Fluorescent Microspheres as Position Markers to Track the Rotation of Listeria monocytogenes during Actin-Based Motility

Material Information

Fluorescent Microspheres as Position Markers to Track the Rotation of Listeria monocytogenes during Actin-Based Motility
Copyright Date:


Subjects / Keywords:
Actins ( jstor )
Aldehydes ( jstor )
Bacteria ( jstor )
Incubation ( jstor )
Listeria ( jstor )
Listeria monocytogenes ( jstor )
Microfilaments ( jstor )
Monomers ( jstor )
Nucleation ( jstor )
Polymerization ( jstor )

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright Catherine A. Marcinkiewcz. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Embargo Date:
Resource Identifier:
436098713 ( OCLC )


This item is only available as the following downloads:

Full Text






ACKNOWLEDGMENTS I would like to thank Will Zeile for his support in preparing this document, for providing materials and protocols vital to this study, and for his boundless patience throughout the editorial process. I would also like to thank Richard Dickinson for his guidance and technical assistance during this project and for editorial advice in shaping the final document. And finally, I thank Daniel Purich for countless ideas and inspirations for this project, and for his many suggestions on how to overcome various obstacles that were encountered during the course of this study. iii


TABLE OF CONTENTS page ACKNOWLEDGMENTS .................................................................................................iii LIST OF TABLES LIST OF FIGURES ..........................................................................................................vii ABSTRACT .......................................................................................................................ix CHAPTER 1 INTRODUCTION........................................................................................................1 Overview of Actin Dynamics.......................................................................................1 Actin Filament Nucleation and Polymerization...........................................................3 Actin-Based Motility of Bacterial Pathogens...............................................................5 2 LITERATURE REVIEW...........................................................................................10 In Vitro Motility Systems...........................................................................................10 Biophysical Models of Force Generation...................................................................12 Long-length Scale Rotation of Listeria monocytogenes.............................................16 3 RESEARCH DESIGN AND METHODS..................................................................19 Summary of Research Objectives...............................................................................19 Labeling Bacteria with Fluorescent Beads.................................................................19 Motility Assays...........................................................................................................25 Long Length-Scale Rotation.......................................................................................30 4 MATERIALS AND METHODS................................................................................31 Bacteria Preparation....................................................................................................31 Fluorescent Labeling of Bacteria................................................................................32 Preparation of Extracts...............................................................................................33 Motility Assays...........................................................................................................34 Microscopy and Imaging Software.............................................................................34 5 RESULTS AND DISCUSSION.................................................................................35 iv


Fluorescent Labeling of Bacteria................................................................................35 Motility Assays...........................................................................................................40 6 CONCLUSIONS AND IMPLICATIONS..................................................................57 Bacteria Labeling with Fluorescent Particles.............................................................57 Recapitulation of Motility with Labeled Bacteria......................................................59 Implications for Tracking Listeria Rotation...............................................................60 Implications for Actin-Based Motility........................................................................61 APPENDIX REACTION MECHANISMS FOR COVALENT COUPLING METHODS...................63 Carbodiimide Coupling..............................................................................................63 Aldehyde Coupling.....................................................................................................64 LIST OF REFERENCES...................................................................................................66 BIOGRAPHICAL SKETCH.............................................................................................71 v


LIST OF TABLES Table page 3-1: Surface saturation of bacterial surface with microspheres.........................................21 5-1. Calculated parameters for different coupling methods...............................................43 5-2. Determination of doubling time for Listeria in log phase..........................................44 vi


LIST OF FIGURES Figure page 5-1. Motile bacteria labeled with 20 nm particles by ionic adsorption.............................36 5-2. Motile bacteria labeled with 50 nm particles by aldehyde coupling.........................36 5-3. Motile bacteria coupled to 500 nm particles by ionic adsorption.............................41 5-4. Motile bacteria coupled to 500 nm particles by covalent coupling...........................42 5-5. Growth curve for Listeria monocytogenes...............................................................44 5-6. Time-dependent changes in bacterial count per FOV..............................................45 5-7. Time-dependent changes in motility........................................................................46 5-8. Bacterial count per FOV as a function of optical density...........................................47 5-9. Motile number per FOV as a function of optical density..........................................47 5-10. Motile fraction as a function of optical density........................................................48 5-11. Variation of tail length with incubation time on ice................................................49 5-12. Motile number per FOV as a function of 4 C incubation time...............................49 5-13. Motile fraction as a function of 4 C incubation time.............................................50 5-14. Tail length as a function of incubation time...........................................................50 5-15. Number of bacteria per FOV as a function of dilution factor.................................52 5-16. Motility as a function of dilution factor...................................................................53 5-17. Relationship between average tail length and dilution factor.................................54 5-18. Distribution of tail lengths (1:1).............................................................................55 5-19. Distribution of tail lengths (1:2).............................................................................55 5-20. Distribution of tail lengths (1:3).............................................................................56 vii


6-1. Spiral trajectories of actin tails attached to L. monocytogenes..................................62 A-1. Reaction scheme for carbodiimide coupling...........................................................63 A-2. Reaction scheme for aldehyde coupling...................................................................65 viii


Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science FLUORESCENT MICRSPHERES AS POSITION MARKERS TO TRACK THE ROTATION OF Listeria monocytogenes DURING ACTIN-BASED MOTILITY By Catherine A. Marcinkiewcz May, 2005 Chair: Richard B. Dickinson Major Department: Biomedical Engineering Listeria monocytogenes is a facultative, intracellular pathogen with the ability to polymerize actin filaments on its surface through the bacterial protein ActA. These filaments form a crosslinked network that allows the bacterium to propel itself through the host cell cytoplasm. This property of Listeria has been exploited in vitro to study the molecular mechanisms of actin-based motility. It has been previously reported that motile bacteria rotate in a clockwise direction with a constant period, suggesting that a subset of actin filaments remain persistently attached to the bacterial surface during monomer addition. Our goal was to recapitulate this experiment and correlate the rotational movement of motile bacteria with the spiral trajectory of their associated actin tails to demonstrate that filaments are clamped to the bacterial surface during motility. Listeria were labeled with fluorescent beads which were to serve as positional markers on the bacterial surface. As the bacteria move, the distance between the bead and the longitudinal axis of the bacterium should correspond to the degree of rotation, so a ix


stable attachment between bead and bacterium was critical. Three different coupling methods were tested, and it was determined that ionic adsorption yielded the highest proportion of labeled bacteria and had the least inhibitory effect on actin tail assembly. We optimized the growth conditions for live bacteria and found that measurement of optical density of the growing culture was the best indicator of tail-forming ability. Reaction conditions such as incubation time at 4 C in the motility medium and extract concentration were also optimized in terms of motile fraction and mean tail length. Overall, we found that motility and average tail length decrease with 4 C incubation time, indicating that actin gel thickening may impede symmetry breaking at the bacterial surface. The effect of incubation time on motility can be reversed by diluting the extract to 1/3 the original concentration. Finally, we determined that motility peaks at 50% extract concentration, whereas average tail length decreases with dilution factor. This allows us to increase the time window for observing motility while simultaneously enhancing motility by diluting the extract to 50%. x


CHAPTER 1 INTRODUCTION Overview of Actin Dynamics Actin is a cytoskeletal protein that can polymerize under certain conditions to form thin, flexible filaments that form a structural scaffold which determines cell shape and regulates cellular movements. There is a dynamic equilibrium between filament assembly and disassembly which is guided by changes in monomer concentration, intracellular cations, and cell signaling molecules (Goldberg, 2001). Internal processes such as actin sequestration, membrane potential, and the expression of certain actin-binding proteins can influence the rate of polymerization, and these in turn are controlled by extracellular signals from the surrounding medium (Goldberg, 2001). For instance, NMDA receptor stimulation can induce the formation of dendritic spines in neurons, which is mediated by actin polymerization at the post-synaptic terminal. Conversely, AMPA receptors stabilize dendritic spines by inhibiting actin polymerization at the cell surface. This is the fundamental basis of synaptic plasticity and long-term potentiation in the central nervous system (Fischer et al., 2000). Actin filaments form a tight helix of uniformly oriented actin molecules (Gor globular actin) which are essentially polar in that one end is designated as the plus or “barbed” end and the other as the minus or “pointed” end (Alberts et al., 1994). Filament elongation almost always occurs from the plus end, since the rate of polymerization is 10 times greater than at the minus end (Carlier & Pantaloni, 1997). Furthermore, filament disassembly typically occurs at the minus end as the filament matures from ADP-P i -actin 1


2 to ADP-actin. Phosphate release from the actin monomer triggers a conformational change that increases the binding affinity of cofilin, a protein that severs actin filaments and precipitates the dissociation of monomers from the pointed end. This shortens the average length of filaments while increasing the free pool of monomeric actin, thus accelerating the rate of polymerization at the barbed end (Goldberg, 2001). In order for actin polymerization to occur, ATP must be available along with a host of other monovalent and divalent cations such as Ca 2+ , Mg 2+ and K + . Initially, there is a lag phase due to the kinetic barrier to filament nucleation, which requires a trimer of actin monomers in the absence of any other polymerization factors. This is followed by a linear phase in which actin monomers are added successively to the growing end of the filament, a process that proceeds much faster than filament nucleation. Depolymerization at the minus end competes with this process, so a critical concentration of 0.2 M is required to maintain actin filaments (Pantaloni et al., 1984). At concentrations above this critical level, filament elongation will occur. The rate of elongation should remain constant unless actin monomers are depleted from solution, which rarely happens under normal physiological conditions since it is continuously recycled through depolymerization and filament severing. Also, the concentration of monomeric actin in the cell typically varies between 50-200 mM, but the majority of this is sequestered by actin monomer-binding proteins such as thymosin-4 and profilin, which act as a molecular capacitor that maintains the free pool of monomeric actin close to 0.2 M (Weber et al., 1992).


3 Actin Filament Nucleation and Polymerization The rate of filament nucleation is expedited by the Arp2/3 complex both in vitro and on the surface of bacteria pathogens. In fact, one end of the Arp2 and Arp3 subunits is homologous to the barbed end of an actin monomer, while the other end is structurally similar to the pointed end, although this tends to be more divergent. Recent findings indicate that these “barbed ends” of the Arp2/3 complex serve as a template for actin filament nucleation (Kelleher et al., 1995). However, the Arp2 and Arp3 subunits require an additional five subunits in order to form a stable complex, each present in 1:1 stoichiometry. Once this complex is activated, it undergoes a conformation change that allows Arp2 and Arp3 to dimerize within the complex. This exposes the barbed ends of Arp2/3 so that new actin filaments can be initiated on its surface (Goldberg, 2001). The Wiskott-Aldrich syndrome protein (WASP) family acts in synergy with the Arp2/3 complex to facilitate actin filament nucleation in cells. Included in this family is WASP, which is expressed only in hematopoeitic cells (Derry et al., 1994); N-WASP which is ubiquitously expressed and especially pronounced in the brain (Miki et al., 1996), Scar (WAVE) (Machesky et al., 1999) and the yeast isoform Las17p/Bee1p (Li, 1997). Patients that have inherited the mutant form of WASP suffer immunodeficiency, thrombocytopenia, eczema, and hematopoitetic malignancies (Goldberg, 2001). The acidic domains of N-WASP are directly involved in binding Arp2/3 in vivo (Marchand et al., 2001; Rohatgi et al., 1999), and the verprolin and cofilin homology domains have been implicated in actin binding (Miki & Takenawa, 1998). There is also a proline-rich region that mediates interactions with profilin, which is amino-terminal to the WH2 domain. The proline-rich region has been shown to augment Arp2/3 complex activation by the WH2 and acidic domains of N-WASP (Machesky et al., 1999), yet despite these


4 findings, the role of profilin in actin filament nucleation remains unclear. For instance, deletion of the proline-rich repeat region of Scar has no discernible effect on filament assembly (Miki et al., 1998), and profilin actually inhibits actin polymerization in the presence of a Scar fragment containing the proline-rich region, WH2 and acidic domains (Machesky et al., 1999). However, profilin does enhance filament assembly in the presence of Cdc42 and the C-terminal region of N-WASP, which does not contain the proline-rich region (Yang et al., 2000). These results suggest that the synergistic effect of profilin on actin assembly does not require profilin binding to N-WASP. Filament crosslinking is also facilitated by Arp2/3, which initiates new branches at 70 angles to a preexisting filament (Goldberg, 2001). The role of capping protein is to regulate the growth of this branched network which, if unrestricted, would generate a fishbone-like web that fails to direct movement of a motile object or cell membrane (Pantaloni et al., 2000). Also, the frequency of filament branching was found to increase with the concentration of capping protein (Wiesner et al., 2003), which is congruent with the observation that branch elongation is limited by capping protein. It has also been suggested that the inhibition of filament branching is necessary to preserve the free pool of monomeric actin which, according to Brownian ratchet models of actin motility, is necessary to drive polymerization. However, capping protein also been shown to block filament assembly at the motile surface. It was also shown that vasodilator-stimulated protein (VASP) acts as an antagonist to this process by lowering the frequency of filament branching in vitro. This seemed to concur with the findings of Bear et al. (2002), which indicated that VASP increased the rate of actin-based movement in lamellipodia yet simultaneously inhibited filament branching. They concluded that VASP enhanced


5 actin filament polymerization at barbed ends by competitive inhibition of capping protein, resulting in faster elongation rates and longer filaments. This supports the idea that barbed ends of filaments are clamped to the motile surface by VASP and can persist for long periods of time without interference by capping protein. This proposition was later challenged by Samarin et al. (2003), who claimed that VASP does not directly compete with capping protein at the barbed ends but instead generates a new class of unbranched filaments that are independent of ActA-Arp2/3 filament nucleation (Skoble et al., 2001). However, there is still ample evidence to suggest a role for VASP in the mediation of filament binding to the motile surface. Actin-Based Motility of Bacterial Pathogens Certain pathogenic organisms such as Listeria monocytogenes, Shigella flexneri, and Rickettsia conorii that have evolved mechanisms to exploit the cell motility apparatus to their own selective advantage. Once the bacterium replicates within the cell, its progeny are able to assemble actin filaments on their surface through an integral membrane protein (ActA in Listeria; IcsA in Shigella), which colocalizes with the site of actin assembly (Gouin et al., 1999). And although ActA and IcsA do not share any sequence homology, their polar distribution on the bacterial surface allows for unidirectional force generation and movement. Furthermore, neither ActA nor IcsA have any intrinsic polymerization activity and must amass cellular proteins on the bacterial surface in order to mobilize actin filaments. Also, there is no indication that ActA or IcsA can bind directly to actin filaments, which implies another role for actin-binding proteins to mediate the interaction between bacterial surface proteins and actin filaments.


6 Mutant strains of Listeria have been engineered to overexpress ActA and are useful tools to probe the biochemical basis of actin dynamics and the biophysical foundations of actin-based motility (Lasa et al., 1995). Listeria monocytogenes Listeria monocytogenes is commonly known as a food-borne pathogen that causes septicemia, meningoencephalitis, and spontaneous abortion in pregnant women. It is also classified as a gram-positive, facultative organism with the ability to coordinate the unilateral assembly of actin filaments on its surface via ActA, a membrane-bound bacterial protein (Goldberg, 2001). Through a complex series of biochemical interactions, ActA is able to recruit and utilize host cell factors such as VASP, Arp2/3, ADF, -actinin and profilin to form a highly crosslinked and stable network of actin filaments with the microscopic appearance of a comet tail (Gouin et al., 1999). Once assembled, these actin tails can direct the intracellular movement of the bacterium through the cytoplasm until it reaches the host cell membrane. From there, the bacterium becomes encapsulated in a double membrane vesicle which allows it to penetrate into neighboring cells, thus escaping immune detection by the host (Goldberg, 2001). Its unique ability to bypass the cell signaling pathway and harness the actin nucleating potential of Arp2/3 makes this an ideal system to investigate the biomolecular mechanisms of actin-based motility. The N-terminal region of mature ActA (amino acids 31-58) is homologous to the acidic domain of N-WASP (Skoble et al., 2000), enabling it to mimic the activity of this cellular protein and recruit actin filaments to its surface through activation of the Arp2/3 complex. Binding to monomeric actin is mediated by a downstream region (amino acids 85-99 and 121-138) aligning with the first two WH2 domains of N-WASP (Zalevsky et


7 al., 2001). In fact, the N-terminal region of ActA is necessary and sufficient to generate motility, whereas the central region containing the proline-rich domain is not essential but does tend to stimulate movement (Lasa et al., 1995). The C-terminal domain anchors the ActA protein to the bacterial membrane and is not rotationally complaint due to the rigid nature of the cell wall, which in a gram positive organism consists of approximately 20 layers of peptidoglycans stacked together. Deletion mutants in which the C-terminal region of ActA has been removed can be cloned into plasmid vectors and expressed in E. coli. for easy isolation and purification (Lauer et al., 2001). Once assembled, the actin filaments form a gel around the region of highest ActA density, with filaments on either side pushing in opposite directions so that the net displacement in zero (Noireaux et al., 2000). As lateral stresses accumulate within the actin network, these filaments are eventually pushed to the posterior end of the bacterium and form a unidirectional comet tail which allows the bacterium to propel itself through the cell membrane and escape into neighboring cells. Once symmetry breaking has occurred, the actin filaments polymerize at their barbed ends proximal to the bacterial surface (Bernheim-Groswasser et al., 2002). The bacterium advances with each new cycle of actin polymerization while the comet tail remains stationary within the cytoplasm. The more distal regions of the tail are typically crosslinked by -actinin (in Listeria and Shigella, but not in Rickettsia) and will form many stabilizing interactions with the cytoplasm, providing leverage so that the force generated by the growing filament ends can be transduced into the forward motion of the bacterium. Shigella flexneri Shigella flexneri is a gram-negative, facultative intracellular pathogen that causes bacillary dysentery in primates and humans (Labrec et al., 1964). Like Listeria, Shigella


8 also grows and replicates within the host cell cytoplasm and recruits actin filaments to its surface through its integral membrane protein IcsA (Makino et al., 1986). Instead of directly mimicking the activity of N-WASP, the Shigella protein IcsA has evolved to stimulate N-WASP which then goes on to activate the Arp2/3 complex. Another mammalian protein, vinculin, has also been shown to colocalize with sites of actin tail assembly (Suzuki et al., 1996; Laine et al., 1997), although the role of vinculin in actin polymerization remains controversial (Goldberg, 1997). The essential difference between Listeria and Shigella lies in its ability to directly stimulate the Arp2/3 complex without the involvement of N-WASP, whereas Shigella does require N-WASP for filament assembly (Suzuki et al., 1998). Also from a biophysical standpoint, IcsA is rotationally compliant due to the fact that it is embedded in the membrane of a gram-negative bacterium, whose cell wall is far less rigid than that of Listeria (Robbins & Theriot, 2003). Rickettsia conorii In contrast to Listeria and Shigella, the biochemical basis of actin-based motility in Rickettsia conorii remains unclear. It is known that Rickettsia is a gram-negative, obligate intracellular pathogen that causes Mediterranean spotted fever in humans and can replicate via binary fission in the cytoplasm of host cells. It was previously reported that VASP and -actinin colocalized with actin tails in Rickettsia conorii, whereas Arp2/3, cofilin, erzin, N-WASP and capZ did not (Gouin et al., 1999). This suggested that Rickettsia employs a different mechanism of actin-based motility than both Listeria and Shigella. However, since de novo protein synthesis is necessary for actin-based motility, it has been argued that a Rickettsial protein which is similar in function to ActA and IcsA may be involved in actin tail formation (Fehrenbacher et al., 2003). The


9 absence of Arp2/3 from comet tails in Rickettsia implied that some other protein mediates the interaction between actin filaments and the bacterial surface during motility. A more recent study showed that Rickettsia recruits Arp2/3 to the site of bacterial entry into the host cell (Gouin et al., 2004). In fact, the Arp2/3 binding domain of Scar was found to inhibit Rickettsia uptake by Vero cells, which indicates that Arp2/3 is instrumental in this process. Furthermore, Rickettsial infection requires certain cell signaling factors such as Cdc42, PI 3-kinase, c-Src, cortactin and tyrosine-phosphorylated proteins, which modulate the activity of Arp2/3.


CHAPTER 2 LITERATURE REVIEW The unique ability of intracellular pathogens such as Listeria monocytogenes and Shigella flexneri to mobilize actin filaments on their surface has been studied extensively to probe the complex workings of the cytoskeleton. Several key features of these organisms have lent themselves to the study of actin-based motility, including the expression of surface proteins that recruit cellular components and initiate filament assembly and comet tail formation. However, the precise mechanism of force generation by these cellular components or “end-tracking proteins” is still under debate, and the answer to this question will be pivotal to our understanding of the actin cytoskeleton. In Vitro Motility Systems Cell-Free Extracts The actA gene in Listeria and the icsA gene in Shigella are both necessary and sufficient to induce comet tail formation and bacterial movement in cell-free extracts (Goldberg & Theriot, 1995; Kocks et al., 1995). This greatly simplifies the system and allows for more precise control over experimental conditions, such as the depletion of specific factors or the addition of purified proteins, antibodies, and protease inhibitors. Commonly used extracts are derived from Xenopus laevis oocytes, human platelets, bovine and rat brain, and bovine thymus. A standard protocol for the preparation of each type of extract is available, but the concentration of various components may still vary from one preparation to the next. Also, the freezing and storage conditions may affect the integrity of the extract. For this reason, glycerol is often added to the extract before flash 10


11 freezing in liquid nitrogen, after which the extract can be stored indefinitely at -70C or for several months at -20C. Brain extracts differ from those derived from Xenopus laevis oocytes in that VASP is replaced by Mena, which may have some affect on the resistance of actin filaments to depolymerization by cofilin. In fact, the comet tails of Listeria that have been incubated in brain extract tend to be more stable and have a higher average length than those grown in Xenopus extracts (W. L. Zeile, unpublished). It has been suggested that Mena, which is present in high concentrations in the brain, may have evolved to stabilize actin filaments in dendritic spines in order for long-term potentiation to occur (D. L. Purich, personal communication). ATP-Regeneration Systems The hydrolysis of ATP during actin polymerization raises the need for some method of recycling ATP from its hydrolyzed product, ADP. This is often accomplished using creatine kinase, with the reaction kinetics depicted below: ATP + creatine ADP + phosphocreatine The rephosphorylation of creatine tends to increase the rate of phosphate release from actin filaments, making them more susceptible to depolymerization by cofilin. The acetyl kinase system works in a similar manner, but the ATP/ADP ratio is 100 times greater (3000:1 compared to 30:1). This greatly reduces the rate of depolymerization and hence the rate of filament turnover, allowing filaments to persist for longer periods of time (D. L. Purich, personal communication). Reconstitution of Motility Using Purified Proteins The recruitment of cellular proteins (e.g. actin, Arp2/3, cofilin/ADF, VASP, profilin, capZ) to the site of filament assembly is one of the limiting factors in the


12 reconstitution of actin-based motility in vitro. Their concentration in cytoplasmic extracts is often variable and inconsistent, depending on the species from which they were derived, the health and age of the animal tissue, and the method of preparation (Goldberg, 2001). This problem was overcome by the identification of a minimal set of actin-binding proteins that are essential for motility (Loisel et al., 1999). The Arp2/3 complex was found to be necessary but not sufficient to activate motility, and required the addition of actin depolymerizing factor (ADF or cofilin) and capping protein (capZ) to maintain a high steady-state concentration of monomeric actin. Other factors were found to promote motility but were not essential for movement, including profilin, -actinin and VASP (for Listeria). These findings are in agreement with those of Gouin et al. (1999), who found that VASP, -actinin, erzin, Arp3, cofilin and capZ colocalized with actin tails in both Listeria and Shigella. Interestingly, only VASP and -actinin were observed in the actin tails of Rickettsia conorii, which may indicate that VASP plays a key role in actin-based motility. Biophysical Models of Force Generation Elastic Brownian Ratchet Model (EBR model) The behavior of actin filaments in vitro has sparked an ongoing debate regarding the specific mechanism of force generation at the motile surface. The EBR model (Mogilner & Oster, 1996) explains this phenomenon in terms of the thermal fluctuations of the filament, which is estimated to allow for intercalation of new monomer subunits to the growing end of the filament as it breaks contact with the motile surface. The filament then diffuses back to the surface and makes transient contact, exerting a force as it pushes against the surface. The resultant force of many randomly oriented filaments on the posterior end of a bacterium produces a forward motion. Since there are many filaments


13 acting in parallel, there will always be bound and unbound filaments. However, it excludes the possibility of single filament propulsion of a motile object, since this would diffuse away once the filament became detached (Dickinson et al., 2004). The Brownian ratchet model also relies upon the free energy of monomer addition to drive motility, and one must assume a free actin concentration of 13.75 M at the leading edge in order for this to be feasible. This approximation was based upon a model that predicts the spatial distribution of G-actin complexes, in which the concentration of ADP-actin is relatively constant throughout the cell while ATP-actin is on a continuum from front to rear (20-12 M). As ATP-actin is incorporated into filaments at the leading edge, it is quickly replenished by the flux of monomers down their concentration gradient. The total G-actin concentration in the cell is estimated to be 40 M and is assumed to be available for polymerization or transported to the leading edge (Mogilner & Edelstein-Keshet., 2002). However, this does not account for actin sequestration at other locations, which reduces the free actin concentration to 0.2 M, a level that is insufficient to drive motility. The EBR model was modified to incorporate the results of Kuo and McGrath (2000), who showed that the effective diffusion constant of motile bacteria was a few orders of magnitude less than that of free pathogens. To explain this phenomenon, it was proposed that there are two separate phases of actin polymerization; one in which the filaments are transiently attached to the motile surface through the Arp2/3 complex, and the other consisting of filaments that are dissociated from the membrane (Mogilner & Oster, 2003). The association of nascent filaments with the Arp2/3 complex resists the forward motion of the bacterium, which is driven by the rapid polymerization of the detached filaments against the surface. Eventually, these filaments are capped and break


14 contact with the surface. However, since growing filaments must always be under compression, this configuration precludes the rapid elongation of filaments orthogonal to the motile surface (Dickinson et al., 2004). Clamped-Filament Model The clamped filament model (Dickinson & Purich, 2002) asserts that all filaments remain tethered to the motile surface through a protein such as VASP (which also contains binding sites for profilin and actin), N-WASP or formins. Profilin has been shown to bind actin monomers in their ATP-bound state, which would increase the local concentration of ATP-actin near the motile surface. As ATP-actin monomers are released from their profilin binding sites, they are subsequently added to the growing end of a filament which is bound to the motile surface by VASP . Monomer addition triggers ATP hydrolysis on the penultimate monomer, inducing a conformational change which reduces its affinity for the clamp molecule. The clamp then detaches from the penultimate monomer and shifts to the new terminus. Since VASP is bound to the motile surface, its translocation produces a force which is driven by the energy of ATP hydrolysis. Summarized below are the key features of the clamped filament model that differ from Brownian-ratchet type mechanisms: 1) Force generation is mediated by clamping proteins 2) Actin-binding proteins harness the energy of ATP hydrolysis rather than the free energy of actin polymerization 3) All filaments are tethered to the motile surface through an end-tracking protein, rather than a subset of filaments.


15 Monomer-sized steps The persistence of rapidly growing, tethered filaments is congruent with the stepwise motion of Listeria reported by Kuo and McGrath (2000). Using sophisticated laser-tracking techniques, the average step size was found to coincide with that of an actin monomer (5.4 nm) over several cycles of monomer addition. This data, which has been corroborated by other groups, excludes the possibility of a Brownian ratchet-type mechanism, since monomer insertion would cause both the tail and the bacterium to move in opposite directions, thus reducing the apparent step size. Also, the Brownian fluctuations of the tail, if sufficient to allow for monomer intercalation, would obscure the step size distribution so that it would resemble a continuum. The experimental conditions necessary to observe individual steps would not allow for thermal fluctuations of this magnitude, so the processive motion of Listeria can only be accounted for by a strong binding interaction between the filament and the motile surface. Single-filament propulsion Reconstitution of motility using polystyrene beads coated with ActA protein has been demonstrated by Cameron et al. (2001). Several factors were found to influence filament density within the comet tail, including bead size, surface density of ActA on the bead surface, and the concentration of the extract. Manipulation of these parameters produced a single filament associated with a 50 nm bead, with 37.5% ActA surface density in 40% extract. This event would be highly unlikely under the Brownian-ratchet regime in which filaments are only transiently attached to the motile surface, since a 50 nm bead has a high diffusivity and would not remain in the vicinity of the actin filament long enough for motility to occur.


16 Vesicle deformation Lipid vesicles coated with ActA protein assume a tear-drop shape during motility, suggesting that the outermost filaments are pushing against the surface while the inner filaments are under tension, thereby limiting the rate of progression of the motile surface. Furthermore, these inner filaments will eventually break when the tension exceeds the filament tensile strength, causing the vesicle to advance forward and assume a more spherical shape (Upadhyaya & Van Oudenaarden, 2003). This process repeats itself as the severed filaments polymerize and reconnect with the crosslinked filament network. These results suggest that the filaments remain associated with the motile surface during actin-based motility. Long-length Scale Rotation of Listeria monocytogenes This next research study is concerned with the analysis of Listeria trajectories during actin-based motility, during which bacteria were found to rotate with a fixed periodicity in both time and space (Robbins & Theriot, 2003). In this next section, we will review the experimental findings of this study, which imply a continuous interaction between the actin tail and the bacterial surface. Results and Conclusions The bilateral attachment of fluorescent beads to the bacterial surface served as position markers to track the trajectories of motile bacteria. The orthogonal projection of the beads onto the long axis of the bacteria was calculated in time and space, and these data were then used to compute the autocorrelation functions of the orthogonal projections. A plot of the autocorrelation functions of Listeria trajectories with respect to both time and space were shown to have regular periods, suggesting that the bacteria


17 rotate with a constant speed. This is confirmed by the power spectrum of the orthogonal projections, which show a single peak at the dominant frequency. This same experiment was repeated with a mutant DP-L2823 strain of Listeria with its Ena/VASP binding regions deleted. These bacteria also exhibited periodic rotation during motility, but the rotation took significantly longer with respect to time. With respect to distance, however, the rotational period was similar to wild-type Listeria. The lag in rotation with respect to time can be attributed to the slower translational speed of the mutant bacteria, which was found to be half that of the wild-type. These results suggested to the authors that the torque responsible for Listeria rotation is independent of VASP, implying that any interaction between actin filaments and the bacterial surface must be mediated by the Arp2/3 complex instead. Conversely, Y. pseudotuberculosis and E. coli did not show periodic rotation, instead the beads were often stationary with respect to both time and distance. The occasional change in bead position would always coincide with a bacterial turn, as indicated by the change in translational direction of the bacterium. This result implies that rotation is not a feature of all bacteria but may be characteristic of gram-positive organisms. The reason for this behavior has to do with the rotational compliance of the IcsA protein in the plasma membrane of gram-negative bacteria. Since these IcsA is able to freely diffuse in the bacterial membrane, the torque generated by actin comet tails is relieved instead of being transferred to the bacterium. Furthermore, it was determined that there was no correlation between bacterial speed and rotational frequency with respect to distance in Listeria. From this we can conclude that the spatial frequency of rotation is determined by factors other than speed,


18 such as filament density, number of contact points between the tail and the bacterial surface, and the degree of cross-linking in actin filaments. However, it was also found that no correlation exists between speed and rotational frequency with respect to time. Now if we assume that the temporal frequency of rotation is determined jointly by speed and the stiffness of the tail, and that these two factors are independent, it seems reasonable that the temporal frequency would not necessarily correlate with speed. The helical twisting of individual filaments tethered to the bacterial membrane produces a torsional force that translates into the rotation of the bacterium. The observation of continuous rotation does exclude Brownian ratchet models of actin-based motility, in which actin filaments are only transiently attached to the bacterial surface. Correlation between Listeria Rotation and Actin Tail Trajectories Assuming that Listeria rotation arises from the tortuosity of nascent filaments tethered to the bacterial membrane, we expect that the trajectory of the actin tails will reflect the spiral path of the bacterium (Zeile et al., in press). In fact, it is likely that the twisting of actin filaments produces a torque on the entire crosslinked ensemble which is dissipated by the rotation of the bacterium. For this reason, we would expect a tightly coiled helix to correspond to bacteria with a high spatial frequency of rotation, while more loosely coiled tails would be coupled to bacteria with a low spatial frequency of rotation.


CHAPTER 3 RESEARCH DESIGN AND METHODS Based on the clamped filament model of actin motility (Dickinson & Purich, 2002), it follows that Listeria monocytogenes should rotate about its long axis as previously shown by Robbins and Theriot (2003). This model further predicts that the periodicity of rotation should correlate with the spiral trajectory of its associated actin tail, which we intend to demonstrate with our current research. Summary of Research Objectives Our first objective was to find a reliable method of labeling bacteria with fluorescent microspheres without inhibiting motility. Additionally, we needed to optimize microscopic visualization of the beads for documentation and image tracking purposes. The second objective was to find a set of conditions that yielded the highest number of motile bacteria and also to determine which coupling method had the least inhibitory effect on motility. And finally, our third objective was to document bacterial rotations using time-lapse images of motile Listeria. Labeling Bacteria with Fluorescent Beads Particle Dimensions and Fluorescence We have experimented with three different-sized beads (20, 50 and 500 nm) to determine their utility as fluorescent markers for motile Listeria. Several important factors in this assessment include monodispersity, visibility under 40X objective, ability to separate from bacteria in mixed suspensions, and labeling efficiency. Particle aggregation is especially pronounced with smaller beads, which have a higher collision 19


20 rate and, consequently, an increased incidence of hydrophobic interactions. The resolution limit of light microscopy is equal to the wavelength (/2) of light passing through the medium, which under optimal conditions is just under 0.2 m. Therefore, the minimum particle size that can be theoretically detected is also 0.2 m, which would preclude use of the 20 and 50 nm beads. However, we were able to observe monodisperse 50 nm fluorescein beads with a light microscope and a fluorescent lamp, which agrees with the findings of Robbins and Theriot (2003). Given that particle concentration is held constant, the smaller beads have a higher diffusivity which should translate into an increased labeling efficiency. However, we should also keep in mind that saturation with beads is not a desirable feature in our experiments, since we need to track specific points on the bacterial surface. Also, if we assume that the particle suspension behaves in a manner similar to a protein solution, the surface coverage () will only depend on diffusivity in the short term, before the system has reached steady-state. At equilibrium, the surface coverage will depend on the dissociation constant (k d ) between the particle and the surface, which is inversely related to particle size. Smaller particles tend to form fewer bonds with the surface and can thus freely dissociate while larger particles will persist in their association with the surface. There is a flaw in this argument which has to do with the relative dimensions of the system. In a protein solution, it is assumed that the proteins are negligibly small when compared to the surface. With the particle suspension this is not the case, since the diameter of the microspheres (0.02 – 0.5 m) is on the same order of magnitude as the bacterium (0.4 0.5 m in diameter and 0.5 2.0 m in length). The surface area of the bacterium can be approximated as follows:


21 S2ba21a 11a asin1a Where a = radius of bacterium b = semi-length of bacterium If we assume that a = 0.25 m and b = 0.5 m, we get S = 1.569 m 2 . If we assume that each microsphere will occupy an area equal to the square of its diameter, it follows that the number of microspheres required to form a monolayer is limited by particle diameter. Table 3-1: Surface saturation of bacterial surface with microspheres Particle diameter (nm) Saturation # (particles) Surface conc. (particles/ml) Solution conc. (particles/ml) 20 6.276 1.25E+17 5.30E+15 50 627.6 8.00E+15 2.31E+14 500 3922.5 8.00E+12 3.74E+11 In the case of protein solutions, the surface phase is typically more concentrated than the solution phase, since the adsorbed proteins lose much of their mobility. The same holds true for particle suspensions, since we know that surface saturation can be achieved at 1/50 factory concentrations (given in Table 3-1). So if we wanted to adsorb 2-3 beads to the surface of the bacterium, we would need to dilute the bead solution accordingly. Experience working with rhodamine beads has shown that they are difficult to distinguish from rhodamine actin in comet tails, and are often undetectable on the bacterial surface due to the adsorption of rhodamine actin. Also, fluorescein beads appear brighter under fluorescence than the rhodamine beads. For this reason, we chose to work with fluorescein beads over rhodamine beads.


22 Coupling Methods We explore three different coupling methods in our experiments in order to assess their efficacy in labeling bacteria. The first of these methods is the acid-catalyzed conjugation of aldehyde-modified beads to amino groups on Listeria. This results in the formation of a Schiff base which is a moderately stable imine. Reduction of the base with sodium cyanoborohydride will produce a more stable 2’ amine, which can be stored for longer periods. The second method that we applied was very similar to that employed by Robbins and Theriot (2003). Carboxylate beads are ionically adsorbed to the bacterial surface, eliminating the need for lower pH buffers which may potentially cause vital protein domains to unfold, rendering them inactive during motility. The third method employed was carbodiimide coupling, a widely used and highly reliable method of protein conjugation, although it is usually applied as a means of coupling one soluble protein to another (e.g. profilin to actin). In our case, a functionalized microsphere was treated as the primary protein with an abundance of carboxyl groups, while the bacterium was treated as the secondary protein with a large number of available amino groups. The main caveat in this grossly oversimplified system is that protein solutions behave very different than particle suspensions, so many of the problems that we encountered were unforeseen by the original authors of this protocol (Grabarek & Gergely, 1990). We computed the labeling efficiency for each of the coupling methods described above using the following formula. = labeling efficiency (fraction of labeled bacteria) i = number of beads bound


23 A i = number of bacteria with i beads bound N = maximum number of beads bound 1NiAi0NiAi If we let A T = total bacteria count then the equation becomes AT0NiAi 0NiAiAT We can then draw a histogram to represent the distribution of labeled bacteia, and from this we can calculate the mean number of beads bound and the standard deviation from this mean. This will be instrumental in deciding which coupling method is the most effective, since ideally we would like to achieve a high fraction of labeled bacteria and for the average number of beads to fall in the range of 2-3 beads per bacterium. The raw data will consist of the number of bacteria with i beads bound in 5 different fields of view for 3 replicates of each set of reaction conditions. Reaction Conditions The temperature at which coupling reactions are carried out is expected to influence the collision frequency between components in the system and hence the reaction time. Ideally, these reactions could be carried out at room temperature or 37C in order to speed up these reactions, although bacterial proteins such as ActA may be susceptible to denaturation at high temperature. Also, live bacteria will continue to divide even at 4 C, albeit at a much slower rate than they would at 25 C or 37 C.


24 We expect a longer incubation time to inhibit motility, especially for reactions carried out at acidic pH. In this case, the lower temperature regime might actually be counterproductive. For these reasons, we have sought to determine the optimal reaction time and temperature which will yield the highest fraction of motile bacteria. To this end, we have prepared a set of aldehyde coupling reactions in which we varied the concentration (10:1, 20:1), and temperature (37 C for 20 min, 25 C for 2 hrs, or 4 C for 18 hrs). However, the various coupling methods called for different reaction times, which was followed in our experiments. For instance, the incubation time was shortened to 10 minutes when using ionic adsorption as a coupling method (Robbins & Theriot, 2003) and lengthened to 2 hours for the carbodiimide coupling procedure (adapted from Grabarek & Gergely, 1990). The set of reaction conditions given in these standardized protocols was adjusted to meet the demands of our particular system. So far we have discussed the labeling efficiency with regard to coupling method, temperature, and reaction time. However, bead concentration is also a key factor in this process, and together with temperature this will determine the collision frequency of beads with bacteria. The 500 nm beads are in a 3.74 x 10 11 cm -3 particle suspension which is typically diluted to the desired concentration. If the labeling efficiency or the mean number of bound beads is not adequate, we can either adjust the bead concentration or vary the reaction time. Conversely, we may find that the beads are too concentrated such that the average number of beads per bacterium is too high. Also, the incidence of crosslinking between bacteria should increase with the collision frequency between a bound bead and another bacterium, which correlates with the bead concentration and


25 labeling efficiency. A reaction matrix could be used to test out many variations in parallel. Wash and Separation Step Our goal was to determine the optimal centrifugation time and speed that would effectively separate the beads from the bacteria in a mixed suspension. It was already known that the bacteria would pellet after 5 minutes at 7,000 rpm, so we could use this as a frame of reference. Since the beads have a lower density (1.05 g/cm 3 ) and are smaller in size than the Listeria, they should take longer to pellet and will require a higher centrifugation speed. Motility Assays The in vitro motility assay has already been optimized for unlabeled, iodoacetate-fixed Listeria prepared using a standardized protocol (Zeile et al., in press). On the other hand, actin tail assembly in labeled bacteria is somewhat compromised. We took several approaches to enhancing motility, which included varying the incubation time on ice, diluting the motility medium with brain extract buffer, and using fresh Listeria harvested from an overnight culture instead of iodoacetate-fixed Listeria. Effects of Coupling Method and Labeling Efficiency Our goal is to select a set of reaction conditions that yield the highest number of double or triple-labeled bacteria to use in motility assays. Samples from each group (ionic adsorption, aldehyde coupling and carbodiimide coupling) were incubated for 20 minutes on ice in cytoplasmic extract supplemented with rhodamine actin, acetyl kinase, and protease inhibitors. This allows sufficient time for filament nucleation before a small sample (10 l) is pipetted onto a glass slide. Once the actin tails are bound to the slide, they can be readily observed in both phase contract and rhodamine fluorescence.


26 Next, we calculated the fraction of bacteria with actin tails and the fraction of motile bacteria with beads bound. = fraction of bacteria with actin tails = fraction of motile bacteria with beads bound 0NiBi0NiAi 1NiBi0NiBi Again, if we let B T = total number of motile bacteria BT0NiBi Then the equations can be simplified as follows: BTAT 1NiBiBT Again, we can use this data (B 1 , B 2 B N ) to construct a histogram that will illustrate the number distribution of beads on motile bacteria. This should reveal whether or not the beads inhibit motility, enhance motility, or have no effect on motility. Our efforts to recapitulate motility in fluorescently-labeled Listeria may have been complicated by any one of several factors, including low pH treatment (for covalently-modified bacteria), surface effects (e.g. steric hindrance of tail formation by beads bound to amino groups on ActA), or bacterial crosslinking by the beads. One way to distinguish between these possibilities is to compare the motile fraction of labeled bacteria across the


27 two groups (adsorption-labeled vs. covalently-labeled bacteria). If this value is roughly the same for both groups, then low pH treatment has no discernible effect on motility. Surface effects can be determined by comparing the motile fraction of bacteria bound with i < n beads with the motile fraction of bacteria bound with i n beads, where n is some arbitrary value. Again, if these values are approximately equal, then motility is not inhibited by steric factors. = motile fraction of labeled bacteria = motile fraction of bacteria bound with i < n beads = motile fraction of bacteria bound with i n beads 1NiBi1NiAi 1n1iBi1n1iAi nNiBinNiAi Incubation Time for Bacterial Cultures The Theriot group worked with freshly harvested Listeria cultured overnight, which has proven more difficult than we originally thought. In our hands, actin tail assembly was very inconsistent using this method. In order to obtain a more consistent result using freshly harvested Listeria, we grew up four separate cultures and measured the optical density every hour from 12-18 hours. Beyond this point, bacterial growth will be limited by contact inhibition, which results in the release of enzymes and other factors deleterious to bacterial growth. All samples that had reached a minimum OD 600 of 0.1 were harvested by centrifugation and washed three times in PBS, then resuspended in brain extract buffer with 20% glycerol. Motility assays were performed with these bacteria and individual samples were analyzed by phase contrast and rhodamine


28 fluoresence microscopy. We then calculated the fraction of motile bacteria for 5 fields of view (FOV) and computed the average tail length for motile bacteria. Incubation Time on Ice Once the bacteria are added to the motility medium, the entire mixture is incubated on ice to facilitate tail formation at the bacterial surface. This low temperature incubation may accelerate filament nucleation at the bacterial surface by reducing thermal fluctuations and increasing the residence time of individual components. It could also stabilize the filament network and increase the density of the actin gel prior to symmetry breaking. For this reason, we may be able to enhance motility by adjusting the incubation time on ice. In this next experiment, the 4 C incubation time was varied from 10-30 minutes in order to measure the motile fraction. This is expected to correlate with the incubation time so that we can determine the optimal reaction conditions. The average tail length may also correlate with incubation time, since the time to symmetry breaking may vary with the density of the actin gel surrounding the bacterium. Bacterial Velocity and Tail Length Once the optimal conditions for Listeria motility were determined, we sought to increase the average tail length and the time window for observing motility. It has been previously demonstrated that tail length is directly proportional to the velocity of the bacterium (Theriot et al., 1992). In other words, more rapidly moving bacteria appear to have longer tails than slowly moving ones. It was found that filament density decreased exponentially from the bacterial surface to its distal end, and that tails were no longer visible once they fell below 30-50% of their peak density. This has to do with the rapid depolymerization of filaments at their pointed ends, which keeps the filament length relatively constant once polymerization has gotten underway.


29 In our system, however, this may not hold true since acetyl kinase keeps the ATP/ADP ratio at 3000:1, which tends to stabilize actin tails for the entire duration of motility. For this reason, there is a mixed population of tail lengths according to the time of nucleation. So while tail length may not correlate with velocity, it will show us the entire trajectory of bacterial movement, including rotation. This will allow us to correlate the spiral trajectory of the actin tail with the rotation of the bacterium. We propose that the time window for motility can be extended by diluting the cytoskeletal components in the motility medium by varying amounts (1:1, 1:2 1:5) following the 4 C incubation. Previous studies have indicated that the average speed of a motile bead coated with ActA decreases when the concentration of crude extract is reduced. It was also found that the rate of symmetry breaking and hence tail formation increased dramatically when the extract was diluted to 40% of the original concentration (Cameron et al., 2000). It may be that by lowering the density of the actin gel which forms around the particle, the force required to break symmetry is reduced. This effect is also observed when varying the concentration of ActA on the surface of the bead (Noireaux et al., 2000). However, the same principle may not apply to our system since motility could be dependent on the diffusion of cytoskeletal components to the bacterial surface or the spontaneous association of protein complexes in solution. The asymmetric distribution of ActA on the surface of Listeria may actually minimize the effect of actin gel density and symmetry breaking. We then set out to determine what effect, if any, the extract concentration had on the efficiency of the system in terms of motile fraction and average tail length. This was accomplished by measuring the fraction of motile bacteria and the distribution of tail


30 lengths using the trace function in Image J (NIH website: The tail length, if averaged over a large population of filaments, should give us some indication of how fast the bacteria are moving and, inversely, the time window for observing motility. The relationship between actin gel density and the velocity of the motile object has already been established for VCA-coated beads (Bernheim-Groswasser et al., 2002), and it was determined that maximal velocity is achieved at the minimum actin gel density necessary to support motility. Thus, tail length should increase as we dilute the extract, until some critical concentration is reached at which point the tail length will begin to taper off. Once the bacterium breaks symmetry, the density of randomly-oriented filaments on its surface is largely irrelevant and does not impede the motion of the bacterium. At this point, bacteria velocity increases linearly with the extract concentration. Long Length-Scale Rotation In the final experiment, we will track the rotation of labeled Listeria by taking time-lapse images every 30 seconds in both phase contrast and fluorescein fluorescence for at least 15 minutes. These images will then be compiled into a video file so that we can observe full rotation of a bacteria with beads bound to its surface.


CHAPTER 4 MATERIALS AND METHODS Bacteria Preparation Overnight Culture Four separate cultures of Listeria monocytogenes was grown simultaneously in BHI (Brain-Heart Infusion) media supplemented with 5 g/ml erythromycin and 7 g/ml chloramphenicol at 37 C with 200 rpm shaking (Zeile et al., in press). For bacteria that were subsequently inactivated, they were then regrown 1:100 in two separate flasks containing 100 ml BHI media for an additional 6 hours or until the OD 600 reached a value of 0.6. At this point, bacteria were transferred to 4 plastic Ti-35 tubes and centrifuged at 7,000 rpm for 10 minutes with a JA-18 rotor. The bacterial pellets were then resuspended in 25 ml PBS and combined with another, to give two separate 50 ml bacterial suspensions. These were then washed twice in filtered PBS. After the second wash, the supernatant was removed and the bacterial pellets were resuspended in 50 ml 10 mM iodoacetate/PBS. The suspensions was incubated for 15 minutes at room temperature with shaking, then spun down in the JA-18 rotor at 7,000 rpm for 5 minutes. The supernatant was removed and the pellets washed twice in brain extract buffer (Yarar et al., 1999). After the second wash, the pellets were resuspended in 1/5 volume brain extract buffer with 20% glycerol, aliquoted, and stored at -70 C. All centrifugations were carried out at 4 C. For live bacteria, they were harvested immediately by centrifuging at 7,000 rpm for 5 minutes. The bacteria were then washed three times in volume PBS and resuspended 31


32 in 1/10 original volume brain extract buffer with 20% glycerol and stored at 4 C or used immediately (Zeile, unpublished). Fluorescent Labeling of Bacteria Ionic Adsorption Fluorescently labeled, carboxylated 500 nm polystyrene beads were diluted to 1/20 the original concentration in brain extract buffer. The bead mixture was then added in a 1:1 ratio with the bacteria, which was further diluted 2:3 in brain extract buffer. The entire suspension was then incubated for 10 minutes at room temperature (Robbins & Theriot, 2003). The bacteria were then washed twice in brain extract buffer by centrifugation (5 min, 7,000 rpm) or else added directly to the motility media. All centrifugations were carried out at 4 C. Aldehyde Coupling to Amine Groups Fluorescently labeled, 400 nm aldehyde-modified beads were diluted to 1/20 the original concentration in ddH 2 O and spun down at 10,000 rpm for 10 minutes in the Eppendorf centrifuge. The supernatant was removed and an equal volume of 20 mM sodium phosphate buffer (pH 6.5) was added to the pelleted beads. Bacteria were spun down for 5 minutes at 7,000 rpm and the supernatant removed so that an equal volume of sodium phosphate buffer could be added to the pelleted bacteria. The beads and bacteria were then combined in a 1:1 ratio and diluted 2:3 in sodium phosphate buffer. The entire suspension was incubated for 30 minutes at 4 C and the bacteria were washed twice in brain extract buffer by centrifugation (5 min, 7,000 rpm). Again, all centrifugations were carried out at 4 C.


33 Carbodiimide Coupling to Amine Groups Fluorescently labeled, carboxylated 500 nm beads were diluted to 1/20 the original concentration in ddH 2 O and spun down at 10,000 rpm for 10 minutes. The supernatant was removed and the beads were resuspended in an equal volume of 20 mM MES (2-[N-morpholino] ethane sulfonic acid) buffer (pH 6.0) supplemented with 0.4 mg/ml EDC and 0.6 mg/ml NHS. The mixture was incubated for 15 minutes at room temperature with rotation. The EDC was quenched with 0.2 M -mercaptoethanol for 5 minutes at room temperature with rotation. Meanwhile, the bacteria were spun down for 5 minutes at 7,000 rpm and resuspended in an equal volume of MES buffer. The beads and bacteria were then combined in a 1:1 ratio and diluted 2:3 in MES buffer. The mixture was incubated at 4 C for 2 hours with rotation (Grabarek & Gergely, 1990). The bacteria were then washed twice in brain extract buffer by centrifugation at 7,000 for 5 minutes. All centrifugations were carried out at 4 C. Preparation of Extracts Crude Extract Frozen rat or bovine brains were homogenized in liquid nitrogen with an equal amount (w/v) of brain extract buffer supplemented with 0.5 mM ATP, 1 mM leupeptin, 1 mM chymostatin, 1 mM pepstatin-A, and 0.5 mM DTT. The lysate was then transferred to pre-chilled tubes and centrifuged at 100,000 x g for 15 minutes at 4 C. The supernatant was then aliquoted, flash-frozen in liquid nitrogen, and stored at -70 C (Zeile et al., in press). High-Speed Supernatant We prepared supplemented brain extract buffer by adding 125 l 100 mM K + ATP, 25 l 10 mg/ml protease inhibitors (leupeptin, chymostatin, and pepstatin A), 25 l 0.5 M


34 DTT and 125 l 0.2 M PMSF to 25 ml brain extract buffer (pH 7.7). Crude extract was then diluted 1:10 in the supplemented brain extract buffer and centrifuged at 90,000 rpm for 1 hour at 4 C to remove cellular debris. The supernatant was transferred to 10 kD Centriplus filters and centrifuged at 4500 rpm for 4 hours or until the volume was just under 1 ml. The flow-through was discarded and the solution recovered by inverting the filter over a plastic tube and spinning at 3000 rpm for 2 minutes. The sucrose concentration was then adjusted to 0.2 M with 2 M sucrose and 1/20 volume acetyl kinase was added to the solution. The extract was then aliquoted (200 l per aliquot) to avoid repeated freezing/thawing which may cause protein degradation over time and stored at -20 C (Zeile et al., in press). Motility Assays A complete extract was prepared by combining 19.5 l HSS (high speed supernatant), 3 l acetyl kinase, 3 l protease inhibitors, 3 l 10 mM PMSF, 1.5 l 100 mM DTT and 3 l 1mg/ml rhodamine actin (Zeile et al., in press). The solution is then added in a 9:1 ratio to the bacteria and incubated at 4 C for 20 minutes, after which a 10 l sample is pipetted onto a glass slide and sealed with a coverslip. Microscopy and Imaging Software All slides were examined in phase contrast with a Nikon Diaphot inverted microscope and a 40X Nikon objective. An Hg Arc lamp (100 W) with FITC and rhodamine filters were used to observe fluorescence. A Hamamatsu chilled CCD camera was used to collect phase images at 0.03 s and fluorescence images at 0.1 s integration time. These images were subsequently viewed with PCTV Vision image software.


CHAPTER 5 RESULTS AND DISCUSSION Fluorescent Labeling of Bacteria Particle Dimensions and Fluorescence The 20 nm microspheres aggregated even when the buffer was diluted to low ionic strength (20 mM). Agitation by repeated vortexing and sonication was insufficient to disrupt these aggregates, which had a “raft” appearance when viewed under a 40X objective. Addition of surfactants increased the monodispersity of the beads but significant aggregation ( 40%) was still present even at 0.1% Tween. Furthermore, the 20 nm fluorescein microspheres showed spectral overlap with rhodamine such that “red beads” were detected even though no rhodamine particles were present in the sample (Figure 5-1). The 50 nm microspheres were monodisperse at low ionic strength (20 mM) and did not require the addition of surfactants. However, we found it difficult to detect these beads under the 40X objective, even in fluorescence. In theory, these beads should be visible and often were at high concentrations (4.62 x 10 12 particles/ml). At reduced concentrations ( 2.31 x 10 11 particles/ml), the beads became very difficult to observe with our microscopic system (Figure 5-2). This can be attributed to the high diffusivity of 50 nm beads, which move in and out of focus very rapidly such that only a small fraction of the bead population will be in focus at any given time. The smaller that population becomes, the lower the chance of observing a bead before it disappears. 35


36 Figure 5-1. Motile bacteria labeled with 20 nm particles by ionic adsorption. Listeria were incubated with 20 nm beads for 10-20 min at 22 C and washed twice in brain extract buffer. The bacteria were then incubated in motility medium for 20 min at 4 C and 10 minutes at 22 C. (A) Phase contrast 40X (B) Phase Contrast 40X (C) Rhodamine fluorescence 40X (D) Rhodamine fluorescence 40X (E) FITC Fluorescence 40X (F) FITC Fluorescence 40X Figure 5-2. Motile bacteria labeled with 50 nm particles by aldehyde coupling. Listeria incubated with 50 nm beads for 10-20 min at 22 C and washed twice in brain extract buffer. The bacteria were then incubated in motility medium for 20 min at 4 C and 10 min at 21 C.(A) Phase contrast 40X (B) Rhodamine fluorescence 40X (C) Fluorescein fluorescence 40X (D) Phase contrast 40X (E) Rhodamine fluorescence 40X We would expect that once a microsphere is fixed to a bacterium, its thermal fluctuations will be attenuated (equal to the thermal fluctuations of the bacterium). This would increase our chance of detecting the beads at lower concentrations, if not for the large size discrepancy between the microspheres and the bacteria (~ 1 m). This requires


37 further dilution of the microspheres in order to achieve an appropriate bead-bacteria (b/L) ratio. Depending upon the reaction conditions (incubation time, temperature, and availability of binding motifs on Listeria), a certain proportion of the microspheres will bind while the remainder are removed in the separation step, thus further reducing the bead concentration. This number will then be lowered 10-fold in motility assays. Given the reasons stated above, and despite various temperature regimes and reaction times, the bead concentration always remained below minimum detection levels. They were, however, observable under 100X fluorescence, but unfortunately due to the misalignment of the phase rings the actin tails were not detectable under the 100X phase objective. The 500 nm microspheres were easily detectable under the 40X objective both in phase contrast and fluorescence. They also remained monodisperse even at higher ionic strength (100 mM) and would pellet after a short incubation time at low speed (10 minutes at 10,000 rpm) in ddH 2 O, making it a relatively simple matter to wash the microspheres prior to incubation with the bacteria. However, this high sedimentation efficiency made more difficult to remove the unbound material during the separation step. These microspheres were better able to bind and stay bound due to the increase in surface area available for adsorption. Since more bonds could form between a single bead and a bacterium, a stable attachment was much more likely. However, we expect the collision frequency of the 50 nm microspheres to be ten times greater than that of the 500 nm microspheres at the same b/L ratio. This effect becomes negligible when the incubation time greatly exceeds the binding time ( B ), which is ~ 1 s for 50 nm microspheres and ~ 10 s for 500 nm microspheres. The exact value of B will vary


38 depending on the b/L ratio. In our experiments, the incubation time exceeded the binding time by several orders of magnitude, so the solution was at equilibrium. After carefully weighing the options, it was decided that the 500 nm microspheres were better suited for this application than the 50 or 20 nm beads. . Coupling Methods Aldehyde coupling yielded between 40-50% labeled Listeria when 500 nm microspheres were used at a b/L ratio of 10X. At higher concentrations, the average number of beads per bacterium would increase but the percentage of labeled beads would not. It was difficult to quantify this for the 20 nm and 50 nm beads since they were often undetectable. Ionic adsorption typically yielded 70-80 % labeled Listeria but had other problems which we will discuss shortly. The carbodiimide coupling procedure yielded 50-60% labeled Listeria, but only a small fraction was motile. Reaction Conditions As mentioned earlier, we expected longer incubation times to inhibit motility even for reactions carried out at 4 C. However, bacteria incubated at 4 C for 18 hours in 20 mM sodium phosphate buffer (pH 6.5) showed a 50 % increase in motility over bacteria incubated for 2 hours at 25 C, so clearly temperature is the determining factor. For this reason, all subsequent reactions were carried out at 4 C unless the incubation time was short (< 30 min). The incubation time was shortened to 10 minutes when using ionic adsorption as a coupling method (Robbins & Theriot, 2003), which had no appreciable effect on the yield of labeled bacteria (70-80 %). Wash and Separation Step A 500 nm polystyrene microsphere should pellet after 15 minutes at 9300 g (10,663 rpm for the Eppendorf centrifuge rotor), but this did not hold true for a particle


39 suspension in 0.2 M sodium phosphate buffer. Even at 14,000 rpm, it would often take hours for the beads to pellet. This posed a significant problem for washing the beads prior to coupling, but it did facilitate separation of the bacteria from the beads after coupling. Eventually, we found that the beads could be washed in ddH 2 O and would form a compact pellet after 10 minutes at 10,000 rpm. Also, the sedimentation rate of beads in low ionic strength buffer (20 mM) approached that in water. However, the microspheres did not pellet after 5 minutes at 7,000 rpm in 20 mM buffer. Free beads were still observed in solution even after repeated washings in low ionic strength brain extract buffer. To solve this problem, we took advantage of the relative low density of polystyrene (1.05 g/cm 3 ) by pipetting the sample onto a 20 % sucrose cushion which would allow only the bacteria to pellet. However, we found that after reacting the bacteria with the beads, their density would be reduced to an extent that they could no longer pellet in 20% sucrose. The bacteria that did pellet were mostly unlabeled, so we had inadvertently enriched the sample for unlabeled bacteria. Since it was not absolutely necessary to remove all the unbound beads in the sample, the sucrose cushion was eliminated and the bacteria washed by centrifugation at 7,000 rpm for 5 minutes. Another problem was encountered during the separation step which had to do with the wall effect. It was noticed that bacterial density was highly attritted following the low speed centrifugation, which was puzzling at first since the bacteria would form a compact pellet at the exact same speed, time, and concentration in 20 mM – 0.2 M buffer. Later on it was discovered that labeled bacteria would adhere to the wall of the tube due to the hydrophobic nature of the beads which were bound to its surface. This material was routinely “wicked off” with a Q-tip under the false assumption that it mainly consisted of


40 unbound beads. After examining a few samples in which the adherent material was not removed but instead resuspended, it was found that the bacterial density was greatly increased. Also, the proportion of labeled bacteria in solution increased dramatically, proving that bacterial loss was a direct result of bound beads adhering to the wall of the tube during centrifugation. It was further shown that this “wall effect” was more pronounced in bacteria labeled with aldehyde beads as opposed to carboxylate beads. In fact, it was originally thought that incubation in acidic buffer caused bacterial lysis, which would explain the fact that aldehyde coupling (carried out at pH 6.5) reduced the concentration of bacteria. So then why was this not observed with bacteria that had been reacted with carboxylate beads via carbodiimide coupling? It was speculated that these bacteria were protected from lysis by the presence of -mercaptoethanol in the reaction mixture (to quench the unreacted EDC). And while this assumption is probably false due to the fact that Listeria monocytogenes is a gram-positive organism with a rigid cell wall that is resistant to lysis, it does show that this effect was noticed prior to determining the true cause of bacterial loss and not merely an ex post facto observation. Motility Assays Coupling Method and Labeling Efficiency In this section we will compare the labeling efficiency and motile fraction obtained for various coupling regimes (adsorption, carbodiimide and aldehyde). In order to distinguish the efficacy of each coupling method, we calculated the fraction of bacteria with actin tails and the fraction of motile bacteria with beads bound. Since the labeling efficiency was previously determined in binding assays, we will investigate whether motility has any effect on the fraction of labeled bacteria.


41 Ionic adsorption Bacteria that had microspheres ionically adsorbed to their surface still exhibited motility which was comparable to unlabeled bacteria (Figure 5-3). Figure 5-3. Motile bacteria coupled to 500 nm particles by ionic adsorption. Phase contrast and fluorescent images of Listeria undergoing motility with beads ionically adsorbed to the surface. Bead adsorption does not appear to inhibit the motile fraction of bacteria. The overall labeling efficiency for bacteria that had been treated in a motility assay was 17%, nearly 60% less than that calculated in the binding assay (Table 5-1). One possible explanation for this is that the beads have a higher affinity for actin filaments than for bacteria, causing them to adhere preferentially to comet tails in the motility medium. Also, the fraction of motile bacteria with beads was calculated to be 56.2%, so beads bound to the bacterial surface do not appear to inhibit motility. The motile fraction of labeled bacteria was 16.1%. A high proportion of bacteria have 1-2 beads bound, while a few have 3 or 5+ beads bound. Also, the vast majority of motile bacteria have only 1 bead bound, with very few having 3 or more beads (data not shown). This may indicate that 3 or more beads could inhibit the formation of protein complexes on the bacterial surface.


42 Covalent coupling This next section deals with Listeria that were labeled using a covalent crosslinking procedure, which was found to alter the morphology of the bacteria under phase contrast microscopy. The bacteria exhibited a rough, abraded surface which appeared to have been chemically degraded by either the crosslinking agents or the buffer in which the Listeria had been incubated. This was thought to correlate with a change in surface properties which may have inhibited motility. The results in Figure 5-4 indicate that both carbodiimide and aldehyde coupling have a negative effect on motility. The majority of labeled bacteria have either 1 bead or 5+ beads attached to their surface, with very few beads having 2-3 beads attached. The vast majority of motile bacteria also had 1 bead bound and a few with 5+ beads bound, which is similar to the trend observed in the general population only more pronounced. Figure 5-4. Motile bacteria coupled to 500 nm particles by covalent coupling. Listeria were reacted with 500 nm carboxylate microspheres using a covalent crosslinking procedure (see Materials and Methods), then incubated in motility medium for 20 minutes at 4 C and 10 minutes at 22 C. (A) Phase contrast 40X, Listeria labeled with carboxylate beads by carbodiimide coupling. (B) FITC fluorescence 40X, Listeria labeled with carboxylate beads by carbodiimide coupling. (C) Phase contrast 40X, Listeria labeled with aldehyde beads. (D) FITC fluorescence, Listeria labeled with aldehyde beads.


43 Table 5-1. Calculated parameters for different coupling methods Label Efficiency (%) labeled/motile (%) motile/labeled (%) Adsorption 16.9 10.8 56.2 41.8 16.1 12.3 Carbodiimide 8.1 3.8 19.1 17.8 24.1 18.9 Aldehyde 4.8 3.9 50 40 75 49 The labeling efficiency of bacteria that were covalently coupled to carboxylate beads and then treated in a motility assay was 8.1 %, which is 50% less than the labeling efficiency in the binding assay and 10% less than the ionically-coupled bacteria. The aldehyde-coupled bacteria had an even lower labeling efficiency of 4.8%. The fraction of motile bacteria with beads was 19.1% for the carboxylate beads and 50% for the aldehyde beads. And finally, the motile fraction of labeled bacteria was 24.1% for covalently-coupled carboxylate beads, a value which is similar to that for ionically-coupled bacteria (16.1%). This seems to suggest that low pH treatment is not a significant impediment to Listeria motility. Optimizing Growth Conditions The ability to assemble actin tails varied wildly from one bacterial preparation to the next, even when individual cultures were incubated for the same length of time. As illustrated by the growth curves shown in Figure 5-5, the optical density (OD 600 ) did not correlate with the incubation time between samples, indicating that optical density, instead of time, should be used as a measure of when to harvest the bacteria to achieve optimal motility. rep = doubling time of Listeria; A 0 = start concentration ODt()A02trep OD t()A02t1.934


44 (hrs)OD(600) Series1 Series2 Series3 Series4 S1a S1b S1c S1d Figure 5-5. Growth curve for Listeria monocytogenes. Clearly the first two samples have achieved a higher density than the latter two, which could result from a discrepancy in growth rate (1/) or starting concentration (A0). Assuming that the replication rate is constant across all samples, which is most likely the case since this is an intrinsic property of the bacterium that should depend only on temperature and the availability of nutrients, we can attribute this batch variability to the starting concentration, which depends on the transfer of bacteria from the inoculation loop to the culture media. Using the trendline function in Excel to compute the slope and intercept of the first two curves in Figure 5-5, we can find analytical solutions to the doubling time for Listeria (see Table 5-2). The two values were then averaged and the result substituted for rep . So we found that the doubling time for Listeria is ~ 2 hours, which allows us to predict how long to incubate the bacteria before they reached a desired OD 600 . Table 5-2. Determination of doubling time for Listeria in log phase slope y-intercept y(14) t1 2y(14) t2 dt Series 1 0.3022 -3.5813 0.635 14 1.27 16.053 2.053 Series 2 0.3009 -3.6747 0.542 14 1.084 15.815 1.815 Series 3 0.2381 -3.1472 1.934 In this next experiment, samples were collected every hour from 12-18 hours and the growth rate was monitored by spectroscopy. Samples that failed to reach a critical density by 12 hours were not discarded but allowed to grow until the end of the


45 incubation period. For every hour until the 18 th hour, we harvested 1 ml from each sample that with an optical density that was 0.1 units. In the end, we wound up with 6 total samples since 2 had failed to reach 0.1 units by 18 hours. It is clear from Figure 5-6 that bacterial density increases exponentially between 12 and 15 hours, then plateaus between 15 and 16 hours. At 16 hours, the density reaches the point where the nutrients become limiting, and the bacteria start to die off rapidly between 16-17 hours. Surprisingly, this sharp decline in bacterial density did not correspond to a drop in motile number per FOV or motile fraction (Figure 5-7). In fact, motile fraction peaks at 13 hours and 17 hours, whereas the motile number per FOV plateaus between 14-17 hours and increases slightly at 18 hours. We also found that there was a negative correlation between bacterial density and motile fraction (-0.96, N = 6) in the 13-18 hour time interval. There is no correlation between bacterial density and motile number per FOV or motile fraction and motile number per FOV. 0204060801001201401601801011121314151617181920time (hrs)Bacteria Count per FOV Figure 5-6. Time-dependent changes in bacterial count per FOV. This plot shows the bacterial count as a function of incubation time. At 12-15 hours the bacteria experience an exponential growth phase followed by a stationary phase in which nutrients are limited. Finally, the density starts to decline as bacteria die off faster than they are replaced.


46 05101520253010111213141516171819time (hrs)Motile #/FOV, % Motile # % Motile Figure 5-7. Time-dependent changes in motility. This plot shows the change in motility with culture time. Motile number per FOV increases between 12 and 14 hours and then plateaus between 14 and 17 hours. Motile fraction (% motile) has two peaks at 13 and 17 hours, and is negatively correlated with bacterial density (-0.96). There was no apparent correlation between motile number per FOV and culture time or motile fraction and culture time over the 12-18 hour period (Figure 5-7), suggesting that time is not a good indicator of motile potential. This variability stems from the fact that bacterial density is a function of the start concentration, which is directly related to the number of bacteria that adhere to the inoculation loop. Bacterial count per FOV increased steadily with optical density as depicted in Figure 5-8, but declined between 1.6-1.8 indicating that optical density is not a perfect indicator of bacterial density during the logarithmic decline phase. Light scattering off of dead bacteria and cellular debris contributes to the optical density of the sample without a corresponding increase in bacterial density.


47 0-0.20.2-0.40.4-0.60.6-0.8 0.8-1.01.0-1.21.2-1.41.4-1.61.6-1.80204060801001201401601801Optical DensityBacteria count/FO V Figure 5-8. Bacterial count per FOV as a function of optical density. In this graph, the bacterial count per FOV was plotted against optical density. There is a general increase in bacterial concentration between 0 and 1.6, after which the number of bacteria per FOV starts to decline. 0-0.20.2-0.40.4-0.60.6-0.8 0.8-1.01.0-1.21.2-1.41.4-1.61.6-1.80510152025301Optical DensityMotile #/FOV Figure 5-9. Motile number per FOV as a function of optical density. The number of motile bacteria per FOV increased between 0 and 0.6, then leveled off between 0.8 and 1.8.


48 0-0.20.2-0.40.4-0.60.6-0.80.8-1.01.0-1.21.2-1.41.4-1.61.6-1.80510152025301Optical Density% motile Figure 5-10. Motile fraction as a function of optical density. The motile fraction was fairly constant between 0.2 and 1.8, indicating that motile fraction is not a good indicator of the robustness of the sample. The number of motile bacteria per FOV increases with optical density between 0 and 0.6 and then reaches a plateau between 0.6 and 1.8 (Figure 5-9). There is a slight increase in the number of motile bacteria per FOV between 1.4 and 1.6, but this may just be statistical noise. Motile fraction, on the other hand, remains constant between 0.2 and 1.8 with a slight peak between 0.6 and 0.8 (Figure 5-10). Overall, this indicates that motile fraction may not correlate with the probability of finding a motile bacterium in the sample. Optimizing Filament Nucleation As stated earlier, the bacteria are incubated in motility medium at 4 C for some arbitrary length of time to increase the rate of actin tail nucleation by allowing the filaments growing on the bacterial surface to congeal into an actin gel. As lateral stresses within the actin gel reach some critical point, the filaments will align themselves in a


49 unilateral direction at the distal end of the bacterium in a process called symmetry breaking. In this experiment, we evaluate the effect the 4 C incubation on both motile fraction and average tail length (Figure 5-11). For each of these regimes, we compared samples in which the extract concentration was diluted 1:3 to undiluted samples. Figure 5-11. Variation of tail length with incubation time on ice. (A) Phase contrast image of a motile bacterium incubated for 10 minutes on ice. (B) Phase contrast image of a motile bacterium incubated for 20 minutes on ice. (C) Phase contrast image of a motile bacterium incubated for 30 minutes on ice. Note that the tail length appears to decrease with increasing incubation time for these bacteria. 0246810121405101520253035time (min)Motile number/FOV Diluted 1:3 Undiluted Figure 5-12. Motile number per FOV as a function of 4 C incubation time. The motile number per FOV was fairly constant as a function of incubation time in the diluted sample, but decreases parabolically with incubation time in the undiluted sample.


50 0510152025303505101520253035time (min)Motile fraction (% ) Diluted 1:3 Undiluted Figure 5-13. Motile fraction as a function of 4 C incubation time. There is no apparent trend in motile fraction as the incubation time is varied, but the curve for the diluted sample has a parabolic shape with its vertex at 20 min. Assuming a parabolic path for the undiluted sample, the vertex would be at 25 min. 05101520253005101520253035timetail length (um) Diluted 1:3 Undiluted Figure 5-14. Tail length as a function of incubation time. We compared the average tail length over the three incubation times. Assuming a parabolic path for the diluted sample, the vertex would be at 17 min, whereas the undiluted sample decreases linearly (R 2 = 0.998) with incubation time.


51 The motile number per FOV decreases in a parabolic manner with incubation time in the undiluted sample, but is a fairly constant function of incubation time in the diluted sample (Figure 5-12). In the undiluted samples, the actin gel that forms around the bacterium becomes more rigid with increasing incubation time, so the rate of symmetry breaking is expected to decrease accordingly. However, in the diluted samples, the actin gel is interrupted by the lowered viscosity of the sample, so there is little correlation between incubation time and symmetry breaking. These observations concur with the results by Cameron et al. (2001), who found that the fraction of crude extract was negatively correlated with the fraction of tails that formed on ActA-coated polystyrene beads. Furthermore, the rate of symmetry breaking of ActA-coated beads was 100 % in a purified protein solution containing all the necessary elements for motility, and the time required for symmetry breaking was 2-3 minutes. In cell extracts, the rate of symmetry breaking dropped to 10 % and the time required was 20 minutes (Bernheim-Groswasser et al., 2002). This effect has been attributed to the increased viscosity of cell extracts over purified proteins after a quantitative analysis of Listeria movement in fibroblasts by Giardini and Theriot (2001). The motile fraction is consistently higher in the diluted sample than the undiluted sample, although both follow a parabolic trajectory with a vertex at 20 min for the diluted sample and 24 min for the undiluted sample (Figure 5-13). Based on the error calculations for this data, this could easily be attributed to random noise in the diluted sample. On the other hand, the motile fraction in the undiluted sample clearly declines after 10 minutes.


52 The average tail length decreased in a linear fashion with incubation time in the undiluted sample (Figure 5-14), which could indicate that the time required for symmetry breaking depends on the incubation time. As the incubation time increases, the actin gel formed around the bacterium thickens and requires more time to break symmetry. The curve for the diluted sample was parabolic with a vertex at 17 min, suggesting that the time required for symmetry breaking is minimized at ~ 17 min under these conditions. Optimizing Extract Concentration In this experiment we varied the extract concentration while holding incubation time constant to determine the effect on motile fraction and average tail length. Our goal is to maximize the average tail length and the total number of long-length tails, which will raise the probability of observing bacterial rotations during motility. On the other hand, we need to reduce the rate of actin tail elongation so there is sufficient time to record bacterial motility. 02040608010012014016018000.511.522.533.544.5Dilution FactorBacterial Denstiy Figure 5-15. Number of bacteria per FOV as a function of dilution factor. The bacterial concentration is inversely related to dilution factor.


53 -505101520253000.511.522.533.544.5Dilution FactorMotile #/FOV,% Motile Number/FOV Motile Fraction (%) Figure 5-16. Motility as a function of dilution factor. The motile number per FOV decreases with dilution factor while the motile fraction peaks around 2. The number of motile bacteria per FOV decreases steadily with dilution factor (Figure 5-15), but not inversely as bacterial concentration does. In fact, the curve of motile number per FOV is concave downward from 1:1 to 1:2 and while the curve of bacterial number per FOV is convex downward from 1:1 to 1:2, indicating that motility actually increases between 1:1 and 1:2. Furthermore, the motile fraction increases between 1:1 and 1:2, then decreases sharply between 1:2 and 1:4 (Figure 5-16). These results suggest that peak motility occurs in the 1:2 dilution group, which is consistent with the results of Cameron et al. (2000) who found that 40% extract increased the rate of symmetry breaking.


54 02040608010012014016018000.511.522.533.544.Dilution FactorAverage Tail Length 5 Figure 5-17. Relationship between average tail length and dilution factor. In this figure, tail length is plotted against dilution factor, which shows a general decline in tail length with dilution factor. This suggests that diluting the extract does in fact slow down the bacteria. Average tail length decreased with dilution factor in a sigmoidal fashion (Figure 5-17), similar to the curve for motile number per FOV. This indicates that the rate of actin polymerization, which is directed related to average tail length (Theriot et al., 1992), is indeed slowed down by diluting the extract. Therefore, the time window for observing motility is increased as the extract concentration is lowered. The next sets of figures are histograms of the tail length distribution for three of the five dilutions. All four samples were lumped together to calculate the distribution, so we can determine which dilution factor is associated with the longest tails. The undiluted sample has a peak between 5-9.9 m and a somewhat smaller peak between 10-15 m and 20-24.9 m, but overall the distribution of tail lengths is fairly wide (Figure 5-18). The sample diluted 1:2 had one very large peak at 10-14.9 m and a more narrow distribution of tail lengths (Figure 5-19). The sampled diluted 1:3 peaked at < 5 m with an even narrower distribution of tail lengths (Figure 5-20). These results indicate that the


55 highest number of long-length tails ( 10 m) is achieved in the 1:2 dilution group, which would afford the best opportunity for observing multiple bacterial rotations during motility. 0123456< 5.05-9.910-14.915-19.920-24.9tail length (um)Number of tails Figure 5-18. Distribution of tail lengths (1:1). The first peak occurs between 5-9.9 m and then falls off with increasing tail length. However, there is a second peak at 20-25 m. 0123456789<5.05-9.910-14.915-19.920-24.9tail length (um)Number of Tails Figure 5-19. Distribution of tail lengths (1:2). A single large peak occurs at 10-14.9m with very few filaments having tail lengths that are shorter or longer.


56 0246810121416<5.05-9.910-14.915-19.920-24.9tail length (um)Number of tails Figure 5-20. Distribution of tail lengths (1:3). This sample has a single large peak at < 5.0 m with no filaments longer than 10 m.


CHAPTER 6 CONCLUSIONS AND IMPLICATIONS The ultimate goal of this study was to optimize bacteria labeling with fluorescent beads and to recapitulate motility with these labeled bacteria for the purpose of tracking their rotational movements. In this next section, we will discuss how successful we were in these efforts, and the resulting implications for biophysical models of actin motility. Bacteria Labeling with Fluorescent Particles There are many factors to consider in attaching a large, biologically inert particle to the surface of a bacterium. The unique arrangement of surface functionalities and proteins on the bacterial surface is essential for actin filament assembly and comet tail formation, so anything that interferes with this process is counterproductive. Our data indicate that 500 nm particles are more frequently observed on the bacterial surface than 20 or 50 nm beads, which may be caused by microscopic limitations rather than some intrinsic property of the microspheres themselves. In this study, we have compared the binding potential of three different coupling methods, each with its own set of advantages and caveats. The first method to be applied was ionic adsorption, which can be carried out at physiological pH but does not form a stable attachment between bead and bacterium. This leads to significant problems when tracking motile Listeria, since the bead would often detach before a full rotation could be documented. We found that ionic adsorption yields the highest proportion of labeled bacteria, and optimal motility is achieved with covalent coupling when the bacteria are incubated with the beads at 4 C for 2 hours. The separation step enriched for unlabeled 57


58 bacteria and reduced the bacterial concentration in the sample, but was necessary to remove unbound particles which could interfere with observing motility. The two covalent coupling methods (carbodiimide and aldehyde) required a low pH incubation to accelerate nucleophilic attack on amino groups at the bacterial surface. This can adversely affect Listeria motility by causing certain proteins on its surface, such as ActA, to denature and become inactive. However, it was determined that low pH treatment did not inhibit motility by comparing the motile fraction of labeled bacteria of the ionic adsorption group to that of the carbodiimide group. These values were roughly the same, and in fact the aldehyde group had a higher motile fraction of labeled bacteria than both the ionic adsorption and carbodiimide group. However, this only proves that the coupling method did not inhibit comet tail nucleation, the effect on average tail length and filament density has yet to be examined. The inhibition of motility may be the result of surface effects, since it was noticed that the number distribution of beads on bacteria was much more narrow among motile bacteria than for the general population. In fact, the distribution seems to focus around 1-2 beads for the motile bacteria, while the general population had a higher proportion of bacteria with 3 or more beads. We also observed bacteria crosslinking at high bead concentrations, since the beads are uniformly coated with functional groups that can bind two bacteria at once. This occurred at lower concentrations for beads that were covalently coupled to bacteria than for those that were ionically adsorbed. Crosslinked bacteria were often removed from solution since they had a tendency to adhere to the walls of the tube. Also, highly crosslinked bacteria-bead aggregates did not form comet tails and looked more like


59 artifacts. The incidence of artifacts in binding and motility assays increased ten-fold in covalently coupled bacteria-bead suspensions than in ionically adsorbed mixtures. Quenching with hydroxylamine had no effect on crosslinking when added to solution after the 2 hour incubation at 4 C. It may be possible to reduce crosslinking by adding the quenching agent along with the reactants, although this may also inhibit binding between the beads and bacteria. Overall, ionic adsorption has shown the most promise as a coupling method for bacteria labeling with fluorescent beads. The main obstacle in tracking motile bacteria with ionically adsorbed beads is their tendency to dissociate from the surface. We may be able to compensate for this by increasing the solution viscosity, which should dampen thermal fluctuations which eventually lead to bead dissociation. We can also reduce the probability of bead dissociation by minimizing the illumination time. This can be accomplished by switching the lamp off after recording an image, and leaving it off for the duration of the time interval between image captures. Recapitulation of Motility with Labeled Bacteria The use of live bacteria in motility assays was designed to enhance bacterial motility. However, given the variability between one bacterial preparation and the next, we needed to determine which parameters are indicative of bacterial performance in motility assays. Then we had to optimize these parameters to achieve reliable motility on each and every trial. The first two parameters analyzed were incubation time and optical density. We determined that incubation time was not a good indicator of motility, but the optical density would often predict how successful the bacteria were in a motility assay. The optical density for which bacterial motility was optimized ranged from 1.4 – 1.6 units.


60 Next, we examined the relationship between the 4 C incubation time and motile fraction as well as average tail length. It was determined that the number of motile bacteria per FOV decreased with incubation time if the sample was undiluted, but remained constant if the sample was diluted. From this we concluded that symmetry breaking is not inhibited by long incubation times on ice if the sample is diluted to 1/3 the original concentration. Average tail length decreases with incubation time in undiluted samples, implying that the time required for actin tail nucleation increases with incubation time. One possible explanation for this is that the actin gel thickness increases with incubation time, which heightens the energy barrier to symmetry breaking and comet tail formation. Furthermore, the average tail length peaks at 17 min in the diluted sample, so the thickness of the actin gel may be optimal for symmetry breaking at this particular time point. And finally, we analyzed the effect of extract concentration of motile fraction and average tail length, as well as the tail length distribution. Motility increases with dilution factor between 1 and 2 but declines sharply thereafter, indicating that symmetry breaking and actin tail nucleation is optimal at ~ 50% extract concentration. The average tail length decreases steadily with dilution factor, which increases the time window for observing motility. The tail length distribution shows that the highest number of long-length tails ( 10 m) occurs when at ~ 50% extract concentration. Implications for Tracking Listeria Rotation Ionically adsorbed beads have a limited residence time on the bacteria surface. We have already suggested a few different approaches to solving this problem, but we may also need to adjust the system further if we are going to observe long-length scale rotation. So far we have been adding the beads to the bacteria before the motility assay,


61 when the bacteria are still free in solution. Once these bacteria are incubated at 4 C in the motility medium, they start to slow down and form comet tails, which may allow them to form a more stable bond with beads via ionic adsorption. One approach may be to incubate bacteria for 20 minutes in the motility medium on ice, then add an equal volume of beads diluted 1:500 in brain extract buffer. After incubating the suspension at room temperature for 10 minutes, we can even spin down the components to remove free beads from the solution, since the bacteria should have already formed actin tails. Then we can simply resuspend the bacteria in extract and observe motility. Otherwise, we can dilute the beads even further in the brain extract buffer and eliminate the separation step. Another approach to this experiment is to incubate the bacteria with an antibody to listeriolysin-O, a protein that disrupts the membrane of endocytotic vesicles and allows the bacterium to escape into the cytoplasm of host cells. Beads that are covalently conjugated to protein A, which binds to the F c region of antibodies, are commercially available and can be used to form a stable attachment between the bead and bacterium. This would allow us to achieve covalent coupling without low pH treatment, and we could eliminate crosslinking by controlling the b/L ratio. If the bacteria concentration is kept relatively low compared to the beads during incubation, the probability of a bound bead encountering another bacterium is negligible. Also, we could quench the beads after incubating them with the bacteria by adding the antibody to solution. Implications for Actin-Based Motility We have been able to track unlabeled bacteria that have right-handed, spiral trajectories and observed multiple iterations of cyclical motion as these bacteria undergo actin-based motility. Since the actin tail is supposed to reflect the trajectory of the motile bacterium, we can assume that bacteria associated with spiral tails are undulating as well.


62 The use of beads as position markers will allow us to track the bacterial positions analytically so that we can compute the sinusoidal variation in the distance of an attached bead from the longitudinal axis of the bacterium. We can then plot the trajectory of the actin tail based on its period and amplitude, which should closely match that of the motile bacterium. Figure 6-1. Spiral trajectories of actin tails attached to L. monocytogenes. The top panels show motile Listeria with corkscrew tails. The bottom panel is an illustration of the right-handed helical trajectory of a bacterium with two beads bound to its surface. It is currently unknown whether the frequency and amplitude of rotation in motile Listeria is determined by filament density, the degree of crosslinking between filaments, the number of contact points between the filaments and the bacterial surface, or some combination thereof. It would be reasonable to assume that a rigid network of actin filaments would impart more resistance to torsion, reducing the frequency of rotation with respect to distance. However, as these forces start to accumulate over many cycles of monomer addition, the resultant “kink” in the filament network would have a higher amplitude than a more compliant actin tail.


APPENDIX REACTION MECHANISMS FOR COVALENT COUPLING METHODS Carbodiimide Coupling Step 1 Step 2 Figure A-1. Reaction scheme for carbodiimide coupling. This figure illustrates the acid-catalyzed carbodiimide coupling reaction. CH 3 CH 2 – N = C = N – (CH 2 ) 3 – N + H + Cl H + + C CH 3 CH 3 OH pH 6.0 O Carboxylate bead EDC CH 3 CH 2 – NH C = N – (CH 3 ) – N + H + Cl N + CH 3 CH 2 + O O H + O C OH pH 6.0 NHS O-acylisourea O + CH 3 CH 2 – NH C NH – (CH 2 ) 3 – N CH 3 CH 3 O O C O N O O More stable in H 2 O C N O H O C O N O NH 2 + NH 2 O 63


64 This reaction is carried out in MES (2-[N-morpholino] ethane sulfonic acid) buffer at pH 6.0 to facilitate nucleophilic attack by a carboxyl group on the imide group of EDC (1-ethyl-3-[3-dimethylaminopropyl]-carbodiimide). This results in the formation of an O-acylisourea which is partially stabilized by the slightly acidic pH, but also requires the addition of NHS (N-hydroxylsuccinimide) which attacks the O-acylisourea and replaces the EDC by nucleophilic substitution in order to achieve stability in aqueous solution. A reducing agent, B-mercaptoethanol, is added to the reaction mixture after a 15 minute incubation at room temperature to quench the unreacted EDC before proceeding to step 2 of the reaction. The NHS-conjugated beads are then reacted with the Listeria in MES buffer at pH 6.0 for 2 hours at 4 C. This reaction was carried out at room temperature in the original protocol but was modified due to the temperature sensitivity of ActA. However, it was never actually shown that a 2 hour incubation at room temperature inhibits motility in this case, it was only extrapolated from past experience with aldehyde coupling. Aldehyde Coupling In our first experiments, we used 500 nm rhodamine fluorescent beads that were modified to have aldehyde groups on their surface. These were then reacted with the Listeria at a fixed ratio, for instance 10 beads per bacterium. Listeria express different proteins on their surface and those proteins have primary amines (lysine, asparagine, arginine, etc.) which will react with aldehyde groups at slightly acidic pH. We used sodium phosphate buffer at pH 6.5 to carry out this reaction, but the ideal pH would have been 4-5. But since we’re using Listeria, we need to work in the physiological pH range to avoid protein degradation, so the reaction takes longer than it would at its ideal pH. So these two come together and form a Schiff base, or an imine.


65 Figure A-2. Reaction scheme for aldehyde coupling. This figure illustrates the acid-catalyzed aldehyde coupling reaction. C O Rhodamine fluorescent NH 2 H NH 2 Listeria + C N H 500 nm Schiff Base – Imine Reduction Step (Optional) C N H Schiff Base NaBH 3 CN + CH 2 N H Stable 2’-amine Sodium Cyanoborohydride pH 6.5 H 3 O +


LIST OF REFERENCES Alberts, B., D. Bray, J. Lewis, M. Raff, K. Roberts, and J. D. Watson. 1994. Molecular biology of the cell. Garland Publishing, New York. Bear, J. E. T. M. Svitkina, M. Krause, D. A. Schafer, J. J. Loureiro, G. A. Strasser, I. V. Maly, O. Y. Chaga, J. A. Cooper, G. A. Borisy, and F. B. Gertler. 2002. Antagonism between Ena/VASP proteins and actin filament capping regulates fibroblast motility. Cell. 100:509-521. Bernheim-Groswasser, A., S. Wiesner, R. M. Golsteyn, M. F. Carlier, and C. Sykes. 2002. The dynamics of actin-based motility depend on surface parameters. Nature. 417:308-311. Cameron, L. A., P. A. Giardini, F. S. Soo, and J. A. Theriot. 2000. Secrets of actin-based motility revealed by a bacterial pathogen. Nature Rev. Mol. Cell Biol. 1:110-119. Cameron, L. A., T. M. Svitkina, D. Vignjevic, J. A. Theriot, and G. G. Borisy. 2001. Dendritic organization of actin comet tails. Curr. Biol. 11:130-135. Carlier M. F. and D. Pantaloni. 1997. Control of actin dynamics in cell motility. J. Mol. Biol. 269:459-467. Derry, J. M., H. D. Ochs, and U. Francke. 1994. Isolation of a novel gene mutated in Wiskott-Aldrich syndrome. Cell. 78:635-644. Dickinson, R. B. and D. L. Purich. 2002. Clamped-filament elongation model for actin-based motors. Biophys. J. 82:605-617. Dickinson, R. B., L. Caro, and D. L. Purich. 2004. Force generation by cytoskeletal filament end-tracking proteins. Biophys. J. 87:2838-2854. Fehrenbacher, K., T. Huckaba, H. C. Yang, I. Boldogh, and L. Pon. 2003. Actin comet tails, endosomes and endosymbionts. EMBO J. 206:1997-1984. Fischer, M., S. Kaech, U. Wagner, H. Brinkhaus, and A. Matus. 2000. Glutamate receptors regulate actin-based plasticity in dendritic spines. Nat Neurosci. 3:887-894. Giardini, P. A., and J. A. Theriot. 2001. Actin-based motility is sufficient for bacterial membrane protrusion formation and host cell uptake. Cellular Microbiology. 3:633-647. 66


67 Goldberg, M. B. and J. A. Theriot. 1995. Shigella flexneri surface protein IcsA is sufficient to direct actin-based motility. Proc. Natl. Acad. Sci. USA. 92:6572-6576. Goldberg, M. B. 1997. Shigella actin-based motility in the absence of vinculin. 1997. Cell Mot. Cytoskel. 37: 44-53. Goldberg, M. B. 2001. Actin-based motility of intracellular microbial pathogens. Microbiol. Mol. Biol. Rev. 65:595-626. Gouin, E., H. Gantelet, C. Egile, I. Lasa, H. Ohayon, V. Villiers, P. Gounon, P. J. Sansonetti, and P. Cossart. 1999. A comparative study of the actin-based motilities of the pathogenic bacteria Listeria monocytogenes, Shigella flexneri and Rickettsia conorii. J. Cell Sci. 112:1697-1708. Gouin, E., C. Egile, P. Dehoux, V. Villiers, J. Adams, F. Gertier, R. Li, and P. Cossart. 2004. The RickA protein of Rickettsia conorii activates the Arp2/3 complex. Nature. 427:457-461. Grabarek, Z. and J. Gergely. 1990. Zero-length crosslinking procedure with the use of active esters. Anal. Biochem. 185:131-135. Kelleher, J. F., S. J. Atkinson, and T. D. Pollard. 1995. Sequences, structural models, and cellular localization of the actin-related proteins Arp2 and Arp3 from Acanthamoeba. J. Cell Biol. 131:385-397. Kocks, C., J. B. Marchand, E. Gouin, H. d’Hauteville, P. J. Sansonetti, M. F. Carlier, and P. Cossart. 1995. The unrelated surface proteins ActA of Listeria monocytogenes and IcsA of Shigella flexneri are sufficient to confer actin-based motility on Listeria innocua and Escherichia coli respectively. Mol. Microbiol. 18:413-423. Kuo, S. C. and J. L. McGrath. 2000. Steps and fluctuations of Listeria monocytogenes during actin-based motility. Nature. 407:1026-1029. Labrec, E. H., H. Schneider, T. J. Magnani, and S. B. Formal. 1964. Epithelial cell penetration as an essential step in the pathogenesis of bacillary dysentery. J. Bacteriol. 88:1503-1518. Laine, R. O., W. L. Zeile, F. Kang, D. L. Purich, and F. S. Southwick. 1997. Vinculin proteolysis unmasks an ActA homolog for actin-based Shigella motility. J. Cell Biol. 138:1255-1264. Lasa, I., V. David, E. Gouin, J. B. Marchand, and P. Cossart. 1995. The amino-terminal part of ActA is critical for the actin-based motility of Listeria monocytogenes; the central proline-rich region acts as a stimulator. Mol. Microbiol. 18:425-436.


68 Lauer, P., J. A. Theriot, J. Skoble, M. D. Welch, and D. A. Portnoy. 2001. Systematic mutational analysis of the amino-terminal domain of the Listeria monocytogenes ActA protein reveals novel functions in actin-based motility. Mol. Microbiol. 42:1163-1177. Li, R. 1997. Bee1, a yeast protein with homology to Wiskott-Aldrich syndrome protein, is critical for the assembly of cortical actin cytoskeleton. J. Cell Biol. 136:649-658. Loisel, T. P., R. Boujamaa, D. Pantaloni, and M. F. Carlier. 1999. Reconstitution of actin-based motility of Listeria and Shigella using pure proteins. Nature. 401:613-616. Machesky, L. M., R. D. Mullins, H. N. Higgs, D. A. Kaiser, L. Blanchoin, R. C. May, M. E. Hall, and T. D. Pollard. 1999. Scar, a WASp-related protein, activates nucleation of actin filaments by the Arp 2/3 complex. Proc. Natl. Acad. Sci. USA. 96:3739-3744. Makino, S., C. Sasakawa, T. Kamata, and M. Yoshikawa. 1986. A genetic determinant required for continuous reinfection of adjacent cells on a large plasmid in Shigella flexneri. Cell. 46:551-555. Marchand, J. B. D. A. Kaiser, T. D. Pollard, and H. N. Higgs. 2001. Interaction of WASP/Scar proteins with actin and vertebrate Arp2/3 complex. Nat. Cell Biol. 3:76-82. Miki, H., K. Miura, and T. Takenawa. 1996. N-WASP, a novel actin-depolymerizing protein, regulates the cortical cytoskeleton rearrangement in a PIP2-dependent manner downstream of tyrosine kinases. EMBO J. 15:5326-5335. Miki, H., S. Suetsugu, and T. Takenawa. 1998. WAVE, a novel WASP-family protein involved in actin reorganization by Rac. EMBO J. 17:6932-6941. Miki, H., and T. Takenawa. 1998. Direct binding of the verprolin-homology domain in N-WASP to actin is essential for cytoskeletal reorganization. Biochem. Biophys. Res. Commun. 243:73-78. Mogilner, A. and G. Oster. 1996. Cell motility driven by actin polymerization. Biophys. J. 71:3030-3045. Mogilner, A. and L. Edelstein-Keshet. 2002. Regulation of actin dynamics in rapidly moving cells: A quantitative analysis. Biophys. J. 83:1237-1258. Mogilner, A. and G. Oster. 2003. Force generation by actin polymerization II: The elastic ratchet and tethered filaments. Biophys. J. 84:1591-1605. Noireaux, V., R. M. Golsteyn, E. Friederich, J. Prost, C. Antony, D. Louvard, and C. Sykes. 2000. Growing an actin gel on spherical surfaces. Biophys. J. 78:1643-1654.


69 Pantaloni, D., M. F. Carlier, M. Coue, A. A. Lai, S. L. Brenner, and E. D. Korn. 1984. The critical concentration of actin in the presence of ATP increases with the number concentration of filaments and approaches the critical concentration of actinADP. J. Biol. Chem. 259:6274-6283. Pantaloni, D. R. Boujemaa, D. Didry, P. Gounon, and M. F. Carlier. 2000. The Arp2/3 complex branches filament barbed ends: functional antagonism with capping proteins. Nat. Cell Biol. 2:385:391. Robbins, J. R. and J. A. Theriot. 2003. Listeria monocytogenes rotates about its long axis during actin-based motility. Curr. Biol. 13:R754-R756. Rohatgi, R., L. Ma, H. Miki, M. Lopez, T. Kirchhausen, T. Takenawa, and M. W. Kirschner. 1999. The interaction between N-WASP and the Arp2/3 complex links Cdc42-dependent signals to actin assembly. Cell. 97:221-231. Samarin, S., S. Romero, C. Kocks, D. Didry, D. Pantaloni, and M. F. Carlier. 2003. How VASP enhances actin-based motility. J. Cell. Biol. 163: 131-142. Skoble, J. D. A. Portnoy, and M. D. Welch. 2000. Three regions within ActA promote Arp2/3 complex-mediated actin nucleation and Listeria monocytogenes motility. J. Cell. Biol. 150:527-537. Skoble, J., V. Auerbuch, E. D. Goley, M. D. Welch, and D. A. Portnoy. 2001. Pivotal role of VASP in Arp2/3 complex-mediated actin nucleation, actin branch-formation, and Listeria monocytogenes motility. J. Cell. Biol. 155:89-100. Suzuki, T., S. Shinsuke, and C. Sasakawa. 1996. Functional analysis of Shigella VirG domains essential for interaction with vinculin and actin-based motility. J. Biol. Chem. 271:21878-21885. Suzuki, T., H. Miki, T. Takenawa, and C. Sasakawa. 1998. Neural Wiskott-Aldrich syndrome protein is implicated in actin-based motility of Shigella flexneri. EMBO J. 17:2767-2776. Theriot, J. A., T. J. Mitchison, L. G. Tilney, and D. A. Portnoy. 1992. The rate of actin-based motility of intracellular Listeria monocytogenes equals the rate of actin polymerization. Nature. 357:257-260. Upadhyaya, A., and A. Van Oudenaarden. 2003. Biomimetic systems for studying actin-based motility. Curr. Biol. 13:R734-R744. Weber, A. V. T. Nachmias, C. R. Pennise, M. Pring, and D. Safer. 1992. Interaction with thymosin -4 with muscle and platelet actin: implications for actin sequestration in resting platelets. Biochemistry. 31:6179-6185.


70 Wiesner, S., E. Helfer, D. Didry, G. Ducouret, F. Lafuma, M. F. Carlier, and D. Pantaloni. 2003. A biomimetic motility assay provides insight into the mechanism of actin-based motility. J. Cell Biol. 160:387-398. Yang, C., M. Huang, J. DeBiasio, M. Pring, M. Joyce, H. Miki, T. Takenawa, and S. H. Zigmond. 2000. Profilin enhances Cdc42-induced nucleation of actin polymerization. J. Cell Biol. 150:1001-1012. Yarar, D., W. To, A. Abo, and M. D. Welch. 1999. The Wiskott-Aldrich syndrome protein directs actin-based motility by stimulating actin nucleation with the Arp2/3 complex. Curr. Biol. 9:555-558. Zalevsky, J., I. Grigirova, and R. D. Mullins. 2001. Activation of the Arp2/3 complex by the Listeria ActA protein. ActA binds two actin monomers and three subunits of the Arp2/3 complex. J. Biol. Chem. 276:3468-3475. Zeile, W. L., F. Zhang, R. B. Dickinson, and D. L. Purich. In press. Listeria’s right-handed helical rocket-tail trajectories: Mechanistic implications for force generation in actin-based motility. Cell Mot. Cytoskel.


BIOGRAPHICAL SKETCH Catherine A. Marcinkiewcz received her B.S. in biomedical engineering from Johns Hopkins University in 2001 and anticipates an M.S. in biomedical engineering from the University of Florida. She will pursue a Ph.D in neuroscience in the near future. 71