting the expression.
For the YidC depletion studies, JS7131 cells were grown in LB media at 37oC for approximately 3 hours with 0.2% arabinose for YidC expression or 0.2% glucose for the depletion of YidC. CM124 was grown for about 8 hours at 37oC in M9 minimal media supplemented with 0.2% arabinose and 0.4% glucose for the expression of SecE or 0.4% 107 glucose in order to deplete SecE. Both JS7131 and CM124 cells were subsequently washed with M9 media and then resuspended in M9 followed by growth for 30 minutes at 37oC before labeling with [35S]-methionine [301]. The pLZ1 vector was used to express the TolQ mutants in CM124, JS7131, and MC1060. The plasmid encoded TolQ gene was under the control of the lacUV5 promoter [240].
Protease accessibility studies
JS7131 Expression of the TolQ proteins encoded on the pLZ1 plasmid was induced by the addition of 1 mM isopropyl 1-thio-β-d-galactopyranoside (IPTG) for 3 minutes. Cells were then labeled with [35S]-methionine for 20 seconds and converted to spheroplasts [240] so that protease mapping studies could be performed with proteinase K. In order to make spheroplasts, the radiolabeled cells were collected by centrifugation followed by resuspension of the pellet in spheroplast buffer (33 mm Tris-HCl, pH 8.0, 40% (m/v)
Sucrose). The resuspended cells were treated with 1 mM EDTA and 10 μg/mL lysozyme on ice for 30 minutes followed by digestion with proteinase K (1.5 mg/mL) on ice for 1 hour. Phenylmethylsulfonyl fluoride (PMSF) was added at a final concentration of 5mM for 5 minutes on ice in order to inhibit the proteinase K. An equal volume of 20% (m/v) trichloroacetic acid (TCA) was then added to the sample followed by incubation on ice for
1 hour in order to precipitate the proteins in the sample. The precipitated protein was then pelleted at 14,000 × g for 10 minutes and washed with 1 mL of ice cold acetone. After removal of the supernatant, the pellets were dried to remove excess acetone and were then solubilized in Tris-SDS buffer (10 mm Tris-HCl, pH 8.0, 2% (m/v) SDS) overnight. The 108 samples were immunoprecipitated with -bungarotoxin, biotin conjugate for TolQ studies or with antiserum to OmpA as a control.
In the rifampicin study, cells were induced with 1 mM IPTG for 15 minutes to express T7
RNA polymerase followed by the addition of 0.25 mg/mL rifampicin and growth for 45 minutes to halt endogenous protein synthesis before the labeling of the cells. For the SecA studies, the cells were treated with sodium azide (3 mM final concentration) for 5 min prior to labeling of the cells to inhibit SecA. For the proton motive force (PMF) studies, the cells were treated with CCCP (final concentration 50 M) for 45 seconds followed by labeling of the cells with [35S]-methionine. Expression of the -5 PClep protein from the pLZ1 plasmid as a positive control for YidC depletion was carried out as described above for the YidC depletion studies. However, immunoprecipitation was performed using antibody to leader peptidase. In all cases, the samples were analyzed by 15% SDS-PAGE and phosphorimaging.
Mutagenesis of TolQ
All of the TolQ mutants were generated by site-directed mutagenesis using the
QuikChange mutagenesis kit (Stratagene) and were confirmed by DNA sequencing.
109
4.5 Figures
Figure 4.1 – Membrane topology of TolQ and expression controls (continued)
110
Figure 4.1 Continued
(A) Schematic of the membrane topology of TolQ with the positively charged amino acids in the periplasmic loops shown in red and the negatively charged residues in blue. (B)
Expression of TolQ in BL21 utilizing rifampicin to suppress endogenous protein synthesis followed by protease mapping of TolQ using proteinase K. The N-terminal fragment (top band) and C-terminal fragment (bottom band) of TolQ are marked with arrows. (C)
Immunoprecipitation and protease mapping of TolQ expressed in JS7131 using the pLZ1 plasmid under arabinose growth conditions as described (see Materials and Methods).
Molecular weight markers are labeled in both B&C and show the approximate molecular weight of the full length TolQ protein “F”, the C-terminal cleavage fragment “C”, and the
N-terminal fragment “N” after treatment with proteinase K. (D) Expression of the -5 PClep mutant (completely YidC dependent) in JS7131 under YidC expression and depletion conditions as a positive control for the effectiveness of YidC depletion. The band labeled
“P” refers to the full length PClep, which is in the cytoplasm. The band labeled “C” is mature coat-lep protein that was inserted into the membrane and the signal peptide was proteolytically removed by leader peptidase. The band labeled “F” is the protease resistant fragment after mature coat-lep was proteolyzed by proteinase K.
111
Figure 4.2 – Expression of the wild-type TolQ in the pLZ1 plasmid under various conditions (continued)
(A) E. coli JS7131 cells expressing TolQ were grown for 3 hours under YidC expression
(0.2% arabinose) or YidC depletion conditions (0.2% glucose) followed by a protease- mapping with proteinase K, as described in the Materials and Methods section. Inhibition of TolQ membrane insertion under YidC depletion conditions shows that YidC facilitates the insertion of the periplasmic loop of TolQ. (B) E. coli CM124 expressing TolQ were grown in SecE expression (0.2% arabinose + 0.4% glucose) or depletion conditions (0.4%
112
Figure 4.2 Continued glucose) and analyzed by protease-mapping (see Materials and Methods). Inhibition of
TolQ membrane insertion under SecE depletion conditions revealed that SecYEG is also able to facilitate the insertion of the TolQ P1 loop. (C) MC1060 cells expressing TolQ from pLZ1 were labeled with [35S]-methionine for 20 seconds and analyzed by PK mapping, as described (Materials and Methods). Where indicated (PMF -), samples were treated with CCCP (50 µM final concentration) for 45 seconds prior to radiolabeling and were then analyzed by protease mapping. Dissipation of the proton motive force (PMF) by the addition of CCCP (50 M final concentration) in the MC1060 strain did not affect the insertion of TolQ. (D) MC1060 cells in the log phase were labeled with [35S]- methionine for 20 seconds and analyzed by protease mapping, as described in the
“Materials and Methods” section. Where indicated (SecA -), samples were treated with sodium azide (3 mM final concentration) for 5 minutes prior to [35S]-labeling and were then analyzed by PK mapping. TolQ insertion was unaffected by the inhibition of SecA via treatment with sodium azide. OmpA data shown in the lower panels for each study indicate the effectiveness of proteinase K cleavage as well as the inhibition efficiency of
OmpA export under SecE depletion conditions or SecA inhibited condition with azide, or when the pmf was abolished. The band labeled as “P” refers to full length pro-OmpA, which is in the cytoplasm. The band marked “M” is mature OmpA that has been exported across the membrane and proteolytically cleaved by leader peptidase.
113
Figure 4.3 – Insertion requirements of positively charged mutants
Determination of the YidC and Sec requirement for membrane insertion of (A) E173R and
(B) E173Q which have a predicted net charge in the P1 loop of +2 and +1, respectively.
Insertion of the TolQ mutants was examined under YidC or SecE depletion conditions as previously described in Figure 4.2. Cells were labeled with [35S]-methionine for 20 seconds and analyzed by proteinase K accessibility (see Materials and Methods).
114
Figure 4.4 – Insertion requirements of negatively charged mutants
Determination of the YidC and Sec requirement for membrane insertion of (A) K160E and
(B) K160N which have a predicted net charge in the P1 loop of -2 and -1, respectively.
Insertion of the TolQ mutants was examined under YidC or SecE depletion conditions as previously described in Figure 4.2. Cells were labeled with [35S]-methionine for 20 seconds and analyzed by proteinase K accessibility (see Materials and Methods).
115
Figure 4.5 – Insertion requirements of charged and more hydrophilic mutants
Determination of the YidC and Sec requirement for membrane insertion of (A) L164E (B)
I171K and (C) L164E/I171E which have a predicted net charge in the P1 loop of -1, +1, and -2, respectively. Insertion of the TolQ mutants was examined under YidC or SecE depletion conditions as previously described in Figure 4.2. Cells were labeled with [35S]- methionine for 20 seconds and analyzed by proteinase K accessibility (see Materials and
Methods). 116
Figure 4.6 – Dependence on the pmf for membrane insertion
The Proton Motive Force (pmf) requirement for membrane insertion of (A) E173R and (B)
E173Q (C) K160E and (D) K160N which have a predicted net charge in the P1 loop of +2,
+1, -2, and -1 respectively. Membrane insertion of the TolQ mutants was examined under conditions where the PMF had been abolished as previously described in Figure 4.2. Cells were labeled with [35S]-methionine for 20 seconds and analyzed by proteinase K accessibility (see Materials and Methods).
117
CHAPTER 5
Conclusions
5.1 Summary of work performed
The primary findings of the work described are as follows:
(i) Proximity relationships between the five conserved transmembrane segments
of the E. coli YidC were proposed and flexibility of the cytoplasmic border
region and cytoplasmic half of the TM segments was observed. Additionally,
it appeared as though one face of each TM segment was oriented toward a
central cavity; this was confirmed by the two x-ray crystallography structures
in the year 2014. Because the study was performed in native membranes with
binding partners and substrates present, I propose that the flexibility is
important for changing conformational states of YidC upon capturing substrate.
However, the observed distances did not match those determined from the E.
118
coli YidC structure (Table 5.1). The deviation of the observed distances could
be due to the inherent flexibility of the TM segments, structural differences
between lipid and detergent solubilized samples, or multiple conformations
being observed in the reported studies because substrates and binding partners
were present.
(ii) The TolQ protein was determined to require YidC and, to a smaller extent,
SecYEG for translocation of the P1 loop into the periplasm. Further, the P1
loop did not require the SecA motor protein for translocation or the proton
motive force. Because of this dual translocase requirement, the current
hypothesis is that TolQ is inserted at the interface of YidC and SecYEG and
that insertion is promiscuous in that the protein can be targeted to YidC or
SecYEG depending on which translocase is available. Changing the overall
charge of the loop from neutral to a net positive or negative or increasing the
polarity of the loop by adding charged residues increased the translocase
requirement. The only exception was the K160E mutant, which showed a
decreased YidC requirement but an increased SecYEG requirement. The major
finding of this study was that TolQ can be a substrate of YidC and that
translocase used for insertion appears to be promiscuous between YidC and
SecYEG.
119
5.2 Future directions
Because the distances determined by the crosslinking studies described in Chapter 3 do not match the distances measured from the E. coli YidC x-ray crystallography structure, further crosslinking studies can be performed in order to confirm the validity of the data. Over- production of YidC could lead to misfolded protein and thus performing studies at basal levels of YidC expression would help to rule out that overproduction caused the large number of crosslinking interactions observed. Crosslinking interactions could also be determined utilizing whole cells instead of membrane vesicles because spheroplasting may cause inner and outer membrane hybrids which could lead to misfolded YidC proteins.
There are also various experiments that can be performed in order to better define the activity of the crosslinking reagents. It is possible that the act of modifying a cysteine with the crosslinker could result in a properly folded protein becoming misfolded and inactive.
Additionally, the cysteines tested could have varying degrees of reactivity (e.g. T361C could be more reactive compared to other cysteine mutants) which could lead to the large number of crosslinks observed over the ten minutes of the experiment. Testing the reactivity, especially of the TM2 mutants, would help to rule out effects due to an increased reactivity which could increase the chance that the second cysteine will react due to the fact that the second reaction essentially becomes intramolecular. Two additional tests could be performed in order to confirm the crosslinking results. First, a reagent that has an intermediate distance between 1,2 and 1,6-MTS could be used to confirm the crosslinking efficiency as well as the distances for mutants that crosslinked with 1,2 and 1,6 MTS.
120
Second, a single cysteine can be used in order to crosslink with p-2’- fluroacetylphenylalanine incorporated at various locations in YidC because this reaction will only occur if the two groups come into close contact.
Future studies of TolQ are required in order to determine the number of charges and/or overall charge required to make the P1 loop completely YidC and Sec dependent. Getting complete cleavage of the P1 loop is also imperative; this could be accomplished by adding a leader peptidase cleavage sequence (Ala – X – Ala) at the periplasmic side of TM2 so that the loop is cleaved by Lep if it is translocated into the periplasm. Once TolQ P1 is completely YidC/Sec dependent, the charge/polarity hypothesis for translocation can be tested in order to collect data on a native protein substrate. An investigation of the translocase requirements for the N-tail of TolQ should also be performed as it is quickly inserted because the formyl-methionine group stays attached to the protein which suggests that it is rapidly inserted as the protein is synthesized. To date, only minor protease cleavage has been observed with amino-peptidase K [293]. Possible strategies for detecting the insertion of the N-tail include: altering its properties to allow cleavage by a protease, disrupting possible secondary structures that could be inhibiting cleavage, performing an indirect cysteine specific gel-shift modification assay, extending the N-tail to promote cleavage by a protease, or any combination of the above. Once the insertion of the tail can be followed, similar studies to those described in chapter 4 can be performed in order to determine the translocation requirements. After the translocation requirements are known, further studies can be carried out in order to probe the validity of the polarity/charge hypothesis for the N-tail of the TolQ protein. 121
Development of a YidC/Sec double depletion strain would benefit these studies and the membrane insertion field in general. This is a difficult task because of the different levels of basal expression; YidC is normally expressed at approximately ten fold higher levels and thus it would be hard to efficiently deplete both using the same promoter for expression. One possibility would be to utilize a tunable expression strain like Lemo21 such that the expression of one protein occurs via the T7 RNA polymerase while the expression of the other is by the endogenous polymerase. The addition of L-rhamnose causes the expression of T7Lys which inhibits the T7 polymerase and allows the tuning of protein expression based upon the concentration of rhamnose added. Traditional methods such as the araBAD promoter can be used to control the expression of the second protein much like the current JS 7131 (YidC) and CM124 (SecE) depletion strains. Construction of this strain would allow us to test if insertion is promiscuous or if TolQ inserts at the interface of YidC and SecYEG in the holo-translocon complex. An alternative to a double depletion strain would be to use an inhibitor of one of the translocases in the depletion strain of the other. In either case, in vitro translocation studies using purified SecYEG and
YidC reconstituted into proteoliposomes can be used to confirm the in vivo results. This would allow us to definitively determine whether insertion is able to occur spontaneously and at the interface of YidC and SecYEG.
122
5.3 Tables
Table 5.1 – Comparison of distances between the E. coli YidC structure and the crosslinking results
The distance between the -carbons as well as the ends of the amino acid side chains for the specified residues are shown for the E. coli YidC structure. The range of crosslinking distances are also reported for comparison sake. Two of the residues were not present in the crystal structure so the nearby 528Q was used for measurement purposes where noted.
123
References
1. Alberts, B., Molecular biology of the cell, 1983. New York: Garland. 2. Krogh, A., et al., Predicting transmembrane protein topology with a hidden Markov model: application to complete genomes. J Mol Biol, 2001. 305(3): p. 567-80. 3. Cooper, G.M. and R.E. Hausman, The cell: ASM press, 2000. 4. Nelson, N. and B.-S. A, The complex architecture of oxygenic photosynthesis. Nat Rev Mol Cell Biol, 2004. 5: p. 971-982. 5. Bukau, B., J. Weissman, and A. Horwich, Molecular chaperones and protein quality control. Cell, 2006. 125(3): p. 443-51. 6. Ellgaard, L. and A. Helenius, Curr Opin Cell Biol, 2001. 13: p. 431-437. 7. Nielsen, H., et al., Identification of prokaryotic and eukaryotic signal peptides and prediction of their cleavage sites. Protein Eng, 1997. 10(1): p. 1-6. 8. Paetzel, M., R.E. Dalbey, and N.C. Strynadka, Crystal structure of a bacterial signal peptidase in complex with a beta-lactam inhibitor [see comments] [published erratum appears in Nature 1998 Dec 17;396(6712):707]. Nature, 1998. 396(6707): p. 186-90. 9. Rizo, J., et al., Conformational behavior of Escherichia coli OmpA signal peptides in membrane mimetic environments. Biochemistry, 1993. 32(18): p. 4881-94. 10. Chupin, V., et al., Biochemistry, 1995. 34: p. 11617-24. 11. Blobel, G., Intracellular protein topogenesis. Proc Natl Acad Sci U S A, 1980. 77(3): p. 1496-500. 12. Koch, H.G., et al., In vitro studies with purified components reveal signal recognition particle (SRP) and SecA/SecB as constituents of two independent protein- targeting pathways of Escherichia coli. Mol. Biol. Cell, 1999. 10(7): p. 2163-73. 13. Beck, K., et al., Discrimination between SRP- and SecA/SecB-dependent substrates involves selective recognition of nascent chains by SRP and trigger factor. EMBO J., 2000. 19(1): p. 134-43. 14. Angelini, S., S. Deitermann, and H.G. Koch, FtsY, the bacterial signal-recognition particle receptor, interacts functionally and physically with the SecYEG translocon. EMBO Rep, 2005. 5: p. 476.
124
15. High, S. and B. Dobberstein, The signal sequence interacts with the methionine- rich domain of the 54- kD protein of signal recognition particle. J Cell Biol, 1991. 113(2): p. 229-33. 16. Hoffmann, A., et al., Concerted action of the ribosome and the associated chaperone trigger factor confines nascent polypeptide folding. Mol Cell, 2012. 48(1): p. 63-74. 17. Lakshmipathy, S.K., et al., Identification of nascent chain interaction sites on trigger factor. J Biol Chem, 2007. 282(16): p. 12186-93. 18. Calloni, G., et al., DnaK functions as a central hub in the E. coli chaperone network. Cell Rep, 2012. 1(3): p. 251-64. 19. Van der Sluis, E.O. and A.J.M. Driessen., Stepwise evolution of the Sec machinery in Proteobacteria. Trends Microbiol., 2006. 14: p. 105-108. 20. Xu, Z., J.D. Knafels, and K. Yoshino, Crystal structure of the bacterial protein export chaperone secB. Nat Struct Biol, 2000. 7(12): p. 1172-7. 21. Randall, L.L., et al., Calorimetric analyses of the interaction between SecB and its ligands. Protein Sci, 1998. 7(5): p. 1195-200. 22. Knoblauch, N.T., et al., Substrate specificity of the SecB chaperone. J Biol Chem, 1999. 274(48): p. 34219-25. 23. Wild, J., et al., DnaK and DnaJ heat shock proteins participate in protein export in Escherichia coli. Genes Dev, 1992. 6(7): p. 1165-72. 24. Crane, J.M., et al., Sites of interaction of a precursor polypeptide on the export chaperone SecB mapped by site-directed spin labeling. J Mol Biol, 2006. 363(1): p. 63-74. 25. Fekkes, P., C. van der Does, and A.J. Driessen, The molecular chaperone SecB is released from the carboxy-terminus of SecA during initiation of precursor protein translocation. Embo J, 1997. 16(20): p. 6105-13. 26. Fekkes, P., et al., Zinc stabilizes the SecB binding site of SecA. Biochemistry, 1999. 38(16): p. 5111-6. 27. Zhou, J. and Z. Xu, Structural determinants of SecB recognition by SecA in bacterial protein translocation. Nat Struct Biol, 2003. 10(11): p. 942-7. 28. Crane, J.M., et al., Mapping of the docking of SecA onto the chaperone SecB by site-directed spin labeling: insight into the mechanism of ligand transfer during protein export. J Mol Biol, 2005. 353(2): p. 295-307. 29. Egea, P.F., R.M. Stroud, and P. Walter, Targeting proteins to membranes: structure of the signal recognition particle. Curr Opin Struct Biol, 2005. 15(2): p. 213-20. 30. Romisch, K., et al., Homology of 54K protein of signal-recognition particle, docking protein and two E. coli proteins with putative GTP-binding domains. Nature, 1989. 340(6233): p. 478-82. 31. Bernstein, H.D., et al., Model for signal sequence recognition from amino-acid sequence of 54K subunit of signal recognition particle. Nature, 1989. 340(6233): p. 482-6. 32. Luirink, J., et al., Signal-sequence recognition by an Escherichia coli ribonucleoprotein complex. Nature, 1992. 359(6397): p. 741-3.
125
33. Valent, Q.A., et al., Early events in preprotein recognition in E. coli: interaction of SRP and trigger factor with nascent polypeptides. Embo J, 1995. 14(22): p. 5494- 505. 34. Luirink, J., et al., An alternative protein targeting pathway in Escherichia coli: studies on the role of FtsY. Embo J, 1994. 13(10): p. 2289-96. 35. Zelazny, A., et al., The NG domain of the prokaryotic signal recognition particle receptor, FtsY, is fully functional when fused to an unrelated integral membrane polypeptide. Proc Natl Acad Sci U S A, 1997. 94(12): p. 6025-9. 36. de Leeuw, E., et al., Anionic phospholipids are involved in membrane association of FtsY and stimulate its GTPase activity. EMBO J., 2000. 19: p. 531. 37. Shan, S.O. and P. Walter, Molecular crosstalk between the nucleotide specificity determinant of the SRP GTPase and the SRP receptor. Biochemistry, 2005. 44(16): p. 6214-22. 38. Reyes, C.L., et al., X-ray structures of the signal recognition particle receptor reveal targeting cycle intermediates. PLoS One, 2007. 2(7): p. e607. 39. Prinz, A., et al., Evolutionarily conserved binding of ribosomes to the translocation channel via the large ribosomal RNA. Embo J, 2000. 19(8): p. 1900-6. 40. Powers, T. and P. Walter, Co-translational protein targeting catalyzed by the Escherichia coli signal recognition particle and its receptor. EMBO J., 1997. 16(16): p. 4880-6. 41. Freymann, D.M., et al., Structure of the conserved GTPase domain of the signal recognition particle. Nature, 1997. 385(6614): p. 361-4. 42. Montoya, G., et al., Crystal structure of the NG domain from the signal-recognition particle receptor FtsY. Nature, 1997. 385(6614): p. 365-8. 43. Bradshaw, N. and P. Walter, The signal recognition particle (SRP) RNA links conformational changes in the SRP to protein targeting. Mol Biol Cell, 2007. 18(7): p. 2728-34. 44. Keenan, R.J., et al., Crystal structure of the signal sequence binding subunit of the signal recognition particle. Cell, 1998. 94(2): p. 181-91. 45. Batey, R.T., et al., Crystal structure of the ribonucleoprotein core of the signal recognition particle. Science, 2000. 287(5456): p. 1232-9. 46. von Loeffelholz, O., et al., Structural basis of signal sequence surveillance and selection by the SRP-FtsY complex. Nat Struct Mol Biol, 2013. 20(5): p. 604-10. 47. Powers, T. and P. Walter, Reciprocal stimulation of GTP hydrolysis by two directly interacting GTPases. Science, 1995. 269: p. 1422-4. 48. Welte, T., et al., Promiscuous targeting of polytopic membrane proteins to SecYEG or YidC by the E. coli Signal Recognition Particle. Mol Biol Cell, 2012. 23(3): p. 464-79. 49. Pohlschroder, M., et al., Protein translocation in the three domains of life: variations on a theme. Cell, 1997. 91(5): p. 563-6. 50. Bieker, K.L. and T.J. Silhavy, PrlA (SecY) and PrlG (SecE) interact directly and function sequentially during protein translocation in E. coli. Cell, 1990. 61(5): p. 833-42.
126
51. Schatz, P.J. and J. Beckwith, Genetic analysis of protein export in Escherichia coli. Annu. Rev. Genet., 1990. 24: p. 215-48. 52. Brundage, L., et al., The purified E. coli integral membrane protein SecY/E is sufficient for reconstitution of SecA-dependent precursor protein translocation. Cell, 1990. 62(4): p. 649-57. 53. Van den Berg, B., et al., X-ray structure of a protein-conducting channel. Nature, 2004. 427(6969): p. 36-44. 54. Cheng, Z., et al., Identification of cytoplasmic residues of Sec61p involved in ribosome binding and cotranslational translocation. J Cell Biol, 2005. 168(1): p. 67-77. 55. Chiba, K., H. Mori, and K. Ito, Roles of the C-terminal end of SecY in protein translocation and viability of Escherichia coli. J Bacteriol, 2002. 184(8): p. 2243- 50. 56. Kuhn, P., et al., The bacterial SRP receptor, SecA and the ribosome use overlapping binding sites on the SecY translocon. Traffic, 2011. 12(5): p. 563-78. 57. Park, E. and T.A. Rapoport, Preserving the membrane barrier for small molecules during bacterial protein translocation. Nature, 2011. 473(7346): p. 239-42. 58. Harris, C.R. and T.J. Silhavy, Mapping an interface of SecY (PrlA) and SecE (PrlG) by using synthetic phenotypes and in vivo cross-linking. J Bacteriol, 1999. 181(11): p. 3438-44. 59. Maillard, A.P., et al., Deregulation of the SecYEG translocation channel upon removal of the plug domain. J Biol Chem, 2007. 282(2): p. 1281-7. 60. Saparov, S.M., et al., Determining the conductance of the SecY protein translocation channel for small molecules. Mol Cell, 2007. 26(4): p. 501-9. 61. Li, W., et al., The plug domain of the SecY protein stabilizes the closed state of the translocation channel and maintains a membrane seal. Mol Cell, 2007. 26(4): p. 511-21. 62. Gumbart, J. and K. Schulten, Structural determinants of lateral gate opening in the protein translocon. Biochemistry, 2007. 46(39): p. 11147-57. 63. Plath, K., et al., Signal sequence recognition in posttranslational protein transport across the yeast ER membrane. Cell, 1998. 94(6): p. 795-807. 64. Cannon, K.S., et al., Disulfide bridge formation between SecY and a translocating polypeptide localizes the translocation pore to the center of SecY. J Cell Biol, 2005. 169(2): p. 219-25. 65. Park, E. and T.A. Rapoport, Bacterial protein translocation requires only one copy of the SecY complex in vivo. J Cell Biol, 2012. 198(5): p. 881-93. 66. Boy, D. and H.G. Koch, Visualization of distinct entities of the SecYEG translocon during translocation and integration of bacterial proteins. Mol Biol Cell, 2009. 20(6): p. 1804-15. 67. Deville, K., et al., The oligomeric state and arrangement of the active bacterial translocon. J Biol Chem, 2011. 286(6): p. 4659-69. 68. Mitra, K., et al., Structure of the E. coli protein-conducting channel bound to a translating ribosome. Nature, 2005. 438(7066): p. 318-24.
127
69. Kaufmann, A., et al., Cysteine-directed cross-linking demonstrates that helix 3 of SecE is close to helix 2 of SecY and helix 3 of a neighboring SecE. Biochemistry, 1999. 38(28): p. 9115-25. 70. Breyton, C., et al., Three-dimensional structure of the bacterial protein- translocation complex SecYEG. Nature, 2002. 418: p. 662-665. 71. Das, S. and D.B. Oliver, Mapping of the SecA.SecY and SecA.SecG interfaces by site-directed in vivo photocross-linking. J Biol Chem, 2011. 286(14): p. 12371-80. 72. Lill, R., W. Dowhan, and W. Wickner, The ATPase activity of SecA is regulated by acidic phospholipids, SecY, and the leader and mature domains of precursor proteins. Cell, 1990. 60(2): p. 271-80. 73. Mori, H. and K. Ito, Different modes of SecY-SecA interactions revealed by site- directed in vivo photo-cross-linking. Proc Natl Acad Sci U S A, 2006. 103(44): p. 16159-64. 74. Miller, A., L. Wang, and D.A. Kendall, SecB modulates the nucleotide-bound state of SecA and stimulates ATPase activity. Biochemistry, 2002. 41(16): p. 5325-32. 75. Tani, K., et al., In vitro analysis of the process of translocation of OmpA across the Escherichia coli cytoplasmic membrane. A translocation intermediate accumulates transiently in the absence of the proton motive force. J Biol Chem, 1989. 264(31): p. 18582-8. 76. Driessen, A.J., SecA, the peripheral subunit of the Escherichia coli precursor protein translocase, is functional as a dimer. Biochemistry, 1993. 32(48): p. 13190- 7. 77. Jilaveanu, L.B., C.R. Zito, and D. Oliver, Dimeric SecA is essential for protein translocation. Proc Natl Acad Sci U S A, 2005. 102(21): p. 7511-6. 78. de Keyzer, J., et al., Covalently dimerized SecA is functional in protein translocation. J Biol Chem, 2005. 280(42): p. 35255-60. 79. Huber, D., et al., SecA interacts with ribosomes in order to facilitate posttranslational translocation in bacteria. Mol Cell, 2011. 41(3): p. 343-53. 80. Wu, Z.C., et al., Competitive binding of the SecA ATPase and ribosomes to the SecYEG translocon. J Biol Chem, 2012. 287(11): p. 7885-95. 81. Singh, R., et al., Cryo-electron microscopic structure of SecA protein bound to the 70S ribosome. J Biol Chem, 2014. 289(10): p. 7190-9. 82. Merz, F., et al., Molecular mechanism and structure of Trigger Factor bound to the translating ribosome. EMBO J, 2008. 27(11): p. 1622-32. 83. Halic, M., et al., Following the signal sequence from ribosomal tunnel exit to signal recognition particle. Nature, 2006. 444(7118): p. 507-11. 84. Schaffitzel, C., et al., Structure of the E. coli signal recognition particle bound to a translating ribosome. Nature, 2006. 444(7118): p. 503-6. 85. Duong, F. and W. Wickner, Distinct catalytic roles of the SecYE, SecG and SecDFyajC subunits of preprotein translocase holoenzyme. EMBO J., 1997. 16(10): p. 2756-68. 86. Gardel, C., et al., The secD locus of E.coli codes for two membrane proteins required for protein export. EMBO J, 1990. 9(10): p. 3209-16.
128
87. Pogliano, J.A. and J. Beckwith, SecD and SecF facilitate protein export in Escherichia coli. EMBO J, 1994. 13(3): p. 554-61. 88. Nouwen, N. and A.J. Driessen, SecDFyajC forms a heterotetrameric complex with YidC. Mol Microbiol, 2002. 44(5): p. 1397-405. 89. Scotti, P.A., et al., YidC, the Escherichia coli homologue of mitochondrial Oxa1p, is a component of the Sec translocase. Embo J, 2000. 19(4): p. 542-9. 90. Tsukazaki, T., et al., Structure and function of a membrane component SecDF that enhances protein export. Nature, 2011. 474(7350): p. 235-8. 91. Frauenfeld, J., et al., Cryo-EM structure of the ribosome-SecYE complex in the membrane environment. Nat Struct Mol Biol, 2011. 18(5): p. 614-21. 92. Becker, T., et al., Structure of monomeric yeast and mammalian Sec61 complexes interacting with the translating ribosome. Science, 2009. 326(5958): p. 1369-73. 93. Menetret, J.F., et al., Ribosome binding of a single copy of the SecY complex: implications for protein translocation. Mol Cell, 2007. 28(6): p. 1083-92. 94. Egea, P.F. and R.M. Stroud, Lateral opening of a translocon upon entry of protein suggests the mechanism of insertion into membranes. Proc Natl Acad Sci U S A, 2010. 107(40): p. 17182-7. 95. Palmer, T. and B.C. Berks, The twin-arginine translocation (Tat) protein export pathway. Nat Rev Microbiol. 10(7): p. 483-96. 96. Sambasivarao, D., et al., Multiple roles for the twin arginine leader sequence of dimethyl sulfoxide reductase of Escherichia coli. J Biol Chem, 2000. 275(29): p. 22526-31. 97. Hutcheon, G.W. and A. Bolhuis, The archaeal twin-arginine translocation pathway. Biochem Soc Trans, 2003. 31(Pt 3): p. 686-9. 98. van der Ploeg, R., et al., Environmental salinity determines the specificity and need for Tat-dependent secretion of the YwbN protein in Bacillus subtilis. PLoS One, 2011. 6(3): p. e18140. 99. van der Ploeg, R., et al., High-salinity growth conditions promote Tat-independent secretion of Tat substrates in Bacillus subtilis. Appl Environ Microbiol, 2012. 78(21): p. 7733-44. 100. Yen, M.R., et al., Sequence and phylogenetic analyses of the twin-arginine targeting (Tat) protein export system. Arch Microbiol, 2002. 177(6): p. 441-50. 101. Aldridge, C., et al., The chloroplast twin arginine transport (Tat) component, Tha4, undergoes conformational changes leading to Tat protein transport. J Biol Chem, 2012. 287(41): p. 34752-63. 102. Koch, S., et al., Escherichia coli TatA and TatB proteins have N-out, C-in topology in intact cells. J Biol Chem, 2012. 287(18): p. 14420-31. 103. Blaudeck, N., et al., Isolation and characterization of bifunctional Escherichia coli TatA mutant proteins that allow efficient tat-dependent protein translocation in the absence of TatB. J Biol Chem, 2005. 280(5): p. 3426-32. 104. Ize, B., et al., In vivo dissection of the Tat translocation pathway in Escherichia coli. J Mol Biol, 2002. 317(3): p. 327-35. 105. Rollauer, S.E., et al., Structure of the TatC core of the twin-arginine protein transport system. Nature, 2012. 492(7428): p. 210-4. 129
106. Graubner, W., A. Schierhorn, and T. Bruser, DnaK plays a pivotal role in Tat targeting of CueO and functions beside SlyD as a general Tat signal binding chaperone. J Biol Chem, 2007. 282(10): p. 7116-24. 107. Holzapfel, E., et al., The entire N-terminal half of TatC is involved in twin-arginine precursor binding. Biochemistry, 2007. 46(10): p. 2892-8. 108. Jong, W.S., et al., Trigger factor interacts with the signal peptide of nascent Tat substrates but does not play a critical role in Tat-mediated export. Eur J Biochem, 2004. 271(23-24): p. 4779-87. 109. Jack, R.L., et al., Coordinating assembly and export of complex bacterial proteins. Embo J, 2004. 23(20): p. 3962-72. 110. Schlesier, R. and R.B. Klosgen, Twin arginine translocation (Tat)-dependent protein transport: the passenger protein participates in the initial membrane binding step. Biol Chem, 2010. 391(12): p. 1411-7. 111. Panahandeh, S., et al., Following the path of a twin-arginine precursor along the TatABC translocase of Escherichia coli. J Biol Chem, 2008. 283(48): p. 33267-75. 112. Alami, M., et al., Differential interactions between a twin-arginine signal peptide and its translocase in Escherichia coli. Mol Cell, 2003. 12(4): p. 937-46. 113. Kreutzenbeck, P., et al., Escherichia coli twin arginine (Tat) mutant translocases possessing relaxed signal peptide recognition specificities. J Biol Chem, 2007. 282(11): p. 7903-11. 114. Strauch, E.M. and G. Georgiou, Escherichia coli tatC mutations that suppress defective twin-arginine transporter signal peptides. J Mol Biol, 2007. 374(2): p. 283-91. 115. Lausberg, F., et al., Genetic evidence for a tight cooperation of TatB and TatC during productive recognition of twin-arginine (Tat) signal peptides in Escherichia coli. PLoS One, 2012. 7(6): p. e39867. 116. Tarry, M.J., et al., Structural analysis of substrate binding by the TatBC component of the twin-arginine protein transport system. Proc Natl Acad Sci U S A, 2009. 106(32): p. 13284-9. 117. Celedon, J.M. and K. Cline, Stoichiometry for binding and transport by the twin arginine translocation system. J Cell Biol, 2012. 197(4): p. 523-34. 118. Alcock, F., et al., Live cell imaging shows reversible assembly of the TatA component of the twin-arginine protein transport system. Proc Natl Acad Sci U S A, 2013. 110(38): p. E3650-9. 119. Orriss, G.L., et al., TatBC, TatB, and TatC form structurally autonomous units within the twin arginine protein transport system of Escherichia coli. Febs Lett, 2007. 581(21): p. 4091-7. 120. Behrendt, J., U. Lindenstrauss, and T. Bruser, The TatBC complex formation suppresses a modular TatB-multimerization in Escherichia coli. FEBS Lett, 2007. 581(21): p. 4085-90. 121. Aldridge, C., et al., Substrate-gated docking of pore subunit Tha4 in the TatC cavity initiates Tat translocase assembly. J Cell Biol, 2014. 205(1): p. 51-65.
130
122. Ma, X. and K. Cline, Mapping the signal peptide binding and oligomer contact sites of the core subunit of the pea twin arginine protein translocase. Plant Cell, 2013. 25(3): p. 999-1015. 123. Fincher, V., C. Dabney-Smith, and K. Cline, Functional assembly of thylakoid deltapH-dependent/Tat protein transport pathway components in vitro. Eur J Biochem, 2003. 270(24): p. 4930-41. 124. Rodriguez, F., et al., Structural model for the protein-translocating element of the twin-arginine transport system. Proc Natl Acad Sci U S A, 2013. 110(12): p. E1092-101. 125. Gohlke, U., et al., The TatA component of the twin-arginine protein transport system forms channel complexes of variable diameter. Proc Natl Acad Sci U S A, 2005. 102(30): p. 10482-6. 126. Beck, D., et al., Ultrastructural characterisation of Bacillus subtilis TatA complexes suggests they are too small to form homooligomeric translocation pores. Biochim Biophys Acta, 2013. 1833(8): p. 1811-9. 127. Baglieri, J., et al., Structure of TatA paralog, TatE, suggests a structurally homogeneous form of Tat protein translocase that transports folded proteins of differing diameter. J Biol Chem, 2012. 287(10): p. 7335-44. 128. Zhang, Y., et al., Structural basis for TatA oligomerization: an NMR study of Escherichia coli TatA dimeric structure. PLoS One, 2014. 9(8): p. e103157. 129. Dabney-Smith, C., H. Mori, and K. Cline, Oligomers of Tha4 organize at the thylakoid Tat translocase during protein transport. J Biol Chem, 2006. 281(9): p. 5476-83. 130. Chan, C.S., et al., The TatA subunit of Escherichia coli twin-arginine translocase has an N-in topology. Biochemistry, 2007. 46(25): p. 7396-404. 131. Walther, T.H., et al., Membrane alignment of the pore-forming component TatA(d) of the twin-arginine translocase from Bacillus subtilis resolved by solid-state NMR spectroscopy. J Am Chem Soc, 2010. 132(45): p. 15945-56. 132. Musser, S.M. and S.M. Theg, Characterization of the early steps of OE17 precursor transport by the thylakoid DeltapH/Tat machinery. Eur J Biochem, 2000. 267(9): p. 2588-98. 133. Teter, S.A. and S.M. Theg, Energy-transducing thylakoid membranes remain highly impermeable to ions during protein translocation. Proc Natl Acad Sci U S A, 1998. 95(4): p. 1590-4. 134. Mori, H. and K. Cline, A twin arginine signal peptide and the pH gradient trigger reversible assembly of the thylakoid [Delta]pH/Tat translocase. J Cell Biol, 2002. 157(2): p. 205-10. 135. Rose, P., et al., Substrate-dependent assembly of the Tat translocase as observed in live Escherichia coli cells. PLoS One, 2013. 8(8): p. e69488. 136. White, G.F., et al., Subunit organization in the TatA complex of the twin arginine protein translocase: a site-directed EPR spin labeling study. J Biol Chem, 2010. 285(4): p. 2294-301. 137. Walther, T.H., et al., Folding and self-assembly of the TatA translocation pore based on a charge zipper mechanism. Cell, 2013. 152(1-2): p. 316-26. 131
138. Berks, B.C., F. Sargent, and T. Palmer, The Tat protein export pathway. Mol Microbiol, 2000. 35(2): p. 260-74. 139. Greene, N.P., et al., Cysteine scanning mutagenesis and disulfide mapping studies of the TatA component of the bacterial twin arginine translocase. J Biol Chem, 2007. 282(33): p. 23937-45. 140. Bruser, T. and C. Sanders, An alternative model of the twin arginine translocation system. Microbiol Res, 2003. 158(1): p. 7-17. 141. Maurer, C., et al., TatB functions as an oligomeric binding site for folded Tat precursor proteins. Mol Biol Cell, 2010. 21(23): p. 4151-61. 142. Samuelson, J.C., et al., YidC mediates membrane protein insertion in bacteria. Nature, 2000. 406(6796): p. 637-41. 143. Dalbey, R.E., P. Wang, and A. Kuhn, Assembly of Bacterial Inner Membrane Proteins. Annu Rev Biochem, 2011. 80: p. 161-87. 144. Facey, S.J., et al., The mechanosensitive channel protein MscL is targeted by the SRP to the novel YidC membrane insertion pathway of Escherichia coli. J Mol Biol, 2007. 365(4): p. 995-1004. 145. Neugebauer, S.A., et al., Membrane protein insertion of variant MscL proteins occurs at YidC and SecYEG of Escherichia coli. J Mol Biol, 2012. 417(4): p. 375- 86. 146. Houben, E.N., et al., Nascent Lep inserts into the Escherichia coli inner membrane in the vicinity of YidC, SecY and SecA. FEBS Lett., 2000. 476(3): p. 229-33. 147. Houben, E.N.G., et al., YidC and SecY mediate membrane insertion of a type I transmembrane domain. J. Biol. Chem., 2002. 277: p. 35880. 148. van der Laan, M., et al., Reconstitution of Sec-dependent membrane protein insertion: nascent FtsQ interacts with YidC in a SecYEG-dependent manner. EMBO Rep., 2001. 2(6): p. 519-23. 149. Wagner, S., et al., Biogenesis of MalF and the MalFGK(2) maltose transport complex in Escherichia coli requires YidC. J Biol Chem, 2008. 283(26): p. 17881- 90. 150. Samuelson, J.C., et al., Function of YidC for the insertion of M13 procoat protein in Escherichia coli: translocation of mutants that show differences in their membrane potential dependence and Sec requirement. J Biol Chem, 2001. 276(37): p. 34847-52. 151. Yi, L., et al., YidC is strictly required for membrane insertion of subunits a and c of the F(1)F(0)ATP synthase and SecE of the SecYEG translocase. Biochemistry, 2003. 42(35): p. 10537-44. 152. Denks, K., et al., The Sec translocon mediated protein transport in prokaryotes and eukaryotes. Mol Membr Biol. 31(2-3): p. 58-84. 153. Hegde, R.S. and R.J. Keenan, Tail-anchored membrane protein insertion into the endoplasmic reticulum. Nat Rev Mol Cell Biol. 12(12): p. 787-98. 154. Frobel, J., P. Rose, and M. Muller, Twin-arginine-dependent translocation of folded proteins. Philos Trans R Soc Lond B Biol Sci. 367(1592): p. 1029-46. 155. Yen, M.R., et al., Phylogenetic and structural analyses of the oxa1 family of protein translocases. FEMS Microbiol Lett, 2001. 204(2): p. 223-31. 132
156. Luirink, J., T. Samuelsson, and J.W. de Gier, YidC/Oxa1p/Alb3: evolutionarily conserved mediators of membrane protein assembly. Febs Lett, 2001. 501(1): p. 1- 5. 157. Funes, S., et al., The Oxa2 protein of Neurospora crassa plays a critical role in the biogenesis of cytochrome oxidase and defines a ubiquitous subbranch of the Oxa1/YidC/Alb3 protein family. Mol Biol Cell, 2004. 15(4): p. 1853-61. 158. Gerdes, L., et al., A second thylakoid membrane-localized Alb3/OxaI/YidC homologue is involved in proper chloroplast biogenesis in Arabidopsis thaliana. J Biol Chem, 2006. 281(24): p. 16632-42. 159. Bonnefoy, N., et al., Roles of Oxa1-related inner-membrane translocases in assembly of respiratory chain complexes. Biochim Biophys Acta, 2009. 1793(1): p. 60-70. 160. Preuss, M., et al., Evolution of mitochondrial oxa proteins from bacterial YidC. Inherited and acquired functions of a conserved protein insertion machinery. J Biol Chem, 2005. 280(13): p. 13004-11. 161. Szyrach, G., et al., Ribosome binding to the Oxa1 complex facilitates co- translational protein insertion in mitochondria. Embo J, 2003. 22(24): p. 6448-57. 162. Funes, S., et al., Evolution of YidC/Oxa1/Alb3 insertases: three independent gene duplications followed by functional specialization in bacteria, mitochondria and chloroplasts. Biol Chem, 2011. 392(1-2): p. 13-9. 163. Tjalsma, H., S. Bron, and J.M. van Dijl, Complementary impact of paralogous Oxa1-like proteins of Bacillus subtilis on post-translocational stages in protein secretion. J Biol Chem, 2003. 278(18): p. 15622-32. 164. Hasona, A., et al., Streptococcal viability and diminished stress tolerance in mutants lacking the signal recognition particle pathway or YidC2. Proc Natl Acad Sci U S A, 2005. 102(48): p. 17466-71. 165. Funes, S., et al., Independent gene duplications of the YidC/Oxa/Alb3 family enabled a specialized cotranslational function. Proc Natl Acad Sci U S A, 2009. 106(16): p. 6656-61. 166. Saaf, A., et al., Membrane topology of the 60-kDa Oxa1p homologue from Escherichia coli. J. Biol. Chem., 1998. 273(46): p. 30415-8. 167. Chiba, S., A. Lamsa, and K. Pogliano, A ribosome-nascent chain sensor of membrane protein biogenesis in Bacillus subtilis. Embo J, 2009. 28(22): p. 3461- 75. 168. Chiba, S. and K. Ito, Multisite ribosomal stalling: a unique mode of regulatory nascent chain action revealed for MifM. Mol Cell. 47(6): p. 863-72. 169. Nakatogawa, H. and K. Ito, The ribosomal exit tunnel functions as a discriminating gate. Cell, 2002. 108(5): p. 629-36. 170. van der Laan, M., N.P. Nouwen, and A.J. Driessen, YidC--an evolutionary conserved device for the assembly of energy-transducing membrane protein complexes. Curr Opin Microbiol, 2005. 8(2): p. 182-7. 171. Bauer, M., et al., PET1402, a nuclear gene required for proteolytic processing of cytochrome oxidase subunit 2 in yeast. Mol Gen Genet, 1994. 245(3): p. 272-8.
133
172. Bonnefoy, N., et al., OXA1, a Saccharomyces cerevisiae nuclear gene whose sequence is conserved from prokaryotes to eukaryotes controls cytochrome oxidase biogenesis. J Mol Biol, 1994. 239(2): p. 201-12. 173. Altamura, N., et al., The Saccharomyces cerevisiae OXA1 gene is required for the correct assembly of cytochrome c oxidase and oligomycin-sensitive ATP synthase. FEBS Lett, 1996. 382(1-2): p. 111-5. 174. van der Laan, M., et al., A conserved function of YidC in the biogenesis of respiratory chain complexes. Proc. Natl. Acad. Sci. USA, 2003. 100: p. 5801. 175. Sundberg, E., et al., ALBINO3, an Arabidopsis nuclear gene essential for chloroplast differentiation, encodes a chloroplast protein that shows homology to proteins present in bacterial membranes and yeast mitochondria. Plant Cell, 1997. 9(5): p. 717-30. 176. Bellafiore, S., et al., Loss of Albino3 leads to the specific depletion of the light- harvesting system. Plant Cell, 2002. 14(9): p. 2303-14. 177. Ossenbuhl, F., et al., Efficient Assembly of Photosystem II in Chlamydomonas reinhardtii Requires Alb3.1p, a Homolog of Arabidopsis ALBINO3. Plant Cell, 2004. 16(7): p. 1790-800. 178. Spence, E., et al., A homolog of Albino3/OxaI is essential for thylakoid biogenesis in the cyanobacterium Synechocystis sp. PCC6803. J Biol Chem, 2004. 279(53): p. 55792-800. 179. Dong, Y., et al., Functional overlap but lack of complete cross-complementation of Streptococcus mutans and Escherichia coli YidC orthologs. J Bacteriol, 2008. 190(7): p. 2458-69. 180. Benz, M., et al., Alb4 of Arabidopsis Promotes Assembly and Stabilization of a Non Chlorophyll-Binding Photosynthetic Complex, the CF1CF0-ATP Synthase. Mol Plant, 2009. 2(6): p. 1410-24. 181. Jiang, F., et al., Chloroplast YidC homolog Albino3 can functionally complement the bacterial YidC depletion strain and promote membrane insertion of both bacterial and chloroplast thylakoid proteins. J. Biol. Chem., 2002. 277: p. 19281. 182. van Bloois, E., et al., The Sec-independent function of Escherichia coli YidC is evolutionary-conserved and essential. J. Biol. Chem, 2005. 280: p. 12996-3003. 183. He, S. and T.D. Fox, Membrane translocation of mitochondrially coded Cox2p: distinct requirements for export of N and C termini and dependence on the conserved protein Oxa1p. Mol Biol Cell, 1997. 8(8): p. 1449-60. 184. Hell, K., et al., Oxa1p mediates the export of the N- and C-termini of pCoxII from the mitochondrial matrix to the intermembrane space. FEBS Lett., 1997. 418(3): p. 367-70. 185. Hell, K., W. Neupert, and R.A. Stuart, Oxa1p acts as a general membrane insertion machinery for proteins encoded by mitochondrial DNA. EMBO J., 2001. 20(6): p. 1281-8. 186. Hell, K., et al., Oxa1p, an essential component of the N-tail protein export machinery in mitochondria. Proc. Natl. Acad. Sci. U S A, 1998. 95(5): p. 2250-5. 187. Bohnert, M., et al., Cooperation of stop-transfer and conservative sorting mechanisms in mitochondrial protein transport. Curr Biol. 20(13): p. 1227-32. 134
188. Jia, L., et al., Yeast Oxa1 interacts with mitochondrial ribosomes: the importance of the C-terminal region of Oxa1. Embo J, 2003. 22(24): p. 6438-47. 189. Jia, L., J. Kaur, and R.A. Stuart, Mapping of the Saccharomyces cerevisiae Oxa1- mitochondrial ribosome interface and identification of MrpL40, a ribosomal protein in close proximity to Oxa1 and critical for oxidative phosphorylation complex assembly. Eukaryot Cell, 2009. 8(11): p. 1792-802. 190. Kaur, J. and R.A. Stuart, Truncation of the Mrp20 protein reveals new ribosome- assembly subcomplex in mitochondria. EMBO Rep. 12(9): p. 950-5. 191. Keil, M., et al., Oxa1-ribosome complexes coordinate the assembly of cytochrome C oxidase in mitochondria. J Biol Chem. 287(41): p. 34484-93. 192. Saracco, S.A. and T.D. Fox, Cox18p is required for export of the mitochondrially encoded Saccharomyces cerevisiae Cox2p C-tail and interacts with Pnt1p and Mss2p in the inner membrane. Mol Biol Cell, 2002. 13(4): p. 1122-31. 193. Fiumera, H.L., S.A. Broadley, and T.D. Fox, Translocation of mitochondrially synthesized Cox2 domains from the matrix to the intermembrane space. Mol Cell Biol, 2007. 27(13): p. 4664-73. 194. Koppen, M. and T. Langer, Protein degradation within mitochondria: versatile activities of AAA proteases and other peptidases. Crit Rev Biochem Mol Biol, 2007. 42(3): p. 221-42. 195. Fiumera, H.L., et al., Translocation and assembly of mitochondrially coded Saccharomyces cerevisiae cytochrome c oxidase subunit Cox2 by Oxa1 and Yme1 in the absence of Cox18. Genetics, 2009. 182(2): p. 519-28. 196. Moore, M., et al., Chloroplast Oxa1p homolog albino3 is required for post- translational integration of the light harvesting chlorophyll-binding protein into thylakoid membranes. J. Biol. Chem., 2000. 275(3): p. 1529-32. 197. Schuenemann, D., et al., A novel signal recognition particle targets light- harvesting proteins to the thylakoid membranes. Proc. Natl. Acad. Sci. U S A, 1998. 95(17): p. 10312-6. 198. Moore, M., et al., Functional interaction of chloroplast SRP/FtsY with the ALB3 translocase in thylakoids: substrate not required. J Cell Biol, 2003. 162(7): p. 1245-54. 199. Jaru-Ampornpan, P., S. Chandrasekar, and S.O. Shan, Efficient interaction between two GTPases allows the chloroplast SRP pathway to bypass the requirement for an SRP RNA. Mol Biol Cell, 2007. 18(7): p. 2636-45. 200. Falk, S., et al., The C Terminus of the Alb3 Membrane Insertase Recruits cpSRP43 to the Thylakoid Membrane. J Biol Chem. 285(8): p. 5954-62. 201. Dunschede, B., et al., Interaction studies between the chloroplast signal recognition particle subunit cpSRP43 and the full-length translocase Alb3 reveal a membrane-embedded binding region in Alb3 protein. J Biol Chem. 286(40): p. 35187-95. 202. Mori, H., et al., Component specificity for the thylakoidal Sec and Delta pH- dependent protein transport pathways. J Cell Biol, 1999. 146(1): p. 45-56. 203. Trager, C., et al., Evolution from the prokaryotic to the higher plant chloroplast signal recognition particle: the signal recognition particle RNA is conserved in 135
plastids of a wide range of photosynthetic organisms. Plant Cell. 24(12): p. 4819- 36. 204. Klostermann, E., et al., The thylakoid membrane protein ALB3 associates with the cpSecY- translocase in Arabidopsis thaliana. Biochem J, 2002. 368(Pt 3): p. 777- 81. 205. Pasch, J.C., J. Nickelsen, and D. Schunemann, The yeast split-ubiquitin system to study chloroplast membrane protein interactions. Appl Microbiol Biotechnol, 2005. 69(4): p. 440-7. 206. Gohre, V., et al., One of two alb3 proteins is essential for the assembly of the photosystems and for cell survival in Chlamydomonas. Plant Cell, 2006. 18(6): p. 1454-66. 207. Chen, M., et al., Direct interaction of YidC with the Sec-independent Pf3 coat protein during its membrane protein insertion. J. Biol. Chem., 2002. 277(10): p. 7670-5. 208. van Bloois, E., et al., F(1)F(0) ATP synthase subunit c is targeted by the SRP to YidC in the E. coli inner membrane. Febs Lett, 2004. 576(1-2): p. 97-100. 209. Aschtgen, M.S., et al., The C-tail anchored TssL subunit, an essential protein of the enteroaggregative Escherichia coli Sci-1 Type VI secretion system, is inserted by YidC. Microbiologyopen. 1(1): p. 71-82. 210. du Plessis, D.J., N. Nouwen, and A.J. Driessen, Subunit a of cytochrome o oxidase requires both YidC and SecYEG for membrane insertion. J Biol Chem, 2006. 281(18): p. 12248-52. 211. van Bloois, E., et al., Distinct requirements for translocation of the N-tail and C- tail of the Escherichia coli inner membrane protein CyoA. J Biol Chem, 2006. 281(15): p. 10002-9. 212. Celebi, N., et al., Membrane biogenesis of subunit II of cytochrome bo oxidase: contrasting requirements for insertion of N-terminal and C-terminal domains. J Mol Biol, 2006. 357(5): p. 1428-36. 213. Kol, S., et al., Subunit a of the F(1)F(0) ATP synthase requires YidC and SecYEG for membrane insertion. J Mol Biol, 2009. 390(5): p. 893-901. 214. Yi, L., et al., Sec/SRP requirements and energetics of membrane insertion of subunits a, b, and c of the Escherichia coli F1F0 ATP synthase. J Biol Chem, 2004. 279(38): p. 39260-7. 215. Price, C.E. and A.J. Driessen, Conserved negative charges in the transmembrane segments of subunit K of the NADH:ubiquinone oxidoreductase determine its dependence on YidC for membrane insertion. J Biol Chem. 285(6): p. 3575-81. 216. Linhartova, M., et al., Accumulation of the Type IV prepilin triggers degradation of SecY and YidC and inhibits synthesis of Photosystem II proteins in the cyanobacterium Synechocystis PCC 6803. Mol Microbiol. 93(6): p. 1207-23. 217. Schulze, R.J., et al., Membrane protein insertion and proton-motive-force- dependent secretion through the bacterial holo-translocon SecYEG-SecDF-YajC- YidC. Proc Natl Acad Sci U S A. 111(13): p. 4844-9.
136
218. Sachelaru, I., et al., YidC occupies the lateral gate of the SecYEG translocon and is sequentially displaced by a nascent membrane protein. J Biol Chem, 2013. 288(23): p. 16295-307. 219. Li, Z., et al., Identification of YidC residues that define interactions with the Sec apparatus. J Bacteriol, 2014. 196(2): p. 367-377. 220. Gotzke, H., et al., YfgM is an ancillary subunit of the SecYEG translocon in Escherichia coli. J Biol Chem. 289(27): p. 19089-97. 221. Zhu, L., et al., Both YidC and SecYEG Are Required for Translocation of the Periplasmic Loops 1 and 2 of TatC, a 6-Membrane-Spanning Protein. J Mol Biol. 222. Urbanus, M.L., et al., Sec-dependent membrane protein insertion: sequential interaction of nascent FtsQ with SecY and YidC. EMBO Rep., 2001. 2: p. 524. 223. Beck, K., et al., YidC, an assembly site for polytopic Escherichia coli membrane proteins located in immediate proximity to the SecYE translocon and lipids. EMBO Rep, 2001. 2(8): p. 709-14. 224. Nagamori, S., I.N. Smirnova, and H.R. Kaback, Role of YidC in folding of polytopic membrane proteins. J Cell Biol, 2004. 165(1): p. 53-62. 225. Zhu, L., H.R. Kaback, and R.E. Dalbey, YidC--a molecular chaperone for LacY protein folding via the SecYEG machinery. J Biol Chem. 226. Jiang, F., et al., Defining the regions of Escherichia coli YidC that contribute to activity. J Biol Chem, 2003. 278(49): p. 48965-72. 227. Oliver, D.C. and M. Paetzel, Crystal structure of the major periplasmic domain of the bacterial membrane protein assembly facilitator YidC. J Biol Chem, 2008. 283(8): p. 5208-16. 228. Ravaud, S., et al., The crystal structure of the periplasmic domain of the Escherichia coli membrane protein insertase YidC contains a substrate binding cleft. J Biol Chem, 2008. 283(14): p. 9350-8. 229. Kohler, R., et al., YidC and Oxa1 form dimeric insertion pores on the translating ribosome. Mol Cell, 2009. 34(3): p. 344-53. 230. Seitl, I., et al., The C-terminal regions of YidC from Rhodopirellula baltica and Oceanicaulis alexandrii bind to ribosomes and partially substitute for SRP receptor function in Escherichia coli. Mol Microbiol, 2014. 91(2): p. 408-21. 231. Kedrov, A., et al., Elucidating the Native Architecture of the YidC: Ribosome Complex. J Mol Biol, 2013. 425(22): p. 4112-24. 232. Kumazaki, K., et al., Structural basis for Sec-independent membrane protein insertion by YidC. Nature, 2014. 509(7501): p. 516-520. 233. Hennon, S.W. and R.E. Dalbey, Cross-Linking-Based Flexibility and Proximity Relationships between the TM Segments of the Escherichia coli YidC. Biochemistry. 234. Wickles, S., et al., A structural model of the active ribosome-bound membrane protein insertase YidC. Elife. 3: p. e03035. 235. Kumazaki, K., et al., Crystal structure of Escherichia coli YidC, a membrane protein chaperone and insertase. Sci Rep. 4: p. 7299.
137
236. Yu, Z., et al., The conserved third transmembrane segment of YidC contacts nascent Escherichia coli inner membrane proteins. J Biol Chem, 2008. 283(50): p. 34635- 42. 237. Klenner, C. and A. Kuhn, Dynamic disulfide scanning of the membrane-inserting Pf3 coat protein reveals multiple YidC substrate contacts. J Biol Chem, 2012. 287(6): p. 3769-76. 238. Chen, Y., et al., The role of the strictly conserved positively charged residue differs among the Gram-positive, Gram-negative and chloroplast YidC homologs. J Biol Chem. 239. Lemaire, C., et al., A yeast mitochondrial membrane methyltransferase-like protein can compensate for oxa1 mutations. J Biol Chem, 2004. 279(46): p. 47464-72. 240. Zhu, L., et al., Charge composition features of model single-span membrane proteins that determine selection of YidC and SecYEG translocase pathways in Escherichia coli. J Biol Chem, 2013. 288(11): p. 7704-16. 241. Soman, R., et al., Polarity and charge of the periplasmic loop determine the YidC and sec translocase requirement for the M13 procoat lep protein. J Biol Chem. 289(2): p. 1023-32. 242. Herrmann, J.M. and N. Bonnefoy, Protein export across the inner membrane of mitochondria: the nature of translocated domains determines the dependence on the Oxa1 translocase. J Biol Chem, 2004. 279(4): p. 2507-12. 243. Gray, A.N., et al., Unbalanced charge distribution as a determinant for dependence of a subset of Escherichia coli membrane proteins on the membrane insertase YidC. MBio, 2011. 2(6). 244. Wang, P. and R.E. Dalbey, Inserting membrane proteins: the YidC/Oxa1/Alb3 machinery in bacteria, mitochondria, and chloroplasts. Biochim Biophys Acta. 1808(3): p. 866-75. 245. Saller, M.J., et al., The YidC/Oxa1/Alb3 protein family: common principles and distinct features. Biol Chem. 393(11): p. 1279-90. 246. Luirink, J., et al., Biogenesis of inner membrane proteins in Escherichia coli. Biochim Biophys Acta. 1817(6): p. 965-76. 247. Samuelson, J.C., et al., Function of YidC for the insertion of M13 procoat protein in E. coli: translocation of mutants that show differences in their membrane potential dependence and Sec requirement. J. Biol. Chem., 2001. 276: p. 34847. 248. Van Der Laan, M., et al., F1F0 ATP synthase subunit c is a substrate of the novel YidC pathway for membrane protein biogenesis. J Cell Biol, 2004. 165(2): p. 213- 22. 249. Xie, K., et al., Different regions of the nonconserved large periplasmic domain of Escherichia coli YidC are involved in the SecF interaction and membrane insertase activity. Biochemistry, 2006. 45(44): p. 13401-8. 250. Klenner, C., et al., The Pf3 coat protein contacts TM1 and TM3 during membrane biogenesis. Febs Lett, 2008: p. in press. 251. Dalbey, R.E. and A. Kuhn, YidC family members are involved in the membrane insertion, lateral integration, folding, and assembly of membrane proteins. J Cell Biol, 2004. 166(6): p. 769-74. 138
252. Wu, J., et al., Site-directed spin labeling and chemical crosslinking demonstrate that helix V is close to helices VII and VIII in the lactose permease of Escherichia coli. Proc Natl Acad Sci U S A, 1996. 93(19): p. 10123-7. 253. Wolin, C.D. and H.R. Kaback, Thiol cross-linking of transmembrane domains IV and V in the lactose permease of Escherichia coli. Biochemistry, 2000. 39(20): p. 6130-5. 254. Sun, J. and H.R. Kaback, Proximity of periplasmic loops in the lactose permease of Escherichia coli determined by site-directed cross-linking. Biochemistry, 1997. 36(39): p. 11959-65. 255. Li, J., et al., Reactions of cysteines substituted in the amphipathic N-terminal tail of a bacterial potassium channel with hydrophilic and hydrophobic maleimides. Proc Natl Acad Sci U S A, 2002. 99(18): p. 11605-10. 256. Struthers, M., et al., Tertiary interactions between the fifth and sixth transmembrane segments of rhodopsin. Biochemistry, 1999. 38(20): p. 6597-603. 257. Rice, W.J., N.M. Green, and D.H. MacLennan, Site-directed disulfide mapping of helices M4 and M6 in the Ca2+ binding domain of SERCA1a, the Ca2+ ATPase of fast twitch skeletal muscle sarcoplasmic reticulum. J Biol Chem, 1997. 272(50): p. 31412-9. 258. Tomoyasu, T., et al., Genetic dissection of the roles of chaperones and proteases in protein folding and degradation in the Escherichia coli cytosol. Mol Microbiol, 2001. 40(2): p. 397-413. 259. Schnaitman, C.A., Solubilization of the cytoplasmic membrane of Escherichia coli by Triton X-100. J Bacteriol, 1971. 108(1): p. 545-52. 260. Miroux, B. and J.E. Walker, Over-production of proteins in Escherichia coli: mutant hosts that allow synthesis of some membrane proteins and globular proteins at high levels. J Mol Biol, 1996. 260(3): p. 289-98. 261. Wagner, S., et al., Tuning Escherichia coli for membrane protein overexpression. Proc Natl Acad Sci U S A, 2008. 105(38): p. 14371-6. 262. Yuan, J., G.J. Phillips, and R.E. Dalbey, Isolation of cold-sensitive yidC mutants provides insights into the substrate profile of the YidC insertase and the importance of transmembrane 3 in YidC function. J Bacteriol, 2007. 189(24): p. 8961-72. 263. Jacob, J., H. Duclohier, and D.S. Cafiso, The role of proline and glycine in determining the backbone flexibility of a channel-forming peptide. Biophys J, 1999. 76(3): p. 1367-76. 264. Bright, J.N., et al., Conformational dynamics of helix S6 from Shaker potassium channel: simulation studies. Biopolymers, 2002. 64(6): p. 303-13. 265. Kumazaki, K., et al., Structural basis of Sec-independent membrane protein insertion by YidC. Nature. 266. Imhof, N., A. Kuhn, and U. Gerken, Substrate-dependent conformational dynamics of the Escherichia coli membrane insertase YidC. Biochemistry. 50(15): p. 3229- 39. 267. Winterfeld, S., et al., Substrate-induced conformational change of the Escherichia coli membrane insertase YidC. Biochemistry, 2009. 48(28): p. 6684-91. 268. Dalbey, R.E., et al., The membrane insertase YidC. Biochim Biophys Acta. 139
269. Duong, F. and W. Wickner, The SecDFyajC domain of preprotein translocase controls preprotein movement by regulating SecA membrane cycling. EMBO J, 1997. 16(16): p. 4871-9. 270. Tsukazaki, T., et al., Structure and function of a membrane component SecDF that enhances protein export. Nature, 2013. 474(7350): p. 235-238. 271. Scotti, P.A., et al., YidC, the Escherichia coli homologue of mitochondrial Oxa1p, is a component of the Sec translocase. EMBO J., 2000. 19: p. 542. 272. Economou, A., et al., SecA membrane cycling at SecYEG is driven by distinct ATP binding and hydrolysis events and is regulated by SecD and SecF. Cell, 1995. 83(7): p. 1171-81. 273. Chatzi, K.E., et al., SecA-mediated targeting and translocation of secretory proteins. Biochim Biophys Acta. 1843(8): p. 1466-74. 274. Kusters, I. and A.J. Driessen, SecA, a remarkable nanomachine. Cell Mol Life Sci. 68(12): p. 2053-66. 275. Schiebel, E., et al., Delta mu H+ and ATP function at different steps of the catalytic cycle of preprotein translocase. Cell, 1991. 64(5): p. 927-39. 276. Lill, R., et al., SecA protein hydrolyzes ATP and is an essential component of the protein translocation ATPase of Escherichia coli. EMBO J, 1989. 8(3): p. 961-6. 277. Serek, J., et al., Escherichia coli YidC is a membrane insertase for Sec-independent proteins. EMBO J., 2004. 23: p. 294. 278. Price, C.E., et al., In vitro synthesis and oligomerization of the mechanosensitive channel of large conductance, MscL, into a functional ion channel. FEBS Lett. 585(1): p. 249-54. 279. Eick-Helmerich, K. and V. Braun, Import of biopolymers into Escherichia coli: nucleotide sequences of the exbB and exbD genes are homologous to those of the tolQ and tolR genes, respectively. J Bacteriol, 1989. 171(9): p. 5117-26. 280. Cascales, E., R. Lloubes, and J.N. Sturgis, The TolQ-TolR proteins energize TolA and share homologies with the flagellar motor proteins MotA-MotB. Mol Microbiol, 2001. 42(3): p. 795-807. 281. Braun, V., Energy-coupled transport and signal transduction through the gram- negative outer membrane via TonB-ExbB-ExbD-dependent receptor proteins. FEMS Microbiol Rev, 1995. 16(4): p. 295-307. 282. Cascales, E., et al., Proton motive force drives the interaction of the inner membrane TolA and outer membrane pal proteins in Escherichia coli. Mol Microbiol, 2000. 38(4): p. 904-15. 283. Riechmann, L. and P. Holliger, The C-terminal domain of TolA is the coreceptor for filamentous phage infection of E. coli. Cell, 1997. 90(2): p. 351-60. 284. Heilpern, A.J. and M.K. Waldor, CTXphi infection of Vibrio cholerae requires the tolQRA gene products. J Bacteriol, 2000. 182(6): p. 1739-47. 285. Clavel, T., et al., Expression of the tolQRA genes of Escherichia coli K-12 is controlled by the RcsC sensor protein involved in capsule synthesis. Mol Microbiol, 1996. 19(1): p. 19-25.
140
286. Lazzaroni, J.C., et al., The Tol proteins of Escherichia coli and their involvement in the uptake of biomolecules and outer membrane stability. FEMS Microbiol Lett, 1999. 177(2): p. 191-7. 287. Llamas, M.A., J.L. Ramos, and J.J. Rodriguez-Herva, Mutations in each of the tol genes of Pseudomonas putida reveal that they are critical for maintenance of outer membrane stability. J Bacteriol, 2000. 182(17): p. 4764-72. 288. Bernadac, A., et al., Escherichia coli tol-pal mutants form outer membrane vesicles. J Bacteriol, 1998. 180(18): p. 4872-8. 289. Balsalobre, C., et al., Release of the type I secreted alpha-haemolysin via outer membrane vesicles from Escherichia coli. Mol Microbiol, 2006. 59(1): p. 99-112. 290. Gaspar, J.A., et al., Surface expression of O-specific lipopolysaccharide in Escherichia coli requires the function of the TolA protein. Mol Microbiol, 2000. 38(2): p. 262-75. 291. Vines, E.D., et al., Defective O-antigen polymerization in tolA and pal mutants of Escherichia coli in response to extracytoplasmic stress. J Bacteriol, 2005. 187(10): p. 3359-68. 292. Gerding, M.A., et al., The trans-envelope Tol-Pal complex is part of the cell division machinery and required for proper outer-membrane invagination during cell constriction in E. coli. Mol Microbiol, 2007. 63(4): p. 1008-25. 293. Kampfenkel, K. and V. Braun, Membrane topologies of the TolQ and TolR proteins of Escherichia coli: inactivation of TolQ by a missense mutation in the proposed first transmembrane segment. J Bacteriol, 1993. 175(14): p. 4485-91. 294. Vianney, A., et al., Membrane topology and mutational analysis of the TolQ protein of Escherichia coli required for the uptake of macromolecules and cell envelope integrity. J Bacteriol, 1994. 176(3): p. 822-9. 295. Lewin, T.M. and R.E. Webster, Membrane insertion characteristics of the various transmembrane domains of the Escherichia coli TolQ protein. J Biol Chem, 1996. 271(24): p. 14143-9. 296. Traxler, B. and C. Murphy, Insertion of the polytopic membrane protein MalF is dependent on the bacterial secretion machinery. J. Biol. Chem., 1996. 271(21): p. 12394-400. 297. Oliver, D.B., et al., Azide-resistant mutants of Escherichia coli alter the SecA protein, an azide-sensitive component of the protein export machinery. Proc. Natl. Acad. Sci. U S A, 1990. 87(21): p. 8227-31. 298. Schuenemann, T.A., V.M. Delgado-Nixon, and R.E. Dalbey, Direct evidence that the proton motive force inhibits membrane translocation of positively charged residues within membrane proteins. J. Biol. Chem., 1999. 274(11): p. 6855-64. 299. Ito, K. and Y. Akiyama, In vivo analysis of integration of membrane proteins in Escherichia coli. Mol Microbiol, 1991. 5(9): p. 2243-53. 300. Stiegler, N., R.E. Dalbey, and A. Kuhn, M13 procoat protein insertion into YidC and SecYEG proteoliposomes and liposomes. J Mol Biol. 406(3): p. 362-70. 301. Miller, J.H., Experiments in Molecular Genetics. 1972, Cold Spring Harbor, NY: Cold Spring Harbor Laboratory. pp. 431.
141