|Table of Contents|
Table of Contents
List of Figures
Chapter 1. Introduction: Actin-based motility of bacterial pathogens
Chapter 2. Recognition of two classes of oligoproline sequences in profilin-mediated acceleration of Actin-based Shigella motility
Chapter 3. Vinculin proteolysis unmasks an ACTA homologue for Actin-based Shigella motility
Chapter 4. Preliminary studies of host cell vinculin and Shigella ICSA binding interactions
Chapter 5. Summary and future directions
List of references
ACTIN-BASED MOTILITY: FUNCTIONAL ANALYSIS OF PATHOGENIC
BACTERIA ANT) HOST CELL PROTEINS
WILLIAM L. ZEILE
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 1997
The work and the contribution to science that this dissertation represents would not exist if it were not for my family. This is dedicated to Patricia Zeile, my wife and our four wonderful children; Kathleen Ruth Zeile, James Walter Zeile, Rachel Euseba Zeile, and Joshua William Zeile. Their joy for living, patience, and love has been an unending source of support and encouragement.
It is a true joy to learn at the hands of great leaders in science. I would liketo give very special thanks to my co-mentors, Dr. Frederick Southwick and Dr. Daniel Purich for allowing me the opportunity to work in their laboratories and study under their direction. I would also like to thank my other committee members: Dr. Gudrun Bennett, Dr. Brian Cain, and Dr. Charles Allen, for their interest and creative suggestions. I would like to especially thank Dr. Fan Kang and Dr. Ron Laine for their help and friendship and for being daily reminders for what is best in experimental science.
Finally, I would like to thank, Patty, my wife of 19 years. We have been through an incredible life journey during those years. At each stage, from the Peace Corps, to dairy farming, to graduate school, Patty has been a true partner. With her clever mind, good sense of humor, love, and encouragement, Patty, who has always recognized what made a fulfilling and happy life, has shared equally with me in reaching our career and life goals. I would like to thank my children, Katie, James, Rachel, and Josh, for all their love, help, and encouragement.
TABLE OF CONTENTS
ACKNOWLEDGMENTS .......................................................................... iii
L IST O F FIG U R E S ................................................................................. vi
A B ST R A C T ........................................................................................ viii
1 INTRODUCTION: ACTIN-BASED MOTILITY OF BACTERIAL
PA T H O G E N S .................................................................................... 1
Specific A im s and Findings ..................................................................... 1
Background and Significance ............................................................... 3
Actin and the Actin-Based Cytoskeleton .............................................. 4
Actin Binding Proteins Important in Actin-Based Motility ........................... 6
Life Cycle and Actin-Based Motility of Intracellular Pathogens ........................ 8
A model for Actin-Based Motility of Bacterial Pathogens ......................... 13
Sum m ary ..................................................................................... 15
2 RECOGNITION OF TWO CLASSES OF OLIGOPROLINE SEQUENCES IN
PROFILIN-MEDIATED ACCELERATION OF ACTIN-BASED SHIGELLA
M O T IL IT Y .................................................................................... 16
Introduction ..................................................................................... 16
Materials and Methods ...................................................................... 21
M aterials .................................................................................... 2 1
Tissue Culture Methods and Infection Procedures ..................................... 21
Microscopy and Microinjection ........................................................ 22
R esults ..................................................................................... ... 27
Characteristics of Shigella Movement and Actin Rocket-Tail Formation in
P tK 2 C ells .............................................................................. 27
Fluorescence Staining of Actin and Alpha-Actinin in Shigella Rocket-Tails ........ 27
Arrest of Shigella Intracellular Movement by the Second Oligo-Proline
Repeat Analogue in Listeria ActA Protein ....................................... 29
Effect of Microinjecting a Binary Solution of Profilin and ActA Analogue on
Shigella Intracellular Movement ................................................... 29
Effects of Microinjection of a VASP Oligoproline Analogue Alone and in
Combination with Profilin on Shigella Intracellular Motility ................... 32
D iscussion ................................................................................ . .. 36
3 VINCULIN PROTEOLYSIS UNMASKS AN ACTA HOMOLOGUE FOR
ACTIN-BASED SHIGELLA MOTILITY ............................................... 43
Introduction .................................................................................. 43
Materials and Methods ...................................................................... 48
M aterials .................................................................................... 4 8
A ntibodies .................................................................................. 4 8
Purification of Recombinant IcsA ..................................................... 49
Affinity Chromatography ............................................................... 50
Solid Phase Overlay Assay ........................................................... 50
Microinjection Experiments ............................................................ 51
Vinculin Proteolysis after Shigella Infection of PtK2 Cells ........................ 51
Indirect Immunofluorescence Microscopy .............................................. 52
R esults ..................................................................................... ... 54
Anti-ActA-peptide Immunofluorescence Microscopy localizes a CrossReactive Protein at the Back of Moving Bacteria ................................ 54
IcsA Affinity Chromatography and Solid Phase Binding Assays Fail to
Isolate or Identify an ActA Mammalian Homologue ........................... 55
Identification of a Cleaved Form of Vinculin, p90 as the ActA Mammalian
H om ologue .......................................................................... 55
Vinculin's Head Domain Localizes to the Surface of Intracellular Shigella ......... 59 The Vinculin Head-Fragment is Generated after Shigella Infection ................. 61
D iscussion .................................................................................... 63
4 PRELIMINARY STUDIES OF HOST CELL VINCULIN AND SHIGELLA
ICSA BINDING INTERACTIONS ......................................................... 69
Introduction .................................................................................. 69
Materials and Methods ...................................................................... 71
Cloning of Glutathione S-Transferase Fusion Protein cDNA Expression
C onstructs ........................................................................... 7 1
Recombinant Protein Purification ..................................................... 72
In vitro Solution Phase Binding Assay: Bead Binding Assay ................... 73
R esults ..................................................................................... ... 74
Construction and Expression of GST-IcsA Fusion Proteins ...................... 74
Solution Phase Binding Assay by Co-Preciptiation with Glutathione
Sepharose B eads .................................................................... 81
Binding of Vinculin to Shigella IcsA Requires the Glycine-Rich Region of
IcsA and Cleavage of Vinculin to the p90 Form ................................. 81
D iscussion .................................................................................... 85
5 SUMMARY AND FUTURE DIRECTIONS ................................................ 89
In vitro Expression Cloning to Isolate a Specific Protease ............................. 90
Biopanning a Phage Display Library ..................................................... 92
Biopanning a Phage Display Library Based on a Biased Oligonucleotide
S equ en ce .................................................................................... 9 3
Another Model System: Vaccinia Virus .................................................. 93
The Actin-Based Motility Complex ...................................................... 95
LIST OF REFERENCES .......................................................................... 99
BIOGRAPHICAL SKETCH ..................................................................... 106
LIST OF FIGURES
1.1. The two steps of actin filament assembly...................................................5
1.2. Intracellular Life-cycle of Listeria ...........................................................9
1.3. Schematic model of ActA/VASP/profilin interactions leading to the generation of
A TP-actin m onom ers............................ ........................................... 14
2.1. Phase images of Shigella rocket tails...................................................... 20
2.2. Comparison of bodipy-phallacidin staining of Shigella and Listeria actin filament
rocket tails ................................................................................ 24
2.3. Anti-alpha actinin immunofluorescence and bodipy-phallacidin fluorescence images
of Shigella rocket tail ....................................................................... 26
2.4. Shigella movement and actin rocket tail formation in PtK2 host cells before and after
microinjection of the synthetic ActA peptide............................................ 28
2.5. Effects of microinjection of the ActA analogue on Shigella motility ................. 30
2.6 A. Time-lapse phase micrograph of Shigella motility in a PtK2 cell before and after
microinjection of an ActA analogue/profilin binary solution............................ 31
2.6 B. Velocities of two bacteria in a PtK2 cell before and after the microinjection of an
ActA analogue/profilin binary solution.................................................. 32
2.7. The effect of microinjection of VASP peptide analogue on Shigella motility in a
P tK 2 cell ................................................................ ..................... 33
2.8. Microinjection of poly-L-proline, profilin and mixtures of VASP analogue and
profilin and mixtures of poly-L-proline and profilin on Shigella motility............. 35
2.9. Working model showing the primary components likely to be involved in the actinbased locomotory unit of Shigella ........................................................42
3.1. Characteristics of anti-ActA antibody ................................................... 47
3.2. Structural organization of human vinculin................................................ 53
3.3. Immunofluorescence microscopy of Shigella-infected PtK2 cells using
anti-vinculin antibody.................................................................... 58
3.4. Effect of Vinc-1 peptide on Shigella speed............................................. 60
3.5. Shigella infection induces the production of the 90 kDa vinculin head-fragment ... 62
3.6. A working model for vinculin proteolysis and assembly of the Shigella actinbased motility complex .................................................................... 68
4.1. Schematic representation of Shigella IcsA .............................................. 74
4.2. Schematic of GST-IcsA fusion proteins used in this study and expression in
E co li .......................................................................................... 7 7
4.3. Purification of GST-IcsA fusion proteins .............................................. 80
4.4. Vinculin p90 binds to full length IcsA and to glycine-rich domain of IcsA ............ 83
5.1. Schematic of Actin-Based Motility Complexes of Listeria and Shigella actin
m otors ..................................................................................... . 98
Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy
ACTIN-BASED MOTILITY: FUNCTIONAL ANALYSIS OF PATHOGENIC BACTERIA AND HOST CELL PROTEINS By
William L. Zeile
Chairman: Frederick Southwick, M.D.
Major Department: Biochemistry and Molecular Biology
The molecular architecture of the eukaryotic cell is, in part, determined by the actinbased cytoskeleton. Understanding the dynamics of this structure is key to understanding many cellular processes and may lead to discoveries for the treatment of diseases that are caused by a breakdown of the control and regulation of actin polymerization. The intracellular bacterial pathogens, Listeria monocytogenes and Shigellaflexneri, during infection of host cells remodel the actin microfilaments; this remodeling supports bacterial motility and spread from cell to cell. We have used the infection of host cells by these organisms as a model system for studying the molecular mechanism of actin-based motility. In so doing we have addressed three hypothesis: 1) Listeria and Shigella share a similar mechanism of actin assembly and parasitism of the host by these organisms requires host cell proteins in addition to host cell actin. 2) For actin-based motility of Shigella, an additional host cell adapter protein is essential. 3) This host cell adapter protein must be activated by proteolysis during infection for productive assembly of the Shigella actin motor. Microinjection of infected mammalian tissue culture cells using specific competitor peptides designed to mimic the binding domains of the Listeria ActA surface protein and the host cell protein, vasodilator-stimulated phosphoprotein (VASP), uncoupled the actin
motors of Listeria and Shigella. These experiments demonstrated a shared mechanism of actin assembly by these bacteria. The surface protein IcsA of Shigella required for actinbased motility has no homology to the Listeria ActA protein yet their mechanisms are similar. We determined that a host cell adapter protein containing ActA-like sequences was required for Shigella motility, and by immunological methods we discovered the cytoskeletal protein vinculin to be this protein. For productive assembly of the actin motor of Shigella, vinculin is proteolyzed to an active state in which the ActA-like sequence is unmasked for binding VASP. Activation of vinculin by a specific protease is likely the result of an apoptotic cascade initiated by the bacterium upon infection. Assembly of a competent actin motor may involve an ordered mechanism; understanding this should teach us how the cell regulates and assembles higher order molecular structures.
INTRODUCTION: ACTIN-BASED MOTILITY OF BACTERIAL PATHOGENS Specific Aims and Findings
We would like to understand how during infection of host cells, the bacterial
pathogens Listeria monocytogenes and Shigellaflexneri take over the actin cytoskeleton, and direct the polymerization of actin. By doing so, they are able to move within the cytoplasm and spread from cell to cell. The ability to spread from cell to cell explains many of the unique clinical manifestations of listeriosis and shigellosis and this step is absolutely required to cause disease. Both of these intracellular pathogens spread from cell to cell by inducing a host cell protein, actin, to polymerize into filaments or "rocket tails", which propel these organisms through the cytoplasm to the peripheral membrane. Here, the rocket tails drive the bacteria into filopods (membrane protrusions) which can be ingested by adjacent cells, allowing the organisms to pass from cell to cell without being exposed to the humoral immune system or extracellular antibiotics. This strategy explains, in particular, Listeria's predisposition to infect individuals with impaired cell-mediated immunity including pregnant women, neonates, organ transplant patients, and persons with AIDS. This same strategy also explains the rapid colonization and resulting devastation of the intestinal epithelial cell layer by Shigella during infection.
The first aim of this research was to address the hypothesis that Listeria and
Shigella share a common component or components in the mechanism of induced actin polymerization or rocket tail formation. Recent studies in our laboratory and others had shown that Listeria's ability to attract VASP from its normal host cell binding sites is the first step in usurping host cell actin assembly (Chakraborty et al., 1995). The oligoproline sequence repeat FEFPPPPTDE of Listeria's ActA surface protein, hereafter known as an
ABM-1 site (for Actin Based Motility 1)(Purich and Southwick, 1997), serves as the critical VASP-binding site. If this sequence was important for Listeria, might a similar docking site be found on a protein necessary for Shigella motility as well? Using synthetic peptides specific to the above sequence, control peptides, and purified cytoskeletal components, microinjection experiments were carried out which demonstrated a shared mechanism, based on the ABM- 1 site, for actin-based motility by Listeria and Shigella.
The second aim of this research was to investigate the hypothesis that the
mammalian host cell provides a component containing an ABM- 1 site that Shigella usurps during infection. We would identify this ABM-1 containing component and try to understand the mechanism by which this host cell protein could become available for actin assembly during infection by Shigella. The cytoskeletal protein, vinculin, was identified as this component essential for Shigella motility. It had the critical ABM-1 site and the mechanism by which it becomes available to the bacterium during infection was found to be cleavage by limited proteolysis. This cleavage results in the unmasking of the critical ABM- 1 site necessary for VASP binding.
The hypothesis that vinculin, when activated by proteolysis, directly binds to the surface protein of Shigella, IcsA, during assembly of the actin-based motor was addressed as the third and final aim of this research. In vitro binding studies and chemical crosslinking demonstrated that host cell vinculin binds IcsA with high affinity and that this protein-protein interaction takes place within the amino terminal glycine-rich region of IcsA. These studies reinforced our earlier findings and demonstrated that vinculin must be activated by limited proteolysis for productive binding to Shigella IcsA.
This research has allowed us to dissect the cascade of protein binding interactions required to generate Shigella's and Listeria's actin-based motor, adding to the knowledge of how virulent pathogenic bacteria invade and spread cell to cell. This promises to advance preventive or remedial measures for diseases caused by these organisms.
Background and Significance
Bacterial infections of humans in which the pathogenic organism invades and
multiplies intracellularly are a leading cause of disease and death in the world population. The bacteria of these infections not only colonize the external linings of the gut and airways, but directly invade the epithelial cells that line these passages causing inflammation and abscesses. Depending on the pathogen these organisms can spread cell to cell, exit the host cells, migrate through the extracellular matrix, and enter the circulation. Some members of this special class of intracellular bacteria are Mycobacterium leprae, the cause of leprosy with 12 million individuals affected worldwide; Mycobacterium tuberculosis, the cause of tuberculosis with 10 million new cases per year; and Chlamydiae trachomatis, a leading cause of blindness. Other members of this class of parasites cause Legionnaire's disease, salmonellosis, psittacosis, tularemia, brucellosis, listeriosis, and shigellosis (Cossart and Mengaud, 1995).
Of these, the last two are responsible for increased morbidity and mortality in the world. Shigellosis is estimated to cause over 100 million cases of dysentery and over 600,000 deaths a year (Gyles, 1993). Gram negative bacteria of the Shigella genus are found in the intestinal tracts of man and animals and are readily spread by fecal contact. Shigellosis is a deadly dysentery of children and the elderly, especially in the less developed countries of the world, where death from the infection comes mainly from dehydration.
Listeriosis to a healthy individual poses little threat, but to pregnant women, the
elderly, and the immuno-compromised (such as AIDS patients), the disease is serious and can lead to bacterial meningitis and death. Listeria monocytogenes, a gram positive rod, the only member of the Listeria genus to infect humans, is a common organism found in soil, water, plants, and the intestinal tracts of many animals. This organism as a contaminant of food is of increasing concern to public health officials as a Listeria infection
has a 23% mortality as opposed to infections from other food-borne organisms which are rarely fatal (Southwick and Purich, 1996).
Understanding the mechanisms by which intracellular pathogens invade and infect man will lead to methods for prevention and cure of the diseases caused by these parasites. To the cell biologist studying the cytoskeleton, the bacteria Listeria monocytogenes and Shigellaflexneri open a unique window into the mechanisms of cytoskeletal remodeling that are operating in many eukaryotic cell types. These intracellular bacteria are members of a distinct subclass of invasive bacteria that upon entry into a host cell are able 1) to usurp the host cell cytoskeleton, 2) to support motility through the cytoplasm and, 3) to invade neighboring cells. This unique adaptation of remaining intracellular bacteria throughout most of their life cycle allows these bacteria to multiply and spread undetected by the host humoral immune system. What is required of the bacteria is the ability to stimulate actin polymerization in an ordered and directed manner to allow the polymers of actin, or microfilaments, to form a solid support in the cytoplasm by which the bacteria can be propelled forward. The obvious question is how do the bacteria accomplish this? Will understanding actin-based motility of bacteria add to our understanding of actin-based cytoskeletal remodeling in highly motile cells such as macrophages and neutrophils and in less motile, but equally dynamic, cells such as fibroblasts and epithelial cells? Much is known of actin dynamics in vitro and many actin associated proteins have been identified and been given functions in the cell. Very little is known about the mechanism of actin assembly in the cell, how that process is regulated, and what part all the various actin associated proteins play in assembling and maintaining the microfilament cytoskeleton. It is for these reasons that we study the actin-based motor of Listeria and Shigella as a model system.
Actin and the Actin-Based Cytoskeleton
The cytoskeleton of eukaryotic cells is made up of three main structural elements: the microtubules, the intermediate filaments, and the microfilaments. Whereas the
microtubules form a framework for molecular transport in the non-mitotic cell and a framework for chromosomal transport in the mitotic cell, and the intermediate filaments form a stable, less dynamic scaffold, the microfilaments support a highly dynamic architecture for cellular remodeling and movement. The microfilaments are polymers of the 43 kD protein, actin, and are found in muscle as well as non-muscle cells. Actin exists as a globular monomer (G-actin) or in filaments (F-actin) in which the monomers are assembled in a head to tail arrangement. Although there are different isoforms in different cell types, monomeric actin has a primary sequence of 375 residues and in tertiary structure is made up of a large and small domain separated by a cleft which serves as a nucleotide binding site and cation binding site (Kabsch et al., 1990). Nucleotide binding to actin is either in the form of adenosine triphosphate (ATP) or adenosine diphosphate (ADP). The myosin heavy chain subfragment S I binds to the small domain of actin and in muscle this binding supplies the necessary contact in muscle contraction.
NU--la i on Eln i ,,',,n
ATUP-A cti n )
onoer > T>r I> E>
S R ep ef iti've
: rFu'k rhomer ,,
Tri ner r
Figure 1.1. The two steps of actin filament assembly
The S I subfragment has been used in electronmicroscopy to decorate actin
filaments which confirmed kinetic studies that postulated two distinct ends to an actin filament. The filament ends have been defined as the plus end or barbed end (from S 1 decoration) and the minus end or pointed end (Figure 1.1). The plus end is the more kinetically active end and has a KD for ATP-actin monomers of 0.15 uM, the minus end is
less active with a KD of 0.5 uM. Both ends have a KD of 6.0 uM for ADP-actin. Actin polymerization is a two step process (Figure 1.1) in which the initial rate limiting step, formation of the unstable nucleus trimer, is followed by rapid filament growth from both ends until the monomer concentration is depleted below the critical concentration for that end or the ends become capped. Once the ATP-actin monomer is incorporated into a filament the ATP is hydrolyzed to ADP, leaving dynamic actin filaments comprised of mostly ADP-actin in which the energy of hydrolysis is stored within the structure of the filament. The unequal kinetics of the two ends for ATP-actin allow for very precise control of the polymerization process, not only allowing for controlled and directed polymerization to take place when the appropriate signal is received, but also controlled depolymerization or disassembly by filament severing and/or uncapping. Actin Binding Proteins Important in Actin-Based Motility
In addition to the kinetic regulation of actin polymerization, further control of polymerization of actin in non-muscle cells rests with the actin binding proteins. The function of some of these critical proteins are described briefly below: monomer sequestering proteins, such as thymosin 34 and profilin; capping proteins which block the plus end, such as villin, gelsolin, CapZ, and CapG; and severing proteins which cleave filaments, such as gelsolin and severin. Some of these actin binding proteins are responsive to calcium and phosphatidylinosititol bis-phosphate (PIP 2), and therefore control of polymerization is responsive to external as well as internal cell signals.
Electron micrographs of nonmuscle cells show that actin filaments are organized into networks where many filaments appear to cross each other at right angles, forming an orthogonal mesh. The cross-linking protein ABP-280, or filamin, is responsible for organizing the right angle networks (Stossel, 1993). In many cell types, parallel bundles of actin filaments can also be seen by electron microscopy. The 105 kDa protein cx-actinin binds to the sides of actin filaments and can link actin filaments into parallel arrays or
bundles. The 120 kDa protein vinculin is a key component of adherins junctions and focal contacts, and it binds and concentrates a-actinin in this region. Vinculin has a 90 kDa head region containing the a-actinin binding site and a 30 kDa tail region (Price et al., 1989). The two regions are linked by three oligoproline sequences, the first of which is homologous to the Listeria surface protein ActA (see below). Protease treatment cleaves the tail from the head. The head region can fold over and bind to the tail, covering up many of the binding domains (Johnson & Craig 1995). It is likely that in the cell, association and dissociation of the head and tail regions serve as a regulatory switch involved in forming focal contacts and assembly of new actin filaments.
Another class of actin regulatory proteins are the monomer binding proteins. Even under conditions expected to polymerize actin completely, unstimulated cells contain high concentrations of monomeric actin. In polymorphonuclear (PMN) leukocytes, for example, 60-70% (or about 200 [tM) of the total actin exists in a monomeric form (Southwick and Young, 1990). It is therefore important for the cell to contain proteins which directly regulate actin monomer function. Thymosin B4, a 5 kDa polypeptide, may account for the high actin monomer concentrations in nonmuscle cells. A second monomer-binding protein is the 15 kDa protein, profilin. This host cell actin regulatory protein is likely to play a key role in bacterial actin-based motility. Profilin also binds G-actin as a one-to-one complex and displays a dissociation constant of 1-10 RM. Profilin catalyzes the exchange of ADPactin with unbound ATP to form ATP-actin (Mockrin & Korn, 1980, GoldschmidtClermont et al. 1991). Because ATP-actin has a higher affinity for actin filament ends, the facilitation of nucleotide exchange should enhance actin filament assembly. Profilin may also serve to enhance the affinity of the barbed ends for actin monomers and stimulate actin assembly by this mechanism (Carlier and Pantaloni, 1997). Finally, profilin binds to poly L-proline (Tanaka & Shibata, 1985), and this binding characteristic is likely to play an important role in concentrating the protein in locations where new actin filaments assemble (Southwick and Purich, 1994a).
The employment of microfilaments and another special class of actin binding proteins supplies the contractile mechanism of muscle cells as well as nonmuscle cells. These proteins are members of the myosin family. In muscle cells, myosin II forms multimers with moveable head domains, which by binding and release (using the energy of ATP hydrolysis) move actin filaments during muscle contraction. In nonmuscle cells, myosin I, acting as a monomer, moves microfilaments using similar contractile processes.
As alluded to earlier, the control of actin assembly and disassembly is necessary to prevent unwanted polymerization and to direct polymerization and depolymerization, both spatially and temporally. Recently, a group of identified monomeric GTPases have been shown to be important in these actin regulatory processes. These GTPases, of which Ras is the prototype, act as molecular switches that are turned on by binding GTP and turned off by GTP hydrolysis. The stimulus for activation comes from a signal cascade initiated at a cell surface receptor. The GTPases can be linked themselves, in signal transduction pathways in which downstream members are stimulated to specifically activate multiple signal pathways (Chant and Stowers, 1995). Identified in the control of the actin cytoskeleton are three GTPases: Rho controls microfilament bundling to produce filopodia and stress fibers; Rac controls microfilament crosslinking to produce lamellipodia extensions; and Cdc42 acts upstream in the pathway to control Rac (Nobes and Hall, 1995). Recently using mutation studies and genetic analysis in Caenorhabditis elegans, Rho was shown to be necessary in cell migration and for nerve cell axon targeting (Zipkin et al., 1997). It is certain other GTPases will be added to this group as more is learned about the regulation of the actin cytoskeleton. Life Cycle and Actin-Based Motility of Intracellular Pathogens
How do Shigella and Listeria gain entry into the host cell's cytoplasm and why is it especially important that cell-mediated immunity is critical for protecting the host against Listeriosis? Studies of Shigella and Listeria infection in tissue culture cells provide a new understanding of the stages in these infections and help to answer these questions. As
shown in Figure 1.2, the organisms are first phagocytosed by neutrophils, enterocytes, macrophages (Dabiri et al. 1990), epithelial cells, and endothelial cells, as well as fibroblasts. Listeria produces the 80 kDa surface protein internalin that enhances bacterial attachment and internalization (Gaillard et al. 1991). Performing the same function in Shigella are the IpaB, IpaC, and IpaD proteins (Menard et al., 1994).
The bacterium then becomes enclosed in a subcellular compartment called the
phagolysosome, a normally hostile and toxic environment for most bacteria. The low pH of this compartment, however, activates listeriolysin-O, a pore forming toxin, of Listeria
proliferatioon/ --- .
spread ; "s
Figure 1.2. Intracellular Life-cycle of Listeria
or a hemolysin of Shigella to lyse the phagolysosome, allowing escape of the bacteria into the cytoplasm. Once within the cytoplasm, the bacteria double every 60 minutes (Dabiri et al., 1990). They become surrounded by actin, and about two hours later, actin filaments begin extending from one end of the bacterium, propelling the organism through the cytoplasm (Dabiri et al.,1990, Tilney and Portnoy, 1989). Many bacteria subsequently induce the formation of filopods which are stabilized actin bundles enveloped in membrane resulting in elongated membrane protrusions, each containing a bacterium at their tip. These
filopods are ingested by adjacent cells, and a characteristic double membrane phagolysosome can be observed (Tilney and Portnoy 1989). After avoiding the hostile extracellular environment of immunoglobulins, complement, and extracellular antibiotics, the life cycle then begins over. This unusual intracellular life style helps to explain many of the unique clinical characteristics of Listeria and Shigella infection (Southwick and Purich, 1996).
A large number of experiments have shown that Shigella and Listeria use a
mechanism of actin polymerization to move through the cytoplasm of infected host cells. These experiments made use of inhibitors of actin polymerization such as cytochalasin D to show that blocking actin assembly also blocked bacterial motility (Dabiri et al., 1990). Additionally, epifluorescence experiments using the fungal peptide, phalloidin, which preferentially binds microfilaments, conjugated to a fluorescent dye, stained actin filaments comprising the rocket tail of moving bacteria. Bacterial F-actin tails can also be clearly visualized by phase-contrast microscopy as dynamic structures that appear as rocket-like tails behind the moving organism.
Mutant strains of Listeria and Shigella have been isolated or generated by
transposon mutagenesis, in which single genes have been shown to be disrupted, and these strains lack actin-based motility. The IcsA gene of Shigella (Bernardini et al., 1989 and Goldberg et al., 1993) and the ActA gene of Listeria (Kocks et al., 1992) were identified to be necessary and sufficient for actin-based motility. IcsAstrains of Shigella and ActAstrains of Listeria are able to invade host cells, but are unable to move within the cytoplasm by actin assembly. ActA and IcsA have been shown to be bacterial surface proteins that are uniquely expressed in a polar orientation from the bacterial membrane (Goldberg et al., 1993).
Immunofluorescence and electron microscopy demonstrate that ActA is expressed on the surface of intracellular Listeria, but is not found in the actin filament rocket tails (Kocks et al. 1992, Niebuhr et al. 1993). The absence of ActA in the actin filament tails is
consistent with in vitro experiments which have shown that the ActA protein does not bind directly to actin (Kocks et al. 1992); nor does ActA stimulate the polymerization of purified actin (Tilney et al. 1990). Examination of the amino acid sequence of ActA reveals four very unusual oligoproline sequence repeats, located at residues 235- DFPPPPTDE, 269FEFPPPPTDE, 304-FEFPPPPTED, and 350-DFPPIPTEE (where the number preceding each sequence designates the first amino acid) (Kocks et al. 1992, Domann et al. 1992). ActA DNA has been transfected into mammalian cells and expression of the ActA protein is associated with the assembly of new actin filaments (Pistor et al. 1995). All evidence suggests that ActA is the only Listeria protein required to stimulate actin assembly in host cells. In-frame deletion of the oligoproline repeats markedly reduces ActA-induced actin assembly, emphasizing the functional importance of these regions (Smith et al., 1996).
There has been a growing recognition of the importance of proline rich sequences in mediating protein-protein regulatory interactions. When a sequence of four or more proline residues are bound in a polypeptide a unique extended structure is produced, the polyproline II helix. Much evidence suggests this extended structure of restricted mobility has unique and favorable binding characteristics with binding partners. Proline-rich sequences, because of their highly restricted mobilities, have a relatively low entropy even before binding; and because binding leads to a smaller drop in entropy than would occur for a polypeptide sequence of more flexibility, binding to proline-rich sequences is favored entropically. There is also evidence to suggest that binding to proline-rich sequences is favored enthalpically as well, because prolines form more electron-rich amide bonds. These two characteristics make binding to proline-rich regions more thermodynamically favorable and result in high affinity complexes (Williamson, 1994). Because of the high affinity of such interactions, proline-rich sequences are often involved in complex multiple protein associations. The actin-based motor utilized by Listeria for intracellular motility appears to be a multi-protein complex, and the proline-rich sequences found in ActA are ideally suited for the assembly of such a motor.
Three classes of proline-rich sequences have been defined (Williamson, 1994): (1) Repetitive short proline repeats [example the (AP)6 sequence found in light chain myosin kinase, shown to bind to actin]; (2) Tandem repeats (example the [SYP(P)Q(P)]5 sequence of Dictyostelium actin-binding protein and Squid rhodopsin sequence [PPQY] 10); and (3) Non-repetitive proline rich sequences (example SH3 binding sequences XPXXPPPXP where x represents nonconserved amino acids and a hydrophobic amino acid). The ActA protein falls into the second category representing a tandem repeat. The sequence found in each repeat is unique compared with other classes of proline-rich proteins. The ActA proline repeats contain a preceding aromatic group, followed by a series of prolines that are flanked by a negatively charged amino acid on each side. This sequence now defines a specific consensus sequence found in a number of cytoskeletal proteins critical for actin-based motility. This consensus sequence is based on a register of 4-5 proline residues preceded on the N-terminal side by amino acids containing acidic and aromatic side chains and on the C-terminal side acidic amino acids. These properties define the ABM- 1 (actin-based motility- 1) homology sequence: (D/F) FPPPPX(D/E)(D/E). ABM-1 sites are found in all four repeats of ActA, Vinculin, and zyxin (Purich and Southwick, 1997).
As previously noted, the surface protein IcsA of Shigella had been identified as the protein required for actin-based motility by these organisms (Bernardini et al., 1989). Further, IcsA was demonstrated to be the only protein necessary to support bacterial motility in cytoplasmic extracts by stable expression of IcsA on E.coli, a bacteria that does not normally assemble actin (Goldberg and Theriot, 1995). Although the actin filament dynamics of Listeria and Shigella intracellular motility appear to be similar (our microinjection experiments outlined in Chapter 2 of this dissertation clearly demonstrated this), the protein, IcsA, shares no homologies with ActA and does not contain an ABM-1 site now believed necessary for efficient actin assembly. It is now believed that Shigella
requires an additional host cell protein, or an ActA mammalian homologue, containing an ABM-1 site that is usurped during infection for assembly of Shigella's actin motor. A model for Actin-Based Motility of Bacterial Pathogens
Based on evidence in the literature and our experimental findings we have
proposed a working model for actin assembly by Listeria and Shigella. Shown in Figure
1.3 is a cartoon of the proposed mechanism of Listeria actin assembly. This model describes a mechanism based on recruitment of the host cell protein, VASP, by the bacterium. In the case of Listeria motor assembly, the ActA protein directly binds VASP by its four registers of ABM- 1 sequences. This binding interaction has been demonstrated in vivo by transfection of cDNAs coding for chimeras of ABM- 1 sites, immunofluorescence (Bubeck et al., 1997) and solid phase binding assays (Niebuhr et al., 1997).
Considerable evidence suggests that VASP acts to concentrate profilin or profilin actin complexes at the bacteria-actin tail interface (Kang et al., 1997). VASP and profilin have been shown to co-localize in fibroblasts (Reinhard et al., 1995a) and profilin has been localized to the bacterial-actin tail interface in moving Listeria (Theriot et al., 1994). VASP exists as a tetramer in the cell and contains mutliple GPPPPP registers to which profilin binds. These oligoproline sequences are characterized by the consensus sequence XPPPPP (X=G, A, L, S), now known as the ABM-2 homology sequence (Purich and Southwick, 1997). These sequences have been found in other actin regulatory proteins such as the human Wiscott-Aldrich Syndrome Protein (WASP), the Drosophilia polypeptide, Ena, and its mammalian homolog, Mena. Because VASP as a monomer has four ABM-2 sequences for profilin binding and therefore, as a tetramer in the cell, has a potential of 16 binding sites for profilin, Kang et al. (1997) postulated that the bacteria
0 ADP-ActinA c A
0 ATP-Actin ADP
Figure 1.3. Schematic model of ActA/VASP/profilin in teractions
leading to the generation of ATP-actin monomers. Left lower
comer depicts the bacterial cell wall containing the ActA protein
must form an activated cluster or polymerization zone in which the necessary components, profilin and profilin/actin complexes, are in very high local concentrations. In addition to high local concentrations of polymerization competent actin, the availability of uncapped plus ends and G-actin turnover by actin depolymerization factor, ADF/cofilin, allows rapid actin polymerization to take place at the bacteria-actin filament interface, propelling the bacteria through the cytoplasm (Carlier and Pantaloni, 1997).
The model for Shigella actin-based motility differs only in one component: a host cell adapter molecule. VASP been shown to be concentrated behind moving Shigella (Chakraborty et al., 1995). Our experiments (Zeile et al., 1996) in which Shigella infected cells were microinjected with synthetic peptides of either the ABM-1 or ABM-2 sequence at submicromolar concentrations, which completely inhibited bacterial motility, indicated a mechanism in which VASP played a central role. We demonstrated (Laine et al., 1997) that although no known protein of Shigella had an ABM-1 site for VASP binding, Shigella recruits the host cell vinculin for this purpose. Vinculin contains an ABM-1 sequence that is normally masked as described previously, but this site is exposed by limited proteolysis during the apoptotic state of the infected cell. The cleaved form of vinculin (p90) with its
unmasked VASP binding site binds IcsA of Shigella which now allows assembly of an actin motor identical to Listeria.
The purpose of the research outlined in this dissertation is to add to our
understanding of intracellular pathogenesis by studying the exploitation of the host cell by bacteria. We will show in the following chapters that our contribution has been in three main areas. 1) We now appreciate that the molecular mechanism of actin motility by Shigella and Listeria is a shared mechanism, and that although these two bacteria are evolutionarily divergent they have evolved similar parasitic mechanisms. 2) We know that in addition to the host cell protein, actin, these bacteria exploit other key cytoskeletal host cell proteins during the infection process. And that in usurping critical host components, these bacteria have exploited different cellular processes, such as apoptosis and proteolysis in the case of Shigella. 3) The molecular mimicry, as revealed by these pathogenic organisms, instructs us about the intracellular environment in general the principles learned can be applied to many cellular processes.
It is hoped that revealing the molecular mechanisms by which bacteria using actinbased processes to invade and spread cell to cell will aid in the discovery of more effective treatments for the infections caused by these organisms. In a more general sense, understanding the molecular basis for actin-based motility in Shigella and Listeria will give a better understanding of the actin cytoskeleton and the processes in which it is continually being remodeled in response to intracellular as well as extracellular signals. This will teach us about the causes of specific diseases of motile cells such as leukocytes, which become impaired for movement; and cancers in which cells that are not normally motile become so. We know these processes are actin-based--now we must understand the complex regulatory cascade by which these processes are controlled.
RECOGNITION OF TWO CLASSES OF OLIGOPROLINE SEQUENCES IN
PROFILIN-MEDIATED ACCELERATION OF ACTIN-BASED SHIGELLA MOTILITY Introduction
In the experiments described below, we addressed the hypothesis that although Shigella and Listeria are evolutionarily divergent organisms they have evolved a similar mechanism for usurping the cytoskeleton of the host cell. We chose to test this by microinjecting synthetic peptides generated toward the putative binding sites of Listeria ActA and VASP, and microinjecting the recombinant protein, profilin, into mammalian cells infected with Shigella.
Microinjection of a synthetic peptide (CFEFPPPPTDE) analogue of the second ActA repeat into PtK2 cells infected with Listeria rapidly and completely blocks Listeriainduced actin assembly at a final intracellular peptide concentration of about 80 nM (Southwick and Purich, 1994b). Microinjection of mosquito oostatic factor, the freely occurring decapeptide YDPAPPPPPP, also inhibits Listeria locomotion (Southwick and Purich, 1995). At similar concentrations, both of these peptides result in the loss of the host cell's normal actin filament structure and retraction of the peripheral membrane. Microinjection of a third peptide analogue DFPPPPTDEELRL derived from first oligoproline repeat in ActA also results in membrane retraction and loss of the normal actin filament architecture. These changes were associated with the dissociation of VASP from focal adhesion plaques and redistribution throughout the cytoplasm (Pistor et al, 1995). These peptides, therefore, are likely to block VASP binding to an ActA-like host protein (possibly vinculin) as well as block VASP binding to the oligoproline regions of ActA.
In addition to binding to ActA, VASP also binds to profilin (Reinhard et al., 1995a). Profilin is a key host cell component responsible for Listeria and Shigella locomotion. Depletion of profilin from Xenopus egg extracts, using beads with covalently bound poly-L-proline, blocked in vitro movement of Listeria, and re-addition of profilin partially restored motility (Theriot et al., 1994). Profilin enhances the exchange of ATP on actin monomers (Mockrin and Korn, 1980; Goldschmidt-Clermont et al., 1991) and may produce higher intracellular concentrations of the more polymerization-competent ATPactin at the bacterium/rocket-tail interface (Southwick and Purich, 1994b). In addition, in the presence of the monomer sequestering protein thymosin B4, profilin may interact with the barbed ends of actin filaments to lower the critical concentration for actin assembly (Pantaloni and Carlier, 1993).
Although the first descriptions of actin filaments being associated with intracellular bacteria were reported with Shigella-infected cells (Bernardini et al., 1989), video microscopy experiments similar to those designed to explore actin-based motility in Listeria have not been performed in live cells infected with Shigella. Here we now performed timelapse studies which reveal that Shigella moves at rates and trajectories similar to Listeria, suggesting these two bacteria stimulate actin based motility by similar mechanisms. Shigella like Listeria has an outer cell wall protein, IcsA, which is necessary for actinbased motility (Bernardini et al., 1989; Goldberg et al. 1993) and is sufficient to support actin-based movement in Xenopus egg extracts (Goldberg and Theriot, 1995). This 120 kDa protein, however, shares no sequence identity with the Listeria ActA protein and lacks oligoproline sequences which might recruit host cell components to facilitate actin filament assembly. To test the possibility that the IcsA protein attracts a host cell oligoprolinecontaining protein to serve in place of ActA, we examined intracellular Shigella motility after the microinjection of two oligoproline analogues derived from ActA and VASP amino acid sequences. Cellular ActA analogue concentrations necessary to inhibit Listeria movement (i.e., in the range of 80-800 nM) blocked Shigella motility as well. The
introduction of an oligoproline peptide based on the VASP sequence, (GPPPPP)3, at considerably higher intracellular concentrations (10 [M) also blocked Shigella movement. Microinjection of an equimolar binary solution of profilin with the ActA or the VASP analogue neutralized the inhibition of Shigella movement. Even more surprisingly, the binary solutions caused a 200% to 300% increase in the velocities of intracellular bacterial migration. These findings provide evidence for a shared mechanism involving certain oligoproline-containing proteins and profilin in actin-based motility of both Shigella and Listeria ; they also suggest that a similar mechanism may regulate actin filament assembly at the cytoskeleton-membrane interface of actively moving nonmuscle cells.
Figure 2.1. Phase images of Shigella rocket tails.
(A-C): Formation of a phase dense rocket tail as a Shigella bacterium migrates upward and to the right through a thin region of the cytoplasm in a PtK2 host cell. Images are taken at approximately 30 sec intervals as indicated by the time stamp. (Panel, top to bottom). (DF): Formation of a phase lucent actin rocket tail as the bacterium in the lower right hand corner of image D migrates through the perinuclear region of a PtK2 cell. The bacterium in D has turned to the right in images E and F, and is migrating toward the top of the micrograph. A thin clear area that displaces subcellular organelles trails behind the bacterium and is best seen in image D. Length of time stamp bar = 12 tm.
I- Albl. j-7
Materials and Methods
Peptides were synthesized by the automated Merrifield method in the University of Florida Protein Sequencing Core Laboratory. For microinjection the peptides were diluted to a stock concentration of 1-1.8 mg/mL in sterile PBS (pH 7.2) and the pH of each peptide solution titrated to a pH of 7.2. Bodipy-phallacidin was obtained from Molecular Probes (Eugene, Oregon). Primary anti-vinculin and anti-alpha actinin antibodies and fluoresceinconjugated anti-IgG antibodies were obtained from Sigma (St. Louis, MO). Profilin was purified from human platelets or from supernatants of E. coli expressing recombinant human profilin (pET expression vector in E.coli strain BL21 kindly provided by Dr. S. Almo, Albert Einstein College of Medicine) using a poly-L-proline Sepharose-4B affinity column as previously described (Southwick and Young, 1990). Tissue Culture Methods and Infection Procedures
The PtK2 cell line (derived from the kidney epithelium of the kangaroo rat Pororous tridactylis ) was seeded at a concentration of 1 x 106 cells per coverslip in 35 mm culture dishes in 3 mL of culture media (MEM with 10% fetal calf serum, 1 % penicillinstreptomycin) and incubated for 72 h at 370C and 5% CO2. Shigellaflexneri M90T wildtype strain was inoculated into brain heart infusion (Difco) and grown overnight at 370C. Bacteria were harvested at mid-log phase and resuspended in MEM without antibiotics to give a final concentration of 1 x 107 or a ratio of 10 bacteria per host cell. Bacteria in 3 mL of culture media were added to each dish followed by centrifugation at 400 x g at room temperature for 10 min and then incubation for 45 min at 370 C and 5% CO2. After incubation, extracellular bacteria were removed by washing three times with Hank's balanced salt (Gibco). The culture media containing gentamicin sulfate (10 ug/mL) was
added back to prevent extracellular growth of bacteria. The monolayers were then incubated for 1-4 h during which microinjection and video microscopy were performed. Microscopy and Microinjection
A Nikon Diaphot inverted microscope was equipped with a charge-coupled device camera (Dage-MTI, Michigan City, IN), and the microscope stage temperature was maintained at 370 C with a MS-200D perfusion microincubation system (Narishige, Tokyo). Digital images were obtained and processed, using an Image-I computer image analyzer (Universal Imaging, West Chester, PA). Velocities of bacterial movement were determined by comparing the images at two time points and measuring the distance traveled by each bacterium using the measure curve length function (Image 1/AT program). Distances were calibrated using a Nikon micrometer. Differences in migration velocities were analyzed using the unpaired Student's t test or the Mann-Whitney nonparametric test. For each bacterium, velocity was determined for 3-4 time points before and 3-4 time points after each microinjection. One to two bacteria were analyzed for each injected cell. In each experiment n indicates the number of velocity measurements. Individual cells were microinjected with peptide using a micromanipulator and microinjector (models 5171 and 5242; Eppendorf, Inc.) as previously described (Southwick and Purich, 1994b).
Immunofluorescence staining using anti-a-actinin antibodies was performed as previously described (Dabiri et al., 1990). In experiments requiring phallicidin staining, PtK2 cells were fixed with 3.7% (vol/vol) formaldehyde in phosphate-buffered saline for 15 min at 25C followed by treatment with 0.4% Triton X-100 and 1.7 X10-7M bodipyphallacidin (Molecular Probes, Eugene, Oregon) for 10 min at 37C. The relative fluorescence intensities of the bodipy-phallacidin stained tails were measured with the Image-1 system using a Genesis I image intensifier (Dage-MTI) in the linear response range. Gain settings were identical for both the Shigella and Listeria rocket tails. The relative intensity was measured at different locations on the tail with a fixed square template
(2 x 2 pixels, brightness function; Image-I/AT). Fluorescence intensity of an identical area adjacent to the actin rocket tail within the cell was measured and subtracted from each value.
Figure 2.2. Comparison of bodipy-phallacidin staining of Shigella and Listeria actin filament rocket tails.
Simultaneous phase (A) and fluorescent micrographs (B) of an intracellular Shigella are shown. Arrows point to the bacterial-actin rocket tail interfaces. Note the faint fluorescence of the actin rocket tails (B) which extend from the back of many of the bacteria. The rocket tails are thin and demonstrate relatively low fluorescence intensity as compared to Listeria actin rocket tails (D). In the phase micrograph of Listeria, phase dense actin rocket tails can be readily visualized (C), and the tails exhibit highest bodipy-phallacidin fluorescence in the region nearest each bacterium (D). Infections were performed simultaneously using the same stock of cells and stained in parallel. Gain settings were identical for both fluorescence images (B, D). Bar, left lower corner of (D) = 10 1m.
Figure 2.3. Simultaneous phase-contrast (A), anti-a-actinin immunofluorescence (B) and bodipy-phallacidin stained fluorescence (C) images of a Shigella rocket tail.
Arrow points to the back of the bacterium which in the phase-contrast image refracts poorly in this region of the cell. Note the bright anti-a-actinin fluorescence as compared to that associated with phallacidin staining (both images were captured with gain settings in the linear response range of the image intensifier). Bar = 10 um
We Oc 26, 199 14:8:1.5
Characteristics of Shigella Movement and Actin Rocket-Tail Formation in PtK2 Cells
Like Listeria, Shigella moves at relatively rapid velocities through the cytoplasm. Although their larger size might be expected to resist migration in a viscous medium, the observed mean rates of Shigella movement in PtK2 cells (0.17-0.05 jtm/sec) were comparable to those of Listeria (0.15 to 0.05 tm/sec) (Southwick and Purich, 1994b; Southwick and Purich, 1995). The maximal velocities of 0.4 Ym/sec attained by Shigella are rarely seen in Listeria-infected PtK2 cells. As observed with Listeria infections, the mean rate of migration varied considerably from day to day. These differences appear to be related to the age of the tissue culture cells at the time of infection, and in all microinjection experiments pre- and post-treatment rates were compared in the same cells. Intracellular movement of Listeria in PtK2 cells is usually associated with the formation of phase-dense rocket-tails on phase contrast micrographs (Sanger et al., 1992; Southwick and Purich, 1994a; 1995). On the other hand, motile Shigella are infrequently associated with phasedense tails (Figure 2.1 A). Bacteria migrating in regions near or within the cell nucleus often display phase-lucent tails (Figure 2.1 E). In most instances, rocket-tails are not seen as the bacteria move through the cytoplasm. Fluorescence Staining of Actin and Alpha-Actinin in Shigella Rocket-Tails
Comparisons of bodipy-phallacidin staining of the actin filament tails reveal that the Shigella-associated structures (Figure 2.2B) have significantly lower fluorescence intensities than Listeria (Figures 2.2D). This observation suggests that Shigella rocket tails have a lower actin filament content than Listeria. As observed in Listeria (Dabiri et al. 1990), the actin filament bundling protein and cross-linking protein ct-actinin also localizes to the Shigella rocket tails (Figure 2.3).
b .... ll
Figure 2.4. Shigella movement and actin rocket tail formation in PtK2 host cells before and after microinjection of the synthetic ActA peptide.
Prior to injection the bacteria are seen to move at 0.12 gm/sec, and maximum tail length is
6.0 vtm (A, B). After injection of an estimated intracellular concentration of 80 nM of ActA analogue (needle concentration 0.8 tM ActA peptide) at 160 sec, bacterial movement stops and the actin tails almost completely disappear (C, D). Times (indicated in sec) are included in the lower left comer of each micrograph. The triangle (drawn by connecting three small phase-dense granules in the cytoplasm) served as a stable reference point. Solid bar = 10 Wm.
Arrest of Shigella Intracellular Movement by the Second Oligo-Proline Repeat Analogue in Listeria ActA Protein
Bacterial motility ceases within 30 sec after injection of the ActA analogue (800 nM needle concentration, estimated intracellular concentration = 80 nM)(Figure 2.4 A-D). Phase-dense actin tails present before injection also disappear within 30 sec. Similar results are shown graphically in Figure 2.5A. Microinjection of this concentration of peptide consistently blocks Shigella movement (mean pre-injection rate of 0.06 0.03 grm/sec, SD n = 47 versus a mean post-injection rate of 0.004 0.01 gmsec, n = 85 velocity measurements) (Table 2.1). At this low intracellular concentration the inhibitory effects of the ActA analogue are not always permanent (Figure 2.5A). One quarter of the bacteria resume migration 2-4 min after microinjection. The rates of movement, however, are in all instances 25-30% of the velocities measured prior to injection (0.0 1-0.02 gm/sec). The inhibitory effects of the ActA are concentration dependent (Figure 2.5B). A lower intracellular concentration (8 nM) of ActA fails to inhibit, while higher intracellular concentrations (400-800 nM) consistently block intracellular movement. In some cells these higher concentrations also cause membrane retraction. Effect of Microinjecting a Binary solution of Profilin and ActA Analogue on Shigella
Although high intracellular concentrations of profilin (10 RM, see below) can
markedly inhibit Shigella movement, microinjection of an 80 nM intracellular concentration of profilin does not have a significant effect on Shigella locomotion (Table 2.1). Nonetheless, microinjection of equimolar binary solutions of the ActA peptide analogue and profilin (needle concentration = 0.8-1.0 gM, corresponding to estimated intracellular concentrations of 80-100 nM) not only neutralizes the analogue's inhibition but significantly increases the velocities by a factor of three (mean rate of movement prior to microinjection 0.09 0.07 jim/sec, n=16 vs 0.3 0.1 gjm/sec, n= 33 postinjection)(Figure 2.6A and B and Table 2.1). The differences in velocities pre- and post-
injection were highly significant on a statistical basis (p< 0.0001). Velocities increased to nearly 0.5 .1m/sec in some instances.
A 80 nM B
ActA peptide B
E o.1 -.1 0.1
> 0.0 0.0-.
0 90 180 270 .001 .01 .1 1
Time (sec) ActA Peptide (M)
Figure 2.5. Effects of microinjection of the ActA analogue on Shigella motility.
(A) Velocity of a single Shigella bacterium in a PtK2 cell before and after microinjection of the ActA analogue. The estimated intracellular ActA analogue concentration was 80 nM (needle concentration 0.8 pLM). The arrow marks the time point at which the peptide was introduced. The graph corresponds to the bacterium shown within triangle of the micrograph shown in Figure 4. (B) Effect of varying intracellular concentrations of the ActA analogue on Shigella intracellular velocity. Horizontal axis is in a log scale. Intracellular concentrations of 8 nM, 80 nM, 400 nM and 800 nM were studied. Bars represent the standard deviation of the mean for 30-80 velocity determinations per concentration.
Introduction of the binary solution also frequently activated stationary bacteria to
move at rapid rates (Figure 2.6B). If the stationary bacteria were included in pre- and postinjection velocity comparisons, the differences were also highly significant (mean pretreatment velocity 0.06 .07 gm/sec, n=25 vs. mean post-treatment velocity 0.25 0.12 .im/sec, n=49, p< .000 1). The dramatic effects of the binary solution are also illustrated in the time-lapse micrographs (Figure 2.6A). A bacterium can be seen to rapidly accelerate in
Figure 2.6 A Time-lapse phase micrograph of Shigella motility in a PtK2 cell before and after microinjection of an ActA analogue/profilin binary solution.
(A) Time-lapse phase micrographs of Shigella motility in a PtK2 before and after microinjection of the ActA/profilin binary solution. This composite photograph depicts the path and distances covered by a single bacterium before and after microinjection of ActA/profilin in an equal molar ratio (estimated intracellular concentration 100 nM, needle concentration 1 tM). Images show the position of the bacterium at 30 sec intervals and are numbered sequentially. The cell was microinjected with the binary solution between images 3 and 4 of the composite. Following microinjection, note the progressive increase in the distance traveled by the bacteria after each time interval. B) The same information in
(A) is depicted graphically as the upper curve of Figure 2.6B.
response to microinjection of a final intracellular concentration of 100 nM of the binary mixture. We have found no other treatment to evoke such a marked enhancement of the bacterial motility. Microinjection of a lower concentration of this equimolar mixture (20 nM) caused a statistically insignificant acceleration of Shigella velocity. (Table 2.1) Analogue/Profilin
0.5 (100 nM each)
0 150 300 450 600
Figure 2.6. B) Velocities of two bacteria in a PtK2 cell before and after the microinjection of an ActA analogue/profilin binary solution (100 nM intracellular concentrations of both reagents, shown in parenthesis; needle concentrations, 1 gM).
Effects of Microinjection of a VASP Oligoproline Analogue Alone and in Combination with Profilin on Shigella Intracellular Motility
Introduction of the VASP analogue (GPPPPPGPPPPPGPPPPP) can also inhibit Shigella motility without causing significant membrane retraction (Figure 2.7A and Table
2.1). This effect is concentration dependent (Figure 2.7B), complete inhibition being seen at intracellular concentrations of 10 gM, while lower concentrations (2 and 6 gM) cause variable inhibition (note the large standard deviation bars at these two concentrations, Figure 2.7B). Introduction of poly-L-proline also causes a dose dependent slowing of bacterial velocity (Figure 2.8A). Microinjection of the same concentration of an unrelated peptide derived from the sequence of MAP-2 had no effect on Shigella migration (Table
2.1). As previously observed with Listeria (Sanger et al. 1995), microinjection of profilin also causes a concentration dependent inhibition of Shigella movement, intracellular concentrations of 10 gM causing nearly total inhibition (Figure 2.8B and Table 2.1). Curiously, introduction of an intermediate intracellular concentration of profilin (6 gM) resulted in a bimodal behavior. Sixty percent of the bacteria stopped moving. The remaining forty percent accelerated their velocity, attaining mean migration rates of 0.19
0.08 gm/sec (n=17). These post-injection velocities were significantly higher than the bacteria's preinjection velocities of 0.14 + 0.05 (n=28, p = 0.039).
A VASP Analogue (10 gM) B
0 90 180270 0 2 4 6 8 10
Time (sec) VASP Peptide (gM)
Figure 2.7. The effect of microinjection of VASP peptide analogue on Shigella motility in a PtK2 cell.
A) Microinjection of VASP peptide analogue, 10 jiM intracellular concentration, into Shigella infected PtK2 cell. The Arrow represents the approximate time of the microinjection. (B) Effect of varying intracellular concentrations of the VASP analogue on Shigella intracellular velocity. The estimated intracellular concentrations of the microinjected peptide are plotted on the horizontal axis. Bars represent the standard deviation of the mean for n = 20-40 velocity measurements per concentration.
The effect of microinjecting a binary mixture of profilin and the VASP oligoproline analogue was also examined (Figure 2.8C). In vitro experiments employing profilin tryptophan fluorescence have recently demonstrated that the (GPPPPP)3 peptide binds to profilin with a KD in the 10-5 M range (Kang et al., 1997). Based on these findings, high equimolar concentrations (10 RM intracellular concentrations) of both profilin and the VASP oligoproline analogue when microinjected (baring interference from other intracellular constituents) should exist as a complex in the cell. We predicted that such a complex might neutralize the inhibitory activity of the two components. In fact, microinjection of this binary mixture accelerates Shigella movement, velocities increasing by a mean of 200% (pre-injection mean velocity: 0.09 0.05 gm/sec, n = 25 vs. postinjection mean velocity: 0.18 0.10, n=61, p< 0.0001) (Table 2.1). Introduction of an equivalent binary mixture of poly-L-proline and profilin inhibits Shigella movement (Figure
2.8D and Table 2.1). Microinjection of a lower equimolar concentration of the VASP analogue and profilin (1 pM intracellular concentrations) fails to accelerate Shigella migration (Table 2.1).
0 1 6 10 14 0 1 2 6 10
PLP (gM) Profilin (KM)
C O PLP/Profilin
0.5 (2.5 lM/10 IM)
W VASP Analogue/
e O.4- Profilin
E (10 RM each) 0.1
7ai 0.1- H
00 -0 0.04
0 180 360 0 90 180 270
Time (sec) Time (sec)
Figure 2.8. Microinjection of poly-L-proline, profilin and mixtures of VASP analogue and profilin and mixtures of poly-L-proline and profilin on Shigella motility.
Dose dependence of (A) poly-L-proline and (B) profilin inhibition of Shigella intracellular motility. The mean velocities of Shigella intracellular migration in PtK2 cells are shown following the microinjection of increasing intracellular concentrations of the two polypeptides. Each point represents the mean of 20-40 velocity measurements. Introduction of an estimated intracellular concentration of 6 RM profilin (needle concentration 60 RM) resulted in a bimodal behavior, 40% of the bacteria accelerating their velocity while 60% stopped moving (see results).
(C) The velocities of a bacterium migrating through a PtK2 cell before and after the microinjection of a VASP analogue/profilin binary solution and (D) before and after the microinjection of binary solution of poly-L-proline and profilin. The values in parenthesis are the estimated intracellular concentrations of the two reagents. Vertical arrows indicate the time when each solution was injected. These individual experiments are representative of numerous experiments for each condition (see results and Table 2.1)
Table 2.1 Effects of Microinjected Peptides on Shigella Intracellular Motility. Additions Intracellular Pre-injection Post-injection Post-injection/ P value
Concentration Velocity Velocity Pre-injection Velocity
(mean, pm/sec, SD)
Act A peptide 80 nM 0.06 0.03 0.004 0.01 0.07 < 0.001
CFEFPPPPTDE (n=47) (n=85)
Profilin 80 nM 0.14 + 0.04 0.12 0.06 0.85 N.S.*
ActA peptide 80 nM/ 0.09 0.07 0.30 0.11 3.33 < 0.001
and Profilin 80 nM (n=16) (n=33)
20 nM/ 0.13 0.05 0.17 + 0.08 1.31 N.S.
20 nM (n=15) (n=12)
VASP peptide 10 gM 0.13 0.05 0.02 0.05 0.15 < 0.001
(GPPPPP)3 (n=40) (n=65)
Profilin 10 .M 0.07 0.03 0.02 0.05 0.28 < 0.001
VASP peptide 10 gM/ 0.09 0.06 0.18 0.10 2.00 = 0.002
and profilin 1OjM (n=25) (n=61)
1 gM/ 0.12 0.06 0.07 0.04 0.58 N.S.
1 gM (n=6) (n=16)
Poly-L-proline 2.5 [M/ 0.14 0.08 0.06 0.11 0.43 < 0.001
and profilin 10 gM (n=29) (n=45)
MAP-2 peptide 10 iM 0.15 0.05 0.15 + 0.07 1.00 N.S.
VKSKIGSTDNIKYZPKGG (n=17) (n=44)
* N.S. = not significant
Dynamic remodeling of the actin cytoskeleton must be controlled (Stossel, 1993; Condeelis, 1993), and bacterial pathogens must utilize these regulatory processes to achieve actin-based motility in host cells in their efforts to evade host defense mechanisms. We compared mechanisms underlying Listeria and Shigella movement in PtK2 host cells. While Shigella rocket tails have a lower F-actin content than Listeria, the average velocities of both pathogens are quite similar. As observed with Listeria, we now find that Shigella rocket-tails also contain the actin bundling and cross-linking protein oc-actinin shown to be
critical for Listeria motility (Dold et al., 1994). These similarities raised the possibility that these two distinct pathogens may be adopting convergent mechanisms to subvert the host cell's actin regulatory system to allow their locomotion within cells and their spread from cell to cell. To explore this possibility, the inhibitory effects of oligoproline peptides based on the sequences in the ActA protein and VASP were examined in cells infected with Shigella. Over the same concentration range that inhibited Listeria intracellular motility (Southwick & Purich, 1994b), the ActA analogue likewise blocked Shigella movement.
We originally hypothesized that the ActA analogue acted by competitively inhibiting profilin binding to bacterial cell wall ActA protein; however, in vitro experiments failed to demonstrate any binding of the ActA oligoproline analogue to profilin (Kang et al., 1997). The discovery that a second host cell actin regulatory protein VASP may serve to link profilin to ActA now provides a self-consistent explanation for our results (Reinhard et al. 1995). It is likely that the ActA oligoproline analogue FEFPPPPTDE dissociates VASP from both Listeria and Shigella. Based on the estimates of Reinhard et al. (1992), the concentration of VASP tetramer in platelets is approximately 0.5-1 tM. The content of VASP in other cells is considerably lower (i.e., approximately 100 nM). The latter value is quite close to the estimated intracellular concentrations of ActA analogue (80 nM) found to arrest Shigella motility. It is noteworthy that Listeria intracellular movement is inhibited by both the ActA analogue and oostatic factor in the identical concentration range. This behavior would be predicted if the peptides interact directly with the limited intracellular pool of VASP.
Dissociation of VASP from the surface of the bacteria would be expected to prevent the concentration of profilin at the bacterial-actin tail interface blocking further actin assembly at this site, thereby preventing bacterial movement (Figure 2.8). Based on our recent studies demonstrating that profilin binds directly to a contiguous triad of GPPPPP repeats spanning positions 172-189 in VASP (Kang et al., 1997), we predicted that microinjection of a synthetic peptide containing this 18 residue triad would block profilin
localization at the bacterial actin interface and prevent bacterial induced actin filament assembly and intracellular movement. Our experiments confirmed this expectation. The intracellular concentrations of peptide required to achieve inhibition of motility were considerably higher than the ActA analogue (10 tM GPPPPPGPPPPPGPPPPP versus 80 nM FEFPPPPTDE), reflecting the higher concentrations of profilin likely to be present in PtK2 cells as compared to VASP and/or a lower affinity of profilin for VASP oligoproline sequence. It is of interest that other investigators have recently demonstrated that the same VASP analogue can dissociate profilin from VASP in vitro (Reinhard et al. 1995), providing further biochemical support for our inferences about the mechanism of action of the VASP analogue in Shigella infected cells. We also find that this same VASP analogue inhibits Listeria intracellular movement at identical concentrations (Kang et al., 1997). Therefore both Shigella and Listeria are likely to utilize VASP and profilin to induce actin assembly in host cells. While all of our results are consistent with the above interpretation, these synthetic peptides may not be entirely specific for the proposed targets, and impaired bacterial movement could represent a nonspecific side effect. Other of our findings argue against such an interpretation. First, introduction of high intracellular concentrations of an unrelated peptide fail to impair motility, excluding a nonspecific toxic effect of synthetic peptides. Second, the ability of equimolar concentrations of profilin to totally reverse the inhibitory effects of the peptides suggests specific protein-protein interactions are responsible for the observed inhibitory effects. Our observations, however, do not exclude the possibility that other host cell actin regulatory proteins in addition to VASP and profilin may play roles in Listeria and Shigella intracellular motility.
What then can be said about the result of our experiments with binary solutions
containing profilin and either of the aforementioned oligoproline sequences? Simultaneous introduction of a profilin and ActA analogue or profilin and the VASP analogue binary solution did more than simply neutralize the inhibitory action. In fact, we were surprised to find that co-injection actually stimulated Shigella to move at rates that were two to three
times greater than their usual velocities. Introduction of the binary solutions even occasionally caused previously quiescent bacteria to commence moving, and these bacteria often reached maximal velocity. This stimulation of movement was observed following the addition of only 80-100 nM concentrations of profilin and the ActA analogue, the same concentration range where microinjection of ActA analogue alone evoked maximal inhibition of both Listeria and Shigella movement. Binding experiments monitoring tryptophan fluorescence of profilin fail to detect binding of the ActA analogue to profilin at concentrations of 100 gM (Kang et al., 1997). Therefore, it is unlikely that these two polypeptides alone form a binary complex before or after microinjection into the cell. They are more likely to form a ternary complex with a third host cell protein, possibly VASP, and this complex in turn could stimulate actin assembly. In vitro binding experiments indicate that the VASP analogue and profilin will associate at the concentrations used in our experiments (10-5M range, Kang et al., 1997). Therefore, the acceleration of Shigella motility by the binary mixtures of VASP and profilin suggests that the profilin-VASP complex can enhance actin assembly in nonmuscle cells. Although further experiments will be required to fully characterize these interactions, the present studies do indicate that under the appropriate conditions profilin can stimulate actin assembly.
Based on our current findings, a working model of how Shigella induces actin
assembly in host cells can be constructed (Figure 2.9). Because the IcsA surface protein of Shigella possesses no ActA oligoproline VASP binding sequence, IcsA protein probably attracts a host cell VASP-binding protein to the bacterial surface to concentrate VASP which in turn binds profilin. Profilin stimulates actin filament assembly behind the bacterium, and this polymerization process propels the bacterium through the host cell cytoplasm. The mechanism(s) by which profilin stimulates actin assembly in cells remain(s) ill-defined. In the presence of the monomer sequestering protein, thymosin 134, profilin can lower the critical concentration of actin filaments (Pantaloni & Carlier, 1993). Profilin also enhances nucleotide exchange on actin monomers (Mochrin & Korn, 1980;
Goldschmidt-Clermont et al., 1991). Under the rapid assembly conditions, 40-200 monomers per sec, associated with Shigella locomotion at rates of 0.1-0.5 pm/sec, nucleotide exchange could prove to be the rate limiting step for new actin assembly and profilin could serve to accelerate this process. In the present model we have illustrated ATP ADP exchange on actin monomers as the most likely explanation for profilin's ability to stimulate host cell actin assembly. While additional biochemical experiments promise a rigorous test of this scheme, a key finding in support of the model is the recent immunofluorescence study demonstrating VASP localization on intracellular Shigella (Chakraborty et al., 1995).
The observation that the ActA analogue can block both Shigella and Listeria actinbased motility suggests that Shigella probably recruits to its surface a host cell protein that contains an ActA-like oligoproline sequence. Kadurugamuwa et al. (1993) suggested that vinculin, itself an oligoproline-containing actin-binding protein, might serve in place of ActA in Shigella actin-based motility. When Shigella infects host cells, vinculin is lost from focal adhesion plaques and could be concentrated on the bacterial surface. Although we clearly observed immunolocalization of vinculin at focal adhesion contacts, we could not demonstrate any accumulation of this protein on the cell wall of intracellular Shigella (data not shown). Such observations do not completely exclude vinculin as the candidate ActA-like host protein, because the amount of vinculin needed on the bacterial surface may be below our detection limit. Alternatively, yet another oligoproline-containing host cell protein may fulfill the requirement for an oligoproline recognition site. Determining the identity of this protein will be of great interest because this ActA-homologue is likely to play a key role in the generation of new actin filaments required for the extension of lamellipods and pseudopods in nonmuscle cells.
In conclusion, our finding that the ActA analogue arrests Shigella motility indicates that its locomotion requires the presence of an oligoproline-containing protein that binds to the bacterium's surface in a manner mimicking the action of Listeria ActA protein.
Moreover, we have demonstrated for the first time that microinjection of a mixture of profilin and the ActA sequence FEFPPPPTDE (or the GPPPPP triad from VASP) can markedly accelerate actin-based motility in living cells. This represents an unprecedented finding that factors introduced by microinjection can actually stimulate directional intracellular actin assembly. These in vivo experiments emphasize the importance of a discrete active pool of profilin that is likely to be responsible for stimulating new actin filament assembly. Shigella and Listeria, two bacterial pathogens with structurally unrelated membrane surface proteins, have thus managed to subvert the host's contractile system to generate force needed for intracellular movement, an evolutionary achievement that allows these pathogens to spread from cell to cell and cause disease. This same system is likely to play a role in promoting localized actin assembly necessary for dynamic remodeling of the leading edge during chemotaxis and phagocytosis.
Actin-Based Locomotory Unit
in Shigella flexneri\ \ \
/T/ D/ D/
Host Cell\ ActA T/ D) D)
I I ADP +LATp
cell wall ATP + )
Figure 2.9. Working model showing the primary components likely to be involved in the actin-based locomotory unit of Shigella.
Shigella contains on its surface the 120 kDa protein IcsA that is likely to attract an ActA-like mammalian protein homologue onto the bacterial surface. This ActA-like protein contains one or more VASP binding sequences (designated as a hatched region) responsible for attracting VASP to the bacterial surface. Because of tetrameric structure, VASP is capable of binding up to 16 profilin molecules, serving to highly concentrate profilin at the bacterial-actin tail interface. Profilin may promote actin filament assembly by increasing the rate of ADP-ATP exchange on actin monomers (chevrons) or profilin may usher actin subunits onto the barbed ends of actin filaments. Microinjection of the ActA peptide FEFPPPPTDE is thought to disrupt VASP binding to the ActA homologue on Shigella and microinjection of the VASP peptide (GPPPPP)3 would be expected to dissociate profilin from VASP. Both peptides act at different steps in Shigella-induced actin assembly and disperse locomotory elements (VASP and/or profilin) from the bacterial surface, thereby blocking actin rocket tail formation and bacterial motility.
VINCULIN PROTEOLYSIS UNMASKS AN ACTA HOMOLOGUE FOR ACTINBASED SHIGELLA MOTILITY
In our recent work (see chapter 2) and work by others (Bernardini et al., 1989;
Dabiri et al., 1990; Heinzen et al., 1993; Cudmore et al., 1995) it had become increasingly clear that the microbial pathogens Listeria monocytogenes, Shigellaflexneri, rickettsia, and vaccinia virus share some of the same components for actin-based motility. Our second aim of this research was to identify the component or components supplied by the host cell during assembly of the actin motor of Shigella. Identifying this component would be of great value in describing the general mechanism of actin assembly that has been conserved in cellular processes of eukaryotic cells.
We hypothesized that Shigella must recruit a Listeria ActA-like molecule or ActA mammalian homologue during its actin-based intracellular spread in order to supply a docking site for VASP. This adapter molecule according to the hypothesis would have an ABM-1 sequence necessary for VASP binding and would also have a domain for binding the surface protein, IcsA of Shigella. To isolate this molecule we would employ a number of approaches based on these unique characteristics.
The first approach probed for the ActA homologue based on its affinity for IcsA. We would generate an affinity column with purified recombinant IcsA covalently bound to N-hydroxysuccinimide-activated sepharose and screen cell lysates for binding partners. A second solid phase assay would be employed in which purified radiolabeled recombinant IcsA would be used to probe cell lysates immobilized to a solid support. Proteins identified by these screens would be submitted for sequence analysis.
Shigella infection had been shown to deplete vinculin from the focal contacts of host cells (Kadurugamuwa et al., 1991), and IcsA is known to bind vinculin and to concentrate vinculin to the back of intracellular bacteria (Suzuki et al. 1996). Vinculin also has an ABM-1 sequence, EPDFPPPPPDLE, which might be recognized by an antibody (FS-1) that we had generated against the synthetic ActA peptide, FEFPPPTDE. The second approach was to demonstrate, by co-localization experiments using immunofluorescence, that the FS-1 antibody recognized an epitope on the rear-ward pole of a moving bacteria and in addition, at the same interface, vinculin could also be localized using a anti-vinculin monoclonal antibody.
Further experimental evidence was gathered by microinjection of Shigella infected PtK2 cells with the FS-1 antibody and a synthetic peptide, Vinc-1, of the sequence, EPDFPPPPPDLE, from the first oligoproline repeat of vinculin. It was hypothesized that either of these moieties could uncouple Shigella's actin motor and this would be further evidence for involvement of vinculin as the VASP binding adapter molecule in Shigella actin-based motility.
Concurrent to my work, Dr. Ron Laine using the FS-1 antibody discovered that platelet extracts contained one or more cross-reactive proteins. A major protein band separated by isoelectric focusing and identified by western blotting was excised and submitted for sequencing and this was identified as a 90 kDa vinculin head-fragment, which retains after cleavage, an ABM-1 sequence at its carboxy-terminus.
Our data suggested that Shigella infection results in the proteolysis of intact 120 kDa vinculin, thereby generating a p90 polypeptide that specifically binds to IcsA and concentrates on the bacterial surface. Infection by Shigella initiates the apoptotic cascade (Zychlinsky et al., 1996) in which cellular proteases are activated. It was reasonable to expect that vinculin might be a target for one of these proteases which could release it from the focal contacts to be available to the bacteria, as p90 with its unmasked ABM-1 site. To
test this, we designed an experiment to analyze whole cell lysates of PtK2 cells for cleavage of full length vinculin to the p90 adapter molecule during infection.
Additional evidence for the role of vinculin p90 as the adapter molecule in the
Shigella actin motor was obtained by microinjection experiments designed and implemented by Dr. F.S. Southwick. Human platelet profilin was purified from platelets by Dr. Fan Kang and aliquots were proteolyzed by thermolysin digestion and the p90 vinculin head and p30 vinculin tail were separated and purified by FPLC. Microinjection of the p90 polypeptide, but not intact vinculin, into Shigella-infected PtK2 cells accelerated intracellular motility of the bacteria by a factor of three.
Our findings indicate that the 90 kDa head-fragment of vinculin can serve as the ActA homologue required for Shigella actin-based motility, and vinculin proteolysis is likely to serve as a molecular switch that unmasks this protein's ABM-1 oligoproline sequence to bind VASP on the bacterial surface and to promote the assembly of an actinbased motor.
Figure 3.1. Characteristics of anti-ActA peptide antibody: Immuno-localization, inhibition of motility, and identification of human platelet p90 polypeptide.
(A) Fluorescence image of Shigella-infected PtK2 using bodipy-phalloidin to label polymerized actin. The thin white bars demarcate the interface between the bacterium and the trailing actin rocket tail. Bar = 10 gm
(B) Phase-contrast image of the same field shown in Panel A.
(C) Indirect immunofluorescence image of the same cells using the FS-1 antibody raised against the FEFPPPPTDE sequence in Listeria ActA protein.
(D) Speed measurements of Shigella in PtK2 cells before and after microinjection of FS- 1 anti-ActA-peptide antibody. The dashed line indicates the time of microinjection of antibody (40 nM calculated intracellular concentration; needle concentration 0.4 gM). Bars represent the standard error of the mean (SEM) of 13 different bacteria at each time point. In order to compare different bacteria, values were graphed as the ratio V(t)/V(0), where V(t) is the velocity at each time point and V(0) is the initial velocity V(0) at t = 0 sec. Comparisons of actual pre- and post-injection speeds also demonstrated a highly significant inhibition of Shigella motility following introduction of the FS- 1 antibody. (mean preinjection speed: 0.11 0.01 m/sec SEM n = 46 vs. post-injection: 0.02 0.01 grm/sec n
- 48, p<0.0001). This same concentration of antibody also significantly inhibited Listeria intracellular motility (mean pre-injection velocity 0.11 0.01 jm/sec n = 14 vs. 0.02 +
0.01 pm/sec n = 18, p<.0001).
(E) Ponseau-S staining of an electroblot of a two-dimensional isoelectric focusing/SDS electrophoresis gel of platelet membrane extracts. The boxed area shows the major spot identified by FS-1 antibody (raised against ActA peptide).(from Laine et al., 1997)
(F) Same electroblot stained with the FS-1 antibody. Two major cross-reactive polypeptides are identified: the 90 kDa polypeptide selected for microsequencing, and a 53 kDa polypeptide. (from Laine et al., 1997)
antibody (FS-1) I 40 nM intracellular
0 60 120
kDa _ _ _ _ _
Materials and Methods
PtK2 kangaroo rat kidney cells were grown and infected with Shigellaflexneri
strain M90T, serotype 5, or Listeria monocytogenes 10403S, virulent strain serotype-1, as previously described (Dabiri et al., 1990; Zeile et al., 1996). Poly-L-proline (MWaverage = 5600), aprotinin, leupeptin, pepstatin A, PMSF, and DTT were obtained from Sigma Chemical Co. (St. Louis, MO). NBD-Bodipy-phalloidin was purchased from Molecular Probes, Inc. (Eugene, Oregon). ActA peptide (CFEFPPPPTDE) and vinculin Vinc-1 peptide (PDFPPPPPDL) were synthesized and HPLC-purified in the University of Florida Protein Sequencing Core Lab.
Anti-vinculin 11-5 mouse monoclonal antibody (Vin 11-5) was obtained from
Sigma (St Louis, MO). The rabbit serum containing anti-VASP polyclonal antibody was a kind gift of Dr. Ulrich Walter (Medical University Clinic, Wurzburg, Germany). Polyclonal rabbit anti-ActA-peptide antibody (FS-1) was raised by immunization with a peptide (CFEFPPPPTDE), corresponding to the ActA's second oligoproline repeat (Southwick and Purich, 1994b) and coupled to keyhole limpet hemocyanin (Cocalico Biologicals, Reamstown, PA). Monospecific IgG was then isolated by immuno-affinity chromatography on CFEFPPPPTDE peptide coupled to cyanogen bromide-activated Sepharose 4B (Laine et al., 1988; Laine and Esser, 1989). Commercial secondary antibodies conjugated to rhodamine, fluorescein, alkaline phosphatase, or horse radish peroxidase were used without further purification.
Purification of Recombinant IcsA
Recombinant Shigella IcsA was cloned from isolated virulence plasmid (Casse et al., 1979) of Shigella by PCR using forward primer 5'-CTG ATA ATA TAG CAT ATG AAT CAA ATT CAC-3' and reverse primer 5'-CAA GCT GTG AAC TAG GAT CCC GAC TAC TCA-3'. After digestion with Nde I and BamHI, the PCR product was subcloned into pET15b expression vector(Novagen, Inc., Madison, WI.) which would upon expression generate a fusion protein of IcsA with 6 N-terminal histidines. Bacterial strain BL2 1 (DE3)(Novagen, Inc) was transformed with the expression vector pET 15b/IcsA-histag. Correct sequence of the insert was confirmed by DNA sequencing performed by the DNA Sequencing Core facility at the University of Florida. IcsA-histag protein was expressed by induction of the transformed strain at an A600=0.6 to 0.4 mM IPTG final concentration and growth for 3 h post induction. The bacteria were harvested by centrifugation at 4400 xg for 10 min at 4oC and the cell pellet frozen in liquid nitrogen. The cell pellet was thawed and resuspended in solubilization buffer, 0.02M Tris-HC1, pH
7.9, 0.5 M NaC1, 6 M Urea and the solution mixed for 1 h at room temperature, then sonicated to shear the DNA. The solution was centrifuged at 27,000 xg for 30 min at 4oC. The recovered supernatant was filtered and then applied to Pharmacia 16/60 G-200 Superdex gel filtration column equilibrated in solubilization buffer at a flow rate of 0.5 mL/min. Pooled gel filtration column fractions were then applied to 5 mL bed volume Ni2+-chelating sepharose column (Chelating Fast Flow Sepharose from Pharmacia) charged with 100 mM NiSO4 and equilibrated in solubilization buffer. The column was washed and bound proteins were eluted with a gradient from 20 mM to 500 mM imidazole. Column fractions were collected, analyzed by SDS-PAGE, and pooled. Pooled fraction of IcsA were refolded by step dialysis in steps from 6 M to 4 M to 2M to 1 M to 0.5 M to 0 Urea over 36 h at 4oC. Protein concentrations were determined by Bio-Rad protein assay. From 1 liter of bacterial culture, 5 mg of purified IcsA was obtained.
From PtK2 cells grown on 150 mm plates, subcellular fractions were prepared by differential centrifugation, a cytosolic fraction (100,000 xg supernatant) and a membrane fraction (100,000 xg pellet). An IcsA affinity column was prepared by binding purified IcsA to a HiTrap NHS-activated 1 mL affinity column (Pharmacia Biotech,Uppsala, Sweden) according to manufacturer's instructions with a calculated 95% binding efficiency. Subcellular fractions were applied to the column equilibrated in 0.1 M NaPi, pH 7.2, 0.05 M NaC1 at a flow rate of 0.1 mL/min. After washing, bound proteins were eluted in high salt at a flow rate of 0.2 mL/min, and fractions collected for analysis by SDS-PAGE and western blotting.
Solid Phase Overlay Assay
The Shigella protein, IcsA was 35S-labeled by metabolically labeling growing E. coli expression cells containing the pET15b/IcsA-histag vector. Methionine and cysteine radiolabeling was accomplished by adding Trans 35S-label (methionine/cysteine) from ICN Pharmaceuticals, Inc.(Irvine, Ca.) to the growing cultures. Purification of the histidine tagged fusion protein was done as above, except the gel filtration step was omitted. Purified IcsA-histag was used in overlay assays in a concentration of 66 ug/mL, specific activity 0.13 uCi/mg to 1.23 mg/mL, specific activity 0.68 uCi/mg. Subcellular fractions of mammalian cells were prepared as above, assayed for total protein concentration, separated by SDS-PAGE and transferred to polyvinylidene difluoride (PVDF)(Bio-Rad, Hercules, Ca.) membrane using electroblotting by tank transfer in 0.01 M MES, pH 6, 20% methanol overnight at 4oC, 35 volts (constant voltage). Additional lanes on the blot included purified full length vinculin and the cleaved p90 form. After transfer the membrane blots were then incubated in blocking buffer; 5% milk, TBS (0.1 M Tris-HC1, pH 8.0, 0.15 M NaC1), 0.1% NaN3 for 1 h to eliminate non-specific binding and then
incubated with radiolabeled IcsA in blocking buffer overnight at room temperature with gentle shaking. The blots were then washed 2 min for a total of three washes in TBS. The membranes were dried and exposed to X-ray film (Kodak X-OMAT AR) overnight at 700C.
Individual PtK2 cells were microinjected with the antibody and peptides and velocities were measured and data analyzed as previously described in chapter 2 (Southwick and Purich, 1994b; Southwick and Purich, 1995; Zeile et al., 1996). Vinculin Proteolysis after Shigella Infection of PtK2 Cells
To quantitate the p90 content of PtK2 cells, 150 mm culture dishes containing
semiconfluent PtK2 cells were infected with Shigella as described above. At various times (lh, 2h and 3h) after initiating infection, cells were harvested into a cocktail of protease inhibitors (2 gg/mL aprotinin, 1 gg/mL leupeptin, 1 gg/mL pepstatin A, 1 mM PMSF, 1 mM EGTA, 1 mM EDTA, and 1 mM DTT in 0.1% Triton X-100). The cells were mildly sonicated at a power setting of 2, 80% duty cycle for 1 min and then centrifuged at 27,000 xg for 30 min at 4oC. The samples were concentrated with a Centriprep-10 membrane concentrator (Amicon Corporation), standardized with respect to protein concentration and then suspended in 2 x SDS sample buffer, boiled for 5 min, and subjected to SDS-PAGE and transferred to PVDF membrane as described above. Uninfected cells and cells exposed to E. coli were extracted and processed in an identical manner. For western blot analysis, membranes were incubated in blocking buffer; phosphate-buffered saline (0.137 NaC1,
0.003 M KC1, 0.0043 M Na2HPO4, 0.0014 KH2PO4, pH 7.3), 1% BSA, 0.05 % Tween20 during the blocking and antibody binding steps. Primary antibody anti-vinculin clone 11-5 (Sigma) and a-actinin (Sigma) were used at a dilution of 1:1000 and the polyclonal FS-1 anti-ActA was used at a dilution of 1:500. Secondary antibodies used included goat anti-rabbit IgG and rabbit anti-mouse IgG conjugated to horse radish peroxidase. Bound
secondary antibody was visualized using the SuperSignal enhanced chemiluminescence method (Pierce, Rockford, IL.). Membrane blots were reprobed by stripping with strip buffer; 2 % SDS, 0.0625 M Tris-HC1, pH 6.8, 0.1 M b-mercaptoethanol at 60oC for 30 min, washing with PBS and then blocking and repeating with new antibody as above.
Indirect Immunofluorescence Microscopy
PtK2 cells were infected with Shigella as described in chapter 2. When it was
determined by microscope, that the majority of the bacteria were moving by actin assembly, the cells were fixed in 3.7 % formalin in a standard salt solution (0.1 M KC1, 0.01 M KPi, 0.001 M MgC12 pH 7.0), and permeabilized with 0.2 % (v/v) Triton X-100. Different sets of the fixed and permeabilized cells were incubated with different primary antibodies; monoclonal anti-vinculin clone 11-5 at 1:200 dilution, monoclonal anti-vinculin hVin-1 at 1:200 dilution, polyclonal FS-1 at 1:50 dilution, polyclonal vinculin tail antibody (Dr. S. Craig, Johns Hopkins University) at 1:500 dilution, in blocking buffer (10% BSA, standard salt, pH 7) for 1 h at 37oC. The cells were washed 2 times with blocking buffer and then secondary antibody was added at 1:200 dilution and incubated for 1 h at 370C. Secondary antibodies used were; anti-mouse or anti-rabbit IgG conjugated to fluorescein isothiocyanate (FITC), or tetramethylrhodamine isothiocyanate (TRITC). F-actin was stained with either FITC or TRITC conjugated phalloidin added to the cells after washing to a final concentration of 0.2 uM. The cells were then washed 3 times and Fluoromount-G (Fisher), an anti-quenching preservative, was placed on the cell surface and then the cell layer sealed with a glass coverslip. Microscopy and image analysis was as previously described in chapter 2.
Vinculin: General Structural Organization,
Bi-Ann aing ite Bi 5ndi d Sit Acning Sites~f
Vinc -1 840 -PEPPPPPD4-849
Figure 3.2. Structural organization of human vinculin (from Laine et al., 1997)
Full-length vinculin has a rigid 90 kDa head region and a 30 kDa tail region defined by a proteolytic cleavage site. As shown, the tail region is folded to indicate the latent binding site for F-actin as proposed by Johnson and Craig (1995). Below is a linearized representation of human vinculin indicating the reported binding regions and the proteolytic cleavage site that generates the p90 head and p30 tail fragments. The head-fragment binds monoclonal anti-vinculin 11-5 antibody (Kilic & Ball, 1991), talin (Price et al., 1989), and cu-actinin (Wachsstock et al. 1987). The location of the IcsA binding site on vinculin's head fragment has not been determined. The tail fragment contains the sites for binding F-actin (Johnson and Craig, 1995) and paxillin (Turner et al. 1990). Human vinculin also has an ABM-1 sequence (residues 840-849) named Vinc-1 located at the carboxy-terminus of the p90 fragment generated from vinculin by limited digestion with thermolysin, and included for comparison is the second ABM- 1 repeat of Listeria ActA (residues 269-278). Also shown are the residues (including the conservative substitution of D for E) shared by the two sequences. The p30 tail contains two other oligoproline sequences (Vinc-2 and Vinc-3 that fail to fulfill the consensus features of ABM-1 homology sequences.
Anti-ActA-Peptide Immunofluorescence Microscopy Localizes a Cross-Reactive Protein at the Back of Moving Bacteria
VASP was recently shown to concentrate on the surface of intracellular Shigella
(Chakraborty et al., 1995). To identify the host cell ActA-like protein that binds VASP, we used an antibody, designated FS-1, raised against the ABM-1 FEFPPPPTDE sequence of Listeria ActA (Southwick and Purich, 1994b). This rabbit polyclonal antibody specifically reacted with the ActA surface protein in Listeria cell wall extracts and bound to the surface of both intra- and extra-cellular Listeria, as demonstrated by immunofluorescence microscopy. The same antibody did not cross-react with Shigella or with cell-free extracts of Shigella grown in bacterial culture (data not shown). Immunofluorescence micrographs of Shigella-infected PtK2 cells demonstrated the presence of a cross-reactive protein or proteins on intracellular bacteria (See Figures 3.lA-C). Moving bacteria with F-actin rocket tails were identified by TRITC-phalloidin staining (Figure 3. 1A, rhodamine channel) and by phase contrast microscopy (Figure 3.1B). The same bacteria were also stained with the anti-ActA peptide antibody FS- 1 (Figure 3.1 C, fluorescein channel). Of 56 bacteria with actin filament rocket tails, 54 demonstrated focal staining with the FS- 1 anti-ActA antibody. As previously reported (Goldberg et al. 1993, Suzuki et al. 1995, D'Hauteville et al. 1996), the IcsA surface protein is shed from the bacterial surface into the actin tails, and the cup-like staining pattern observed in Figure 3. 1C reflects the same distribution that is observed with IcsA (i.e., the greatest concentration was found immediately behind the motile bacteria). These observations suggested that the intracellular Shigella bacterium does attract a mammalian ActA homologue to its surface, as suggested by our earlier inhibition experiments, chapter 2, with ABM-1 and ABM-2 oligoproline peptides (Zeile et al. 1996).
Dr. F.S. Southwick also found that microinjection of the FS-1 antibody (at
intracellular concentrations as low as 40 nM) into Listeria-infected PtK2 cells rapidly halted bacterial locomotion, suggesting that the antibody binds to an ActA region that is critical for
actin-based Listeria motility (unpublished findings). The same intracellular concentration of the anti-ActA antibody also rapidly blocked Shigella motility in PtK2 cells (Figure
3.1D). Microinjection of the same concentration of FS-1 antibody preincubated with a 2 x molar concentration of the Listeria ActA ABM-l peptide (FEFPPPPTDE) failed to block bacterial intracellular movement (data not shown), thereby excluding nonspecific inhibition by IgG.
IcsA Affinity Chromatography and Solid Phase Binding Assays Fail to Isolate or Identify
an ActA Mammalian Homologue
It was initially hypothesized that the ActA mammalian homologue would bind the surface protein IcsA of Shigella with high enough affinity that the application of affinity chromatography and/or solid phase assays would allow for its rapid isolation and identification. With this in mind, a recombinant IcsA histidine tagged fusion protein using an E. coli pET expression system was purified for use in both assays, see materials and methods. In the solid phase assay, IcsA was metabolically radiolabeled and used to probe total proteins from subcellular PtK2 fractions immobilized on PVDF membranes. In repeated assays, we were unable to verify a specific protein interaction with the labeled IcsA probe. In addition, when subcellular fractions of mammalian cells were applied to the prepared IcsA affinity column, no proteins were specifically eluted. Although at first frustrating, it was later appreciated that the binding partner of IcsA or the ActA mammalian homologue, must go through a processing step before it is competent to bind with high affinity and that even after processing it is present in very low concentrations in the cytoplasm of infected cells, see below.
Identification of a Cleaved Form of Vinculin. p90 as the ActA Mammalian Homologue
The ActA-binding protein VASP is abundant in platelets (Reinhard et al., 1992), and we expected that one or more ActA-homologue(s) might be present in higher abundance in platelets relative to PtK2 cells. Concurrent with my work, Dr. Ron Laine used the FS- 1 antibody to identify ActA-homologue(s) in platelet membrane extracts that
were subjected to IEF and subsequent SDS-PAGE. After electrotransfer of proteins from the two-dimensional gel, an immunoblot revealed the presence of a major cross-reactive 90 kDa polypeptide with an approximate isoelectric point of 6.0-6.3 (Figures 3.1E & F). A second, less-abundant 53 kDa protein (with a slightly more acidic isoelectric point) reacted more weakly with the FS-1 antibody.
The 90 kDa polypeptide (designated hereafter as p90) recognized by FS-1 anti-ActA peptide antibody was collected by excising the Ponseau S-stained protein from the twodimensional electrophoresis blot by Dr. Laine. Upon his finding that the N-terminus was blocked, the p90 species was treated treated with Lys-C protease. Subsequent gas-phase microsequencing yielded five different peptides (281-GXLRDPSAXPGDAG; 315ERREILGTXK; 417-IAELXDDPK; 607-LLAVAATAPPTDAPNREEVF; 815SFLDSGYRILGA-826), and all corresponded exactly to the numbered positions within human vinculin (Weller et al. 1990).
As a further test of the FS-1 antibody's specificity for the 90 kDa vinculin
fragment, Western blots of thermolysin-cleaved and intact vinculin were performed by Dr. Laine. Johnson and Craig (1994; 1995) reported that vinculin's tail region is folded over and many binding sites for known actin-regulatory proteins, as well as internal epitopes, are masked by interactions between the head and tail protein (See Figure 3.2). Dr. Laine found anti-ActA-peptide antibody only weakly cross-reacted with full-length vinculin, yet reacted strongly with the 90 kDa head-fragment. The C-terminal region of the p90 vinculin domain also contains an ABM-1 oligoproline sequence (PDFPPPPPDL, designated Vinc1) that bears sequence homology to the Act-A peptide used to generate the FS-1 antibody. Two other oligoproline sequences (named Vinc-2 and Vinc-3) are also found in vinculin p30 tail (See Figure 3.2). These proline-rich sequences are distinctly different from Vinc-1, because they have intervening basic amino acids that were not recognized by the FS-1
Figure 3.3. Immunofluorescence microscopy of Shigella-infected PtK2 cells using antivinculin antibody.
(A) Fluorescence image of Shigella-infected PtK2 cells using bodipy-phalloidin to label polymerized actin. The thin white lines demarcate the junction between the bacterium and the actin rocket tail. The asterisk identifies a bacterium that has a small focal cluster of Factin. Bar = 10 im.
(B) Phase-contrast image of the same field as shown in Panel A.
(C) Indirect immunofluorescence micrograph of the same cells using the anti-vinculin 11-5 antibody (monoclonal antibody directed against the head-fragment). Note that this antivinculin antibody localizes to the F-actin tails and to the focal F-actin cluster at one end of the bacterium identified by the asterisk.
(D) An additional fluorescence image of Shigella-infected PtK2 cells using bodipyphalloidin.
(E) Same cell visualized by indirect immunofluorescence using the vini 1-5 antibody.
antibody, even on Western blots. In this work, he concluded that the full-length vinculin refolds to mask its ABM-1 site, after transfer to the PVDF membrane and exposure to physiologic buffer.
Vinculin's Head Domain Localizes to the Surface of Intracellular Shigella
To learn whether the vinculin head-fragment serves as a mammalian ActAhomologue involved in Shigella motility, bacteria moving in PtK2 cells were studied by immunofluorescence microscopy using a monoclonal antibody (Vin 11-5) whose epitope has been mapped to the 90 kDa head of vinculin (Figure 3.2). Moving bacteria were identified by TRITC-phalloidin staining (rhodamnine channel), and the vinculin headfragment was localized (Figure 3.3) by using Vin 11-5 antibody staining (fluorescein channel). As previously observed with the FS-1 antibody, nearly all moving bacteria (identified by the presence of actin filament rocket tails) also were stained with this antivinculin antibody. Combined with the results with the FS-1 antibody, this suggests that moving bacteria have usurped a form of vinculin containing the unmasked Vinc-1 oligoproline sequence. We had shown that the monoclonal Vin 11-5 antibody did not cross-react with any Shigella proteins and it specifically cross-reacted with intact vinculin and the p90 head-region on Western blot analysis of PtK2 cell extracts (see Figure 3.5).
Microinjection experiments were designed on the basis of our past experiments (chapter 2), to look at the potential functional significance of the ABM- 1 sequence in vinculin p90 in Shigella motility. We examined by microinjection of Shigella infected PtK2 cells, the inhibitory properties of the peptide,Vinc- 1, a synthetic peptide based on the vinculin ABM-1 oligoproline sequence PDFPPPPPDL. As noted above, this sequence is located at the carboxyl-terminus of the p90 head-fragment and shows homology to the oligoproline sequences of Listeria ActA (Figure 3.2). In chapter 2 (Zeile et al., 1996) it was observed that microinjection of the ActA ABM- 1 peptide FEFPPPPTDE arrested intracellular Shigella motility at submicromolar concentrations. The Vinc-1 peptide also inhibited intracellular bacterial movement in PtK2 cells. Complete inhibition was observed
at an estimated intracellular concentration of 800 nM Vinc- 1 peptide (mean velocity preinjection 0.11 0.05 gm/sec, mean and standard deviation, n = 44 versus 0.00 gmlsec, n
- 44 post-injection) (Figure 3.4). Further examination revealed that a ten times higher concentration of Vinc- 1 peptide was required (inset to Figure 3.4) to produce the same level of inhibition observed with the ActA ABM- I peptide (i.e., half-maximal inhibition was observed at 0.5 gM for Vinc-1 versus 0.05 iM for the ActA ABM-1 peptide). Introduction of poly-L-proline (intracellular concentration, I gM) failed to inhibit motility, thereby excluding any nonspecific inhibitory effect as in chapter 2 (Zeile et al., 1996).
0 400 800
0 90 180 270
Figure 3.4. Effect of Vinc- I peptide on Shigella speed.
Representative experiment showing Shigellaflexneri speed measurements before and after introduction of the Vinc- 1 peptide. Arrow indicates time of microinjection of 800 nM Vinc1 peptide, calculated as the intracellular concentration. This experiment is representative of multiple determinations (mean pre-injection velocities: 0.11 0.01, SEM, n=44 vs. postinjection: 0.00 0.01 n=44, p<0.0001). Insert demonstrates the concentration dependence of Vinc- 1 peptide inhibition. The relative values V(t)/V(0) were determined as described in Figure legend 3.1D. Bars represent the standard error of the means (SEM) for 24-44 observations.
The Vinculin Head-Fragment is Generated After Shigella Infection
Although present in outdated platelets, the p90 vinculin head-fragment was not
detected in extracts of uninfected PtK2 cells by western blotting. To study how a Shigella infection may change the distribution or availability of p90, a time-course of p90 generation upon Shigella infection was examined (Figure 3.5 B). The vinculin p90 polypeptide was formed and persisted throughout the period over which Shigella are typically observed to move within the cytoplasm of PtK2 cells (i.e., 1-3 h after the initiation of infection). Densitometry scans of the autoradiograms revealed that the p90 proteolytic fragment represented 5.6-7.7% of the total vinculin in Shigella-infected PtK2 cells. To exclude the possibility that proteolysis was being stimulated by extracellular bacteria, PtK2 cells were exposed to a similar number of E. coli and then incubated for 3h. While closely related to Shigellaflexneri, E. coli lacks the 220 Kb virulence plasmid that permits Shigella to enter host cells, and therefore remains extracellular. Infection with E. coli failed to generate the p90 fragment in PtK2 cells (Figure 3.5B). Western blots of uninfected and infected cell extracts, using an anti-ct-actinin antibody, indicated that there were no differences in aactinin proteolysis in uninfected versus infected cells (data not shown). These observations indicate that the generation of vinculin p90 did not simply arise from generalized proteolysis in infected PtK2 cells.
Based on Western analysis of Shigella-infected PtK2 cells, it was hypothesized that cells containing slowly moving bacteria may be suboptimal with respect to the intracellular concentrations of the vinculin p90. Dr. F.S. Southwick tested this prediction by microinjecting purified platelet p90 (needle concentration = 1.2 gM, estimated intracellular concentration = 0.12 RM) into Shigella infected PtK2 cells. Within 30 s after microinjection, all moving bacteria began to increase their rates and within 60 s, they reached
velocities that were greater than three times their pre-injection rates This was the first observation of stimulated Shigella locomotion by the addition of a fragment of vinculin.
Shigella Infection Time-course
(-) (+) (-) (+) t=O lh 2h 3h E.coli
Stain Western Blot Anti-Vinculin Western Blot
Figure 3.5. Shigella infection induces the production of the 90 kDa vinculin headfragment (Part A, from Laine et al., 1997).
(A) Western blots of PtK2 extracts from uninfected and infected cells using antivinculin clone 11-5. Left lanes: Coomassie blue stained samples, Right lanes: ECL developed Western blots. A 120 kDa cross-reactive polypeptide (full length vinculin) is evident in both extracts. However the 90 kDa head-fragment is detected only in the infected cell extract. Cells were infected for 3 hours prior to generation of the extract. (from Laine et al., 1997)
(B) Time course of p90 formation after Shigella infection. Note the appearance of the p90 band within 1 h of infection. Far right lane shows extract from PtK2 cells infected for 3h with a similar number of E. coli.
A comparison of the actin-based molecular motors of Shigella and Listeria has
demonstrated marked similarities (Chapter 2); yet it is known that the surface protein, IcsA of Shigella, required for motility, has no homology to the ActA surface protein of Listeria. Transposon mutation studies identified IcsA as a necessary component for actin assembly (Bernardini et al., 1989), and expression of only IcsA on the surface of E. coli stimulates actin-based motility in Xenopus oocyte extracts (Goldberg and Theriot, 1995). However, IcsA contains none of the ABM-1 VASP-binding sites. Charkraborty et al. (1995) demonstrated that Shigella attracts VASP to its surface, and we have confirmed binding of VASP to moving bacteria by using anti-hVASP antibody (R. O. Laine, W. Zeile, F. Southwick, and D. L. Purich, unpublished observations). VASP also has three GPPPPP (or ABM-2) sequences for profilin binding, and microinjection of the synthetic VASP analogue peptide blocks Shigella actin-based motility (Chapter 2)(Zeile et al. 1996). Kang et al. (1997) confirmed that the VASP analogue peptide binds to profilin using fluorescence spectroscopy. These findings indicate that VASP and profilin are key components in Shigella's actin-based motor.
We have proposed that the IcsA protein must bind an ActA-like adapter protein containing one or more ABM-1 sequences to bind VASP. Vinculin is known to bind on the surface of intracellularly motile Shigella (Kadurugamuwa et al., 1991; Suzuki et al., 1996), and IcsA is capable of binding the head-fragment of vinculin (Suzuki et al., 1996). Our experiments indicate a role for the p90 head fragment as an ActA homologue that attracts VASP. Vinculin p90 fulfills the following essential features of an ActA homologue: (a) an ability to bind to IcsA on the surface of Shigella (see Chapter 4), (b) an ActA-like ABM-sequence PDFPPPPPDL for VASP binding (Brindle et al. 1996), (c) the capacity to be generated by vinculin proteolysis in Shigella-infected cells, and (d) the ability to accelerate actin-based motility of Shigella when introduced into infected cells (Laine et al., 1997). These properties suggest a mechanism for assembly of an actin-based motor
(Figure 3.6). First, proteolysis generates the 90 kDa head, allowing vinculin's newly exposed VASP-binding sites to link VASP to Shigella's surface. IcsA has recently been shown to have a considerably higher affinity for the p90 head-fragment than for intact vinculin (Chapter 4)(Suzuki al. 1996) suggesting that vinculin proteolysis may also unmask vinculin's IcsA binding site. Binding of p90 to Shigella's surface could then attract VASP to the same location. Brindle et al. (1996) used GST-fusion protein fragments of vinculin (residues 836-940) to demonstrate binding to VASP, and this region of vinculin contains the ABM- 1 sequence. The oligoproline sequences present in the tail region failed to bind VASP (Brindle et al., 1996).
Sechi et al. (1997) has demonstrated that most actin filaments within bacterial rocket tails are organized into long, cross-linked arrays. Their findings disagree with those of Tilney and Portnoy (1989) and Tilney et al. (1992a, b) who observed much shorter actin filaments in Listeria rocket tails. Based on their findings, Sechi et al. (1997) proposed the existence of a polymerization zone on the bacterial surface that accelerates filament assembly without increasing spontaneous nucleation or the capture of new filaments. Kang et al. (1997) have proposed a cluster model for concentrating VASP and profilin into narrow zone on the trailing pole of motile bacteria. They suggest that profilin tethered within this zone may reach sub-millimolar concentrations. Recently, Perelroizen et al. (1996) suggested that profilin promotes assembly by increasing the efficiency of actin monomer addition to the barbed ends of growing filaments, and they conclude that this process need not involve profilin-catalyzed exchange of actin-bound nucleotide. In Figure
3.6 is proposed a mechanism for using vinculin p90 as an adapter molecule for concentrating VASP and profilin in a similar polymerization zone on the surface of Shigella.
Inhibition by a synthetic Vinc- 1 peptide on Shigella intracellular motility also supports the conclusion that the vinculin ABM- 1 site is important. In Chapter 2 we obtained apparent inhibitory constants (i.e. the concentration needed to inhibit motility by
50%) of 0.05 gM and 6 pgM, respectively, for the Listeria ABM- 1 peptide (Phe-Glu-PhePro-Pro-Pro-Pro-Thr-Asp-Glu), and the ABM-2 peptide (Gly- Pro-Pro-Pro-Pro Pro). In
the experiments here, we obtained an approximate inhibitory constant of 0.5 pM for the vinculin ABM-1 peptide (Asp-Phe-Pro-Pro-Pro-Pro-Pro-Asp-Leu) (Figure 3.4). This
indicates that VASP binds to the Vinc-1 ABM-1 peptide about ten times more weakly than the corresponding ActA sequence. This may explain why Listeria form rocket tails having a greater F-actin content than Shigella rocket tails (see Figure 2.2, Chapter 2).
We believe that the ABM-1 peptides do not inhibit bacterial motility by binding to Src SH3 domains. SH3 domains interact with two classes of oligoproline ligands, designated as the PLR1 and RLP2 sequences: APPPLPRR and RALPPLP, respectively (Feng et al. 1994). Positively charged guanidinium groups on PLR1 and RLP2 interact with an essential carboxyl group located on SH3 domains, and this charge-neutralization is an essential binding interaction. The negatively charged ABM-1 sequences most likely interact with one or more cationic side-chains on VASP.
It is most likely that the vinculin p90 is the adapter protein required for Shigella's actin-based motor. It is also possible that other homologous proteins may also participate as a component in an actin-based motor. The cytoskeletal protein, zyxin has several ABM1 sequences (Reinhard et al. 1995b, Purich and Southwick, 1997), but zyxin is about 1030 times less abundant than vinculin in nonmuscle cells (Beckerle, 1986). We have shown that vinculin p90 is clearly present at the actin tail-bacteria interface of Shigella, but zyxin may also play an analogous role as an adapter. Studies are necessary to further clarify the role zyxin and other ABM-1 containing proteins in this process.
Platelet extract experiments have recently identified a complex involving the actinrelated proteins (or ARPs) as another potential component of the actin motor (Welch et al., 1997). This complex may be attracted by the high concentrations of profilin (Machevsky et al. 1994) concentrated by both VASP and another VASP-like protein Mena (Gertler et
al. 1996) at the back of Listeria. Similarly, Shigella's ability to attract VASP raises the possibility that the ARP complex may be a component of Shigella's actin-based motor.
Identification of the p90 vinculin head-fragment on the surface of intracellular
Shigella also raises questions about the abundance of this proteolytic fragment in host cells. Western blot analysis fails to detect the p90 proteolytic fragment in uninfected PtK2 cells, but this head-fragment is known to accumulate in aging platelets (Reid et al., 1993). The latter finding probably accounts for the observation that zyxin is the only VASP binding protein in fresh platelets (Reinhard, et al. 1995b). Intracellular infection by Shigella appears to be required to generate the vinculin p90 fragment in PtK2 cells, and our studies suggest proteolysis may serve as one molecular switch that unmasks the vinculin ABM- 1 site. The appearance of p90 over the same time-frame as Shigella motility as well as the absence of generalized proteolysis point to vinculin proteolysis as a key step in the generation of Shigella's actin based motor.
Figure 3.6. A working model for vinculin proteolysis and the assembly of the Shigella actin-based motility complex (from Laine et al., 1997).
Intact vinculin contains a set of masked cytoskeletal protein binding sites, including the ABM-1 sequence PDFPPPPPDL (shown as a black square beneath the folded C-terminal region of full-length vinculin). Proteolysis releases the p90 head-fragment with its newly unmasked ABM-1 sequence. Shigella's surface protein IcsA binds the vinculin p90 headfragment. The unmasked ABM-1 sequence then binds the tetrameric protein VASP, which deploys its 20-24 ABM-2 sequences to concentrate profilin within a polymerization zone located near growing actin filaments in the bacterial rocket tail. The overall ABM complex is part of the motor unit which uses bound profilin to usher ATP-actin subunits onto the barbed ends of elongating actin filaments.
2. p90 Binding' t'ado fsA and- Profilin
el i e/
-. .,4 ASEMBLD MOOR UIi
PRELIMINARY STUDIES OF HOST CELL VINCULIN AND SHIGELLA ICSA BINDING INTERACTIONS.
An activated form of the host cell protein, vinculin, is the adapter protein or ActA mammalian homologue in Shigella actin-based motility. Future experiments are planned to demonstrate that an actin motor complex can be assembled in vitro from purified proteins and cell free extracts; so indicating, the minimal requirements for bacterial actin-based motility. We have hypothesized that the minimal components for a Shigella actin motor would be cleaved vinculin p90, VASP, profilin, IcsA, and actin, and given proper conditions these would assemble and form a complex that could be detected biochemically. In vitro experiments of Kang et al. (1997) demonstrated that profilin bound the GPPPPP sequences of VASP with a KDof 84 uM. This would indicate a relatively weak interaction and they speculated that this lower affinity is overcome by the bacteria in the cell by concentrating profilin to a calculated concentration of approximately 1 mM at the bacterial pole-actin tail interface. In this way, the bacterium maintains "an activated cluster "for localizing profilin at the bacterial surface due to the multiplicative effect of tethering tetrameric VASP with as many as 16 potential profilin binding sites. Therefore VASP is the central molecule in the actin motor of either Shigella or Listeria. Because it is comprised of four identical subunits, in addition to supplying multiple binding sites for profilin, it may also display cooperativity in ligand binding. An allosteric conformational change induced by binding of VASP to its target ABM-1 site/s (in the case of Listeria found in the four oligoproline repeats of ActA or in case Shigella found in the one oligoproline sequence of vinculin p90) may increase the binding affinity for profilin or profilin-actin
complexes. We speculate that this binding occurs in a sequential manner by assembly of the actin motor in specific steps, or an ordered mechanism of ligand binding. In vitro assembly of the minimal essential actin motor or actin-based motility complex (ABM complex) requires experimental conditions that are optimized to sustain VASP in its high affinity or active state by maintaining relatively high concentrations of profilin and the addition of components should reflect the order in which they would associate in the cell. We have proposed experiments to test this model, the first have been completed and are described here.
The experiments planned required a form of Shigella IcsA that could be readily purified and studied by in vitro techniques. Initially, Glutathione S-Transferase-IcsA fusion proteins were designed to be purified and then used in solution phase binding and chemical crosslinking assays with vinculin. This would be done to establish the usefulness of solution phase bead binding experiments for studying these interactions and determine a preliminary or approximate dissociation constant for vinculin binding to IcsA. In experiments that would follow, IcsA tagged with GST would be used to co-precipitate actin motor complexes from cell free extracts and/or purified components. GST-IcsA fusion proteins (GIF proteins) were designed to express full length IcsA and various in-frame deletion mutants based on our own preliminary work and a published report by Suzuki et al. (1996). In this report, similar GST tagged IcsA proteins were used to demonstrate vinculin binding domains necessary for actin-based motility. First we hoped to identify more precisely the vinculin binding domain on IcsA with a deletion mutant and, second, such a mutant would be of lower molecular weight and, possibly, more readily purified and manipulated than the full length IcsA. This minimal vinculin binding domain mutant would be used in later experiments for studying the in vitro assembly of the ABM complex.
Here we have characterized, through co-precipitation studies, binding of vinculin to full length IcsA fusion protein and evidence for a specific domain for vinculin binding in
the N-terminal glycine-rich repeat region of IcsA. We also show that modification of vinculin by proteolysis is necessary for productive binding to Shigella IcsA in vitro.
Materials and Methods
Cloning of Glutathione S-Transferase Fusion Protein cDNA Expression Constructs
Four BamHII-EcoRI cDNA fragments were generated by PCR from Shigella virulence plasmid, pWR100 template, coding for either full length IcsA or a in-frame deletion mutation fused in-frame to the coding region for glutathione S-transferase. The cDNA fragment coding for the full length GST-IcsA fusion or GIF construct (GST-IcsA Fusion) used in this study, GIF53-758, was amplified by PCR using forward primer GIFR53, 5'- CAA ATA GCT TTT GGA TCC CCT CTT TCG GGT- 3' and reverse primer GIRV758, 5'- AAG CTG TGA GAA TTC TCA GCG ACT ACT CAT- 3'. This cDNA codes for amino acids 53 to 758, which begins the protein coding sequence at the first amino acid downstream of the N-terminal signal sequence and stops at arginine 758 in the natural cleavage site, SSRRA (Fukuda et al., 1995), upstream of the transmembrane coding sequence. The cDNA fragment coding for the N-terminal glycine-rich containing region, construct GIF53-302, was amplified using forward primer GIFR53 and reverse primer GIRV302 5'- ATATCC GCT ATT GAA TTC TCA CAT ATT GCT- 3'. The cDNA fragment coding for an N-terminal deletion of amino acids 1-301, construct GIF302-758 was amplified using forward primer ICFP3 5'- GGT AGC AAT GGA TCC ATT GCT AAT AGC GGA- 3' and reverse primer GIRV758. The deletion mutant GIF302-418 cDNA fragment was amplified using forward primer ICFP3 and reverse primer GIRV418 5'- TAC ATT GAA TTC TCA TGA CAT ATT TCC AGA-3'. PCR fragments were digested with BamHI and EcoRI and ligated into expression vector pGEX4T-1 (Pharmacia Biotech, Uppsala Sweden). E. coli strain BL21 was transformed with expression vector constructs and target protein expressed as described in chapter 3.
Plasmid DNA was isolated from expression clones for confirmation of correct nucleotide sequence by DNA sequencing at the University of Florida DNA Sequencing Core. Protein expression was monitored by time course analysis and western blotting using an anti-GST antibody (Sigma, St. Louis Mo.)
Recombinant Protein Purification
GIF53-758 was purified from inclusion bodies using a denaturation-renaturation purification protocol. Briefly, after induction and growth of bacterial expression cultures, the harvested bacterial pellet was resuspended in lysis buffer, 50 mM Tris-HC1, pH 7.9, 1 mM EDTA, 1 mM DTT and cells lysed by adding lysozyme to 100 ug/mL and post lysis, sonicating to shear the DNA. The bacterial sonicate was solubilized by adding urea to a final concentration of 6 M and incubated for 1 h at room temperature. The cell solution was then centrifuged at 100,000 xg for 1 h at 4oC. A step dialysis followed to refold the protein as described in chapter 3 to a final buffer of 20 mM Tris-HC1, pH 7.5, 150 mM NaC1, 1 mM EDTA, 1 mM EGTA, 1 mM DTT. After refolding, the protein was applied to a glutathione sepharose 4B affinity column or batch purified using glutathione sepharose 4B resin (Pharmacia Biotech, Uppsala Sweden) equilibrated in Tris buffer. The column or batch resin was washed extensively and bound GST proteins were eluted by applying elution buffer, 20 mM reduced glutathione, 50 mM Tris-HC1, pH 7.5, 20 mM DTT and fractions collected. Fractions were analyzed by SDS-PAGE, pooled, dialyzed in Q buffer, 20 mM Tris, pH 7.5, 20 mM NaC1, 1 mM EDTA, 1 mM EGTA, 1 mM DTT, and applied to a Hi-Trap Q 5 mL column (Pharmacia Biotech, Uppsala Sweden). Proteins were eluted with a salt gradient, fractions collected, analyzed, pooled and applied to Superose 12 gel filtration column equilibrated in GF buffer, 20 mM Tris, pH 7.5, 20 mM NaC1, 1 mM EDTA, 1 mM EGTA, 1 mM DTT. Aliquots were flash frozen in liquid nitrogen and stored at -70oC. GIF53-302 was purified by the same procedure, except that the denaturationrenaturation step was not used as this protein was highly expressed as soluble protein. For the ion exchange chromatography step a Mono Q column (Pharmacia Biotech, Uppsala
Sweden) was used. Crude preparations of GIF302-418 and GIF302-758 were prepared only from batch purification using glutathione sepharose 4B. GST protein was prepared from batch purification with glutathione sepharose 4B and then by gel filtration on Superose 12.
In vitro Solution Phase Binding Assay: Bead Binding Assay
Ligand binding reactions were performed with 0.5 uM and 1 uM vinculin (p120, p90) and 2.5 uM, 5 uM, and 10 uM GIF proteins or GST protein (control). Reactions were in a total volume of 200 uL in binding buffer, 20 mM Hepes-NaOH, 150 mM NaCl,
1 mM EDTA, 1 mM EGTA, 1 mM DTT, and 0.1% BSA. Protein components were mixed and 40 uL of 50% slurry glutathione sepharose 4B beads (equilibrated in binding buffer) was added and reactions incubated 1 h at room temp with mixing every 15 min by flicking the tube. Bound protein complexes were co-precipitated by centrifugation at 500 xg at room temp for 5 min. Beads were washed with binding buffer 1 time, centrifuged as previously, and washed two times more with binding buffer without 0.1% BSA. After the final wash, the supernatant was carefully removed and the beads resuspended in 6X SDSsample loading buffer, boiled 5 min, and loaded on an SDS-PAGE gel. The separated proteins were transferred to PVDF membrane by tank transfer as described in chapter 3 and western analysis was performed using anti-vinculin clone 11-5 antibody and protein complexes visualized by chemiluminescence.
Construction and Expression of GST-IcsA Fusion Proteins
Analysis of Shigella IcsA(see Figure 4.1) (Goldberg et al., 1993)(Fukuda et al., 1995)(D'Hauteville et al., 1996) has revealed important structural features; a signal sequence (amino acids 1-53), a glycine-rich region with repeat motif GGXGGX (amino acids 140-307), a specific protease cleavage site SSRIRA (arginine 758-759) and a transmembrane sequence beginning at amino acid 759 for anchoring IcsA into the bacterial outer membrane. Suzuki et al. (1996) demonstrate that during infection, IcsA is
30 CKLSFSPFVVGASLLLGGPIAFATPLS 56
/ PROTEASE CLEAVAGE SITE
I i I Ii i I
140 GSDLSIINQGMIIGGSGGSGA----DHNGDGGE-AVT 171 172 GDNLFIINGEIISGGHGGDSYS---DSDGGNGGDAVT 205 206 GVNLPIINGKTISGGNGGNNYG -- EGDGGNGGDAIT 239 240 GSSLSVINKGTFAGGNGGAAYGYGYDGYGGN -AIT 273 274 GDNLSVINNGAILGGNGGHWGD --AINGSNMTIANS 307
Figure 4.1 Schematic representation of Shigella IcsA (from Goldberg et al., 1993)
cleaved at arginine 758 which releases a 95 kDa fragment that is shed and incorporated into the actin tail and that lack of cleavage at one bacterial pole establishes polarity necessary for actin assembly in vivo. Although cleavage does occur, full length IcsA on the surface of the bacterium is absolutely required for movement by actin assembly (Goldberg and
Theriot, 1995). Suzuki et al. (1996) further show by deletion mutations of IcsA, that the vinculin binding domain is located between amino acids 53 and 506 and the form of vinculin binding with highest affinity is vinculin p90.
Based on the studies of Shigella IcsA, we designed E.coli expression vectors coding for full length IcsA and in-frame deletion mutants for verification of the binding domain on IcsA for vinculin. After isolation of the large virulence plasmid of Shigella for a template, we used specific primers to generate IcsA cDNAs by PCR for cloning into the pGEX4T-1 expression vector. Figure 4. 2 A is a schematic of the GST-IcsA fusion proteins designed and used in this study. The construct for full length IcsA (GIF53-758) is a GST fusion lacking the 52 amino acid signal sequence and the C-terminal transmembrane anchor sequence. GIF53-302 is a deletion mutant spanning the glycine-rich region of IcsA. If the glycine-rich domain was the vinculin binding site, then a fusion protein was designed which eliminated this region but which retained amino acids 302-758 or fusion, GIF302-758. Suzuki et al. (1996) left open the possibility that the vinculin binding domain was outside the glycine-rich region but before amino acid 506 and so GIF 302-418, a deletion mutant containing 116 amino acids, was constructed.
In Figure 4.2 B panels 1-4 are SDS-PAGE gels of cell lysates from expression of: panel 1 GIF53-758, panel 2- GIF53-302, panel 3- GIF302-418, and panel 4- GIF302758. In each panel, lane 1 is cell lysate from cultures before induction by IPTG, lane 2 is lysate after 3 hours of induction, lane 3 is protein bound to glutathione sepharose 4B beads, lane 4 is a wash fraction and lane 5 is protein after elution from the glutathione sepharose 4B. Only expression of GIF53-302 resulted in protein that was primarily partitioned to the soluble fraction. The yield from a liter culture of GIF53-302 was greater than 10 mg of protein whereas soluble yields from the other three expression cultures yielded less than 250 ug per liter of culture. Protein expression from GIF53-758, GIF302758, and GIF302-418 yielded high levels of protein that was partitioned into inclusion bodies. We experienced problems that are common to many GST fusion expressions:
Figure 4.2. Schematic of GST-IcsA fusion proteins used in this study and expression in E. coli.
A) Diagram of wild type Shigella IcsA and GST-IcsA fusion proteins. Construct 1: Full length construct, IcsA amino acids 53 to 758 fused to GST. Construct 2: in-frame deletion mutant comprising glycine-rich repeats of IcsA, amino acids 53 to 302 fused to GST. Construct 3: in-frame deletion mutant comprising 116 amino acids of IcsA from 302 to 418. Construct 4: in-frame deletion mutant lacking glycine-rich repeats of IcsA from amino acids 302 to 758.
B) Coomassie stained gels of GST-IcsA fusion proteins expressed in E.coli. Panel 1: Expression of GIF53-758, Panel 2: Expression of GIF53-302, Panel 3: Expression of GIF302-418, Panel 4: Expression of GIF302-758. Lane designations for Panels 1-4 are: Lane M, molecular weight marker, Lane 1, bacterial lysate pre-induction with IPTG, Lane 2, bacterial lysate 3h post-induction with IPTG, Lane 3, Soluble fraction from bacterial expression lysates, Lane 4, wash fraction, Lane 5, GST fusion protein eluted from beads with 20 mM glutathione.
A) Shigella IcsA WT
53 Glycine-rich repeats 7OO
4) GIF302-758 3 B)
1) 1 2 3 45 2) 1 2345
I, -- mo
3) M 1 2 3 45 4) 12 34 5
403f 31 RINI
incomplete translation products and proteolytic degradation resulting in a preparation contaminated with truncated species of the target protein. Such a mixture is difficult to separate by standard purification techniques as these truncated proteins share electrostatic and solution phase properties. In addition, it is known that GST itself can associate to form dimers causing difficulties for separating a mixture based on size by gel filtration.
Despite these problems, a method was established for increasing the yield of
GIF53-758 so that it could be further purified by ion exchange chromatography and gel filtration. Briefly, the expression cultures after 3 h of induction were harvested, lysed, and then solubilized in 6 M urea. This procedure denatures and solubilizes all of the cellular proteins. The DNA is sheared by sonication and the cell lysates are centrifuged at high speed to pellet residual insoluble material and membranes. The supernatant is then placed in a step dialysis where the urea concentration is slowly diminished over a period of 24-36 h. During dialysis, the proteins are renatured and some correct refolding occurs. The refolded GST tagged proteins are then trapped on a glutathione sepharose 4B resin either as a column or batch configuration. Correctly refolded GST fusions should bind the glutathione beads, whereas incorrectly folded and proteins lacking the GST domain should not bind. The beads with protein bound are extensively washed and then bound protein eluted with 20 mM glutathione. The yield from 1 liter of culture at this stage is 15 mg of a mixture of target protein and truncated target protein. The eluted protein fractions then are size fractionated by gel filtration. After analysis by SDS-PAGE the fractions are pooled with a protein yield of 5 mg per liter starting culture. Pooled fractions are then further purified by ion exchange chromatography to give a final yield of 0.7-1.0 mg of protein per liter culture. Figure 4.3A is an SDS-PAGE coomassie stained gel of GIF protein purifications. Lane 4 is 15 ug of GIF53-758 after employing the above method. In lanes 2, 3 and 5 are loaded 15 ug of GIF302-418, G1F53-302, and GST, respectively. GIF53302 and GIF302-418 proved to be difficult proteins to purify from the contaminating truncated forms and were subject to degradation by proteolysis. Each major band migrated
Figure 4.3 Purification of GST-IcsA fusion proteins by ion exchange and gel filtration chromatography and verification by western blotting.
A) Coomassie stained SDS-PAGE gel of fusion proteins. Proteins were loaded at 15 ug in the following lanes. Lane 2, GIF302-418, Lane 3, GIF53-302, Lane 4, GIF53-758, Lane 5, GST. Lane 1 and 6 are molecular weight markers.
B) Western blot using anti-GST antibody with development by alkaline phosphatase and visualization by chromogenic substrate (NBT/BCIP). Lane designations are as in (A) above.
A) 1 2 3 4 5 6
B) 1 2 345
45 31 -
at the correct size for each of the target fusion products; GIF53-758 at 109 kDa, GIF53302 at 52 kDa, and GIF302-418 at 40 kDa. Figure 4.3 B is a immunoblot of a membrane from a gel identical to Figure 4.3 A which was developed using an anti-GST antibody with development by alkaline phosphatase and visualization by chromogenic reagents, Nitro Blue Tetrazolium and 5-bromo-4-chloro-3-indolyl phosphate. Sequence analysis of the cDNA construct, the correct molecular weight of the target protein on SDS-PAGE, and western analysis with anti-GST antibody confirm expression and purification of the correct GST-IcsA fusion protein.
Solution Phase Binding Assay by Co-Preciptiation with Glutathione Sepharose Beads
In vitro binding of vinculin to IcsA was first studied by binding full length vinculin p 120 and vinculin p90 to GST fusion proteins bound to glutathione sepharose 4B beads. Figure 4.4 A is a schematic diagram of this assay. After binding of proteins and washing to remove non-specific bound proteins, the complexes bound to the beads were solubilized in SDS-sample buffer, boiled and loaded on a SDS-PAGE gel for transfer to a membrane. The proteins were analyzed by western blotting. The results of three binding experiments are represented in Figure 4.4 B in which complex formation was visualized by using an anti-vinculin antibody. Figure 4.4B is a western blot where we show that vinculin p90 binds specifically to the full length IcsA, GIF53-758. Lane 1 is a control binding reaction in which 5 uM GST was incubated with 0.5 uM vinculin p90. Lane 2 is a binding reaction with 5 uM GIF53-758 incubated with 0.5 uM vinculin p90. From these results, we conclude, that in this assay, full length IcsA as a GST fusion protein specifically binds vinculin p90.
Binding of Vinculin to Shigella IcsA Requires the Glycine-Rich Region of IcsA and
Cleavage of Vinculin to the p90 Form
We next investigated an IcsA domain for vinculin binding. To accomplish this we used the same bead binding assay as previously described, but this time, in addition to full length IcsA, the deletion mutants, GIFS3-302 and GIF302-418 were tested. We believe
Figure 4.4 Vinculin p90 binds to full length IcsA and to glycine-rich domain of IcsA.
A) Schematic of GST bead binding assay.
B) Vinculin p90, 0.5 uM, purified from human platelets was incubated with 5 uM of GIF53-758 and precipitated with glutathione-sepharose. Beads were washed and analyzed by SDS-PAGE and immunoblotting using a monoclonal antibody specific for the 90 kDa head of vinculin (11-5). Lane 1, Binding reaction, GST and vinculin p90, Lane 2, Binding reaction, GIF53-758 and vinculin p90.
C) Binding reactions and visualization as in (B). Vinculin p120 or p90, 1 uM, was incubated with 10 uM of GST fusion proteins. Lane 1, Binding reaction, GST and vinculin p120, Lane 2, Binding reaction, GIF53-758 and vinculin p120, Lane 3, Binding reaction, GIF53-302 and vinculin p120, Lane 4, Binding reaction, GIF302-418 and vinculin p120, Lane 5, Binding reaction, GST and vinculin p120, Lane 6, 50 ng purified p90, Lane 7, Binding reaction, Binding reaction, GIF53-758 and vinculin p90, Lane 8, Binding reaction, GIF53-302 and vinculin p90, Lane 9, Binding reaction, GIF302-418 and vinculin p90.
Solution phase binding by GST bead binding assay 1 2
Glutathione Lig and 116seh97.4 ~sephaGST Fusion protein
1234 5 6 789
66 45 -
that the glycine-rich region containing the GGXGGX motif, which is repeated seven times in the IcsA molecule (see Figure 4.1), may be a region in which multiple vinculin molecules could bind, mimicking the four repeats of Listeria ActA. For Shigella motility, this configuration may be necessary because vinculin p90 has only one ABM-1 site per activated molecule. This multiplicity of binding sites may allow Shigella through the IcsA molecule, vinculin p90, and VASP, a way of increasing the effective concentration of profilin and profilin-actin complexes at its bacterial actin-tail interface. The clustering of profilin would occur in the same manner as it is clustered on the trailing pole of Listeria. We therefore tested GIF53-302, an IcsA glycine-rich domain mutant, for its ability to bind vinculin p90 or vinculin p120. An additional mutant, GIF302-418 was also generated to the sequence 116 amino acids upstream of the glycine-rich repeats. This was done to address the findings of Suzuki et al. (1996), that found vinculin binding in the region of amino acids 53 to 418 of IcsA, using similar assays.
Figure 4.4 C is a western blot in which anti-vinculin 11-5 antibody was used to visualize vinculin p120 and p90 binding during complex formation with the GST fusion proteins. In these binding reactions 1 uM vinculin either as p90 or p120 was incubated with 10 uM of the GST fusion proteins. Comparison of Lane 3 in which the binding reaction contained vinculin p 120, and GIF53-302, to Lane 2 containing vinculin p 120 and GIF53-758, and Lane 4 containing vinculin p120 and GIF302-418, demonstrates that the glycine-rich mutant binds with highest affinity to vinculin p 120 and that the deletion mutant, GIF302-418 does not bind vinculin p120. Lanes 1 and 5 are control lanes containing GST and vinculin p120 binding reactions. Lane 7 is the binding reaction of full length IcsA fusion with vinculin p90, Lane 8 is the binding reaction of the glycine-rich mutant, GIF53-302 and vinculin p90, and Lane 9 is the binding reaction of vinculin p90 and GIF302-418. These results indicate that vinculin p90 binds with highest affinity to the glycine-rich mutant, GIF53-302, and full length IcsA, GIF53-758, but does not bind to GIF302-418. Overall, the results are in agreement with Suzuki et al. (1996), in which they
also found the most likely binding domain on IcsA for vinculin to be the region of glycinerich repeats. These experiments indicate that the presence of the vinculin 30 kDa tail on uncleaved vinculin is sufficient to block binding of full length IcsA and that either by removing the 30 kDa tail to form p90 or truncating of IcsA allows for a higher affinity binding interaction. These results are consistent with our previous findings in which we show that vinculin p120 must be cleaved to form the active p90 form that binds IcsA in vivo.
The pathogenic bacteria, Listeria monocytogenes and Shigellaflexneri are able to multiply and spread in host cells due to their ability to subvert the host cells' cytoskeleton for their own actin-based motility. We have demonstrated that the mechanism that is employed by these bacteria is essentially the same (Chapter 2). These organisms during their intracellular life cycle exploit similar host cell proteins to assemble their actin motors. In both cases, profilin and profilin-actin are required for this process. In both cases tetrameric VASP is required for this process and in both cases a specific bacterial surface protein is required for this process. An important finding is, that despite a conserved mechanism, there is an additional host cell protein that only Shigella requires. This may further demonstrate the conserved nature of directed actin polymerization in which the basic actin motor is conserved, but with minor variations to adapt it to different cell types or organisms. It is known that other organisms such as Rickettsia, enteropathogenic E. coli (EPEC), and the vaccinia virus also direct actin assembly for intracellular movement(Southwick and Purich, 1996)(Cudmore et al., 1995), and although their mechanisms for exploiting host cells is not completely understood, it is likely that they use a similar mechanism to that of Listeria and Shigella (Finlay and Cossart, 1997).
In this last series of experiments we have presented additional evidence that vinculin within the host cell must be activated by proteolytic cleavage for productive binding to
Shigella IcsA during infection. This was demonstrated by glutathione bead binding assays. In these experiments it was shown that the highest affinity binding interaction was between IcsA or a specific domain of IcsA and the activated form of vinculin, vinculin p90. These assays alone are not sufficient proof that such an interaction takes place upon Shigella entry into a host cell; but they support our previous immunofluorescence data, microinjection data, cell culture infection data, and immunological data.
We have also presented evidence that the domain on IcsA to which vinculin p90
binds is the glycine-rich region of the repeated motif GGXGGX. As alluded to earlier, the significance of binding in this region is that multiple vinculin p90 molecules may need to be bound to IcsA so that VASP can be recruited and bound in its active conformation. Sustaining the active conformation of VASP may require that VASP be bound at more than one ABM- 1 site. Quantitative analysis of the binding experiments may reveal multiple molecules of vinculin bound in this region of IcsA which would argue for intramolecular binding site interactions. It is also possible that vinculin binds IcsA in a 1:1 stoichiometry, yet VASP is activated due to intermolecular interactions of vinculin residing on adjacent IcsA molecules. In either case, careful analysis may distinguish between these possibilities. Further experiments are necessary to establish the role of ligand binding in the transition of VASP between inactive to active states during assembly of the actin motor. If this could be demonstrated experimentally, VASP could be added to the growing list of proteins such as, hemoglobin and aspartate transcarbamoylase, that are regulated by allosterism.
Beyond characterizing a binding domain on IcsA for vinculin p90, we have
endeavored to design a system in which in vitro assembly of an actin-based motility (ABM) complex can be demonstrated. See Chapter 5 and Figure 5.1 for a further discussion of the ABM complex and our model for actin-based motility. We now have in place tagged molecules which can be readily manipulated and when supplied with the correct components and with the proper conditions should form an ABM complex. The following
experiments are proposed. To assay for complex formation, native VASP will be coprecipitated from platelet extracts using GST-IcsA fusion protein, GIF53-758, bound to glutathione sepharose beads in the presence of exogenous vinculin p90 and profilin. As the source of native VASP, freshly outdated platelets are lysed and extracted with high salt to enrich for membrane bound proteins. The amount of VASP present in 1 unit of platelets is approximately 100 ug which is sufficient for the proposed experiments. To remove endogenous vinculin, the extract is mixed with a anion exchange resin and then filtered. It would be important at this point to remove endogenous zyxin, but with a basic pI very close to VASP it would not be possible to remove zyxin from the extract without removing VASP as well. Although zyxin contains multiple ABM-1 sites it is about 10-30 times less abundant than vinculin (Beckerle, 1986) in non-muscle cells and therefore may not be a confounding factor is these assays. Before use the prepared platelet extract is centrifuged at high speed to pellet F-actin. This step is required to eliminate false-positive complexes that result in pelleting F-actin and its associated actin-binding proteins. The binding reactions are then assembled with vinculin p90 and GIF53-758 to a final concentration of 5-10 uM in a reaction volume of 200 uL. After a brief incubation, glutathione sepharose beads are added, incubation is continued with intermittent mixing, and then the complexes are pelleted by centrifugation. The pelleted beads with bound complex are washed and after the final wash resuspended in SDS-sample buffer and loaded on a gel. The separated proteins are transferred to a membrane and visualized by western blotting and chemiluminescence.
Assuming a relatively conservative KD of 100 uM a complex of 0.01 ug/uL to 0.1 ug/uL could be formed which would be within the detection limits of western blotting and visualization by enhanced chemiluminescence. Reaction and wash conditions would be varied with respect to the exogenous profilin concentration. A series of replicates would be done for each reaction and wash condition. All components kept at a constant concentration would be; platelet extract, exogenous vinculin, and GST-IcsA fusion protein,
while the exogenous profilin concentration would be varied. A first experimental set would have the components without exogenous profilin in the reaction and washes. A second set would have exogenous profilin added to the reaction but not the washes. The third set would have exogenous profilin in the reaction and the washes. We hypothesize that profilin is required in both the reaction and the washes to maintain VASP in its active and therefore high affinity state throughout the manipulations. If this is true, at lower concentrations of profilin, less complex should form. These conditions would exist either in the experimental sets which have low profilin concentration in the reaction and washes or in the experimental sets in which profilin is omitted from the wash steps.
Misleading results in these assays may arise from a number of causes including, coprecipitating actin and actin binding proteins. As a control, the specific ABM- 1 and ABM2 synthetic peptides can be used as competitive inhibitors in the binding reactions. In this case, positive complex formation can be tested by titration with these synthetic peptides to an appropriate concentration that disrupts the interaction. Our microinjection experiments in which we inhibited bacterial motility with these peptides can be used as a guide to determine the inhibitory concentration of synthetic peptide that should be used. For an additional control, we have available a mutant profilin that is impaired for binding poly-Lproline. This profilin should not bind VASP and therefore in the binding reactions should not form an ABM complex. This would be added proof that the complex formed in the presence of wild type profilin was specific.
If successful, these experiments would support our model for a minimal actin motor consisting of: actin, profilin, VASP, and an ABM-1 site containing protein. In the next chapter we will discuss this further and explore other experimental approaches to investigating actin-based motility and its regulation.
SUMMARY AND FUTURE DIRECTIONS After escape from the phagolysosome, Shigella begins assembly of an actin motor from host cell components. It is known that IpaA, a secreted protein of the Ipa family of invasin proteins of Shigella, interacts with host cell vinculin to mediate efficient internalization of the bacterium and that vinculin is redistributed from the cytoplasmic pool to sites of bacterial entry during early stages of the infection (Tran Van Nhieu et al., 1997). Tran Van Nhieu et al. (1997) also found that vinculin was localized around the bacteria very early in the infection well before actin assembly. Processing of vinculin may be required at this stage, for modifying vinculin to a form with high affinity for IcsA. Shigella has to compete with talin, an anchoring protein of focal adhesions, for vinculin and only by generating a form of vinculin with more favorable binding interactions or lower KD for IcsA, could it assemble its actin motor. Regulation of vinculin unfolding by phosphorylation (Schwienbacher et al., 1996) or by phosphoinositides, namely phosphatidylinositol (4,5)-bisphosphate [PtdIns(4.5)P2] (Gilmore and Burridge, 1996) to unmask F-actin and paxillin binding sites has been proposed for actin remodeling in a normal cell. We now believe that Shigella either directly or indirectly initiate a proteolytic processing pathway for vinculin. Shigella during infection may exploit programmed cell death or apoptosis during which the release of proteolyzed forms of cellular proteins are made available. Bacterial induced apoptosis of macrophages by the invasin, IpaB of Shigella, has been demonstrated in vitro (Zychlinsky et al., 1996). IpaB binds to interleukin-1B converting enzyme (ICE), a cysteine protease, [now referred to as caspase-1 (Cohen, 1997)] known to initiate the proteolytic cascade of apoptosis (Martin and Green,
1995). A number of substrates have been identified for caspases and these are cleaved at an aspartate residue within the cleavage motif, DXXDI X. Amino acid sequence analysis failed to uncover a similar motif in vinculin, but this does not preclude the possibility that a caspase may be involved in cleavage of vinculin. Calpain, another cysteine protease, which is an intracellular, calcium dependent protease (Inomata et al., 1996) is activated and cleaves cytoskeletal proteins under conditions of elevated intracellular calcium or during apoptosis (Martin and Green, 1995) and may also be involved. Cell death can also occur upon Shigella infection by the secretion of the ricin-like depurinase or shiga toxin of Shigella which blocks protein synthesis (Gyles, C.L.,1993). Induction of a proteolytic cascade by apoptosis may be a necessary feature of Shigella actin-based motility.
We would like to identify the protease that specifically cleaves vinculin during
Shigella infection. Identification of this protease would add further evidence for the role of proteolysis in Shigella actin motor assembly and would define a more general mechanism for specific activation of regulators of actin polymerization. The following are two possible experimental approaches for isolating and characterizing this specific protease.
In vitro Expression Cloning to Isolate a Specific Protease
Recently described is a technique known as in vitro expression cloning
(IVEC)(King et al., 1997), which uses biochemical functional assays to screen cDNA expression libraries to clone genes of proteins involved in specific protein-protein interactions. Because many cellular processes involve post-translational modification of protein substrates such as by phosphorylation and proteolysis, common functional assays can be used to screen an expression library for a specific modification activity. In so doing the gene for that specific activity is cloned. Briefly, cloning by this method is as follows. A cDNA expression library is subcloned into an expression plasmid vector or a phagemid cDNA expression library is constructed or purchased. Plasmid DNA is purified and the plasmid library divided into pools of 50-100 clones. The pooled plasmids are expressed in an in vitro coupled transcription-translation system. The expressed protein pools are then
screened for biochemical activity. The screening is repeated until a single cDNA encoding the activity is isolated. Among proteins that have been identified this way are; specific substrates for caspases (Kothakota et al., 1997), phosphorylation substrates, substrates degraded by cell cycle specific proteolysis (King et al., 1997), and bacterial proteases (Shere et al., 1997).
This method may prove useful in screening for the protease that is activated during infection by Shigella. In vitro translated protein pools can be screened for the ability to cleave vinculin p120 to p90. Screening of proteolysis reactions can be done in 96 well microtiter plates in which at first screening a recombinant library of 200,000 clones can be assayed with 40-20 plates at 50-100 clones per well. After screening, the cleavage can be monitored by SDS-PAGE and coomassie staining. Some important assumptions are made in using this assay for identifying biochemical activity. It is assumed that the proteins translated in this system are correctly folded and they require no additional modification for activity. Also, if the proteolytic activity required a multi-subunit protein complex; these proteases would not be found in such a screen. Large numbers of false positives may be another problem with this method. The structure of vinculin makes it susceptible to cleavage by proteases in the region of amino acids 830-890 which spans the oligoproline repeats. To address this possibility, positive clones could be screened in conjunction with the FS-1 anti-ActA antibody. Only vinculin molecules with an intact ABM-1 site would be detected and only those plasmid pools would need to be screened further. Additionally, the assay could be scaled up analyze many more recombinant clones by using an ELISA based assay, in our case, based on the FS-1 antibody and screening with a microtiter plate reader. Another variation of the method would look for proteins that specifically bind vinculin by immobilizing vinculin and assaying for expressed proteins that bind with high affinity.