Kinectin in protein synthesis and membrane dynamics

ROLE OF KINECTIN IN MEDIATING ENDOPLASMIC RETICULUM EXTENSION DURING CELLULAR MIGRATION AND ATTACHMENT HENG KIANG JUSTIN (B.SC (Hons.), NUS) A THESIS SUBMITTED FOR THE DEGREE OF MASTER OF SCIENCE DEPARTMENT OF PHYSIOLOGY NATIONAL UNIVERSITY OF SINGAPORE 2009 ACKNOWLEDGEMENTS I wish to first thank my supervisor Prof Hanry Yu for all the advice he has provided me in both the academic and personal aspects of my life over these few years of study. I dearly wish to thank my parents, my wife and family members for their continued support for me during this phase of my life. I would like to express my sincere gratitude especially to Dr Zhang Xin and Ms Tee Yee-Han for the many years of working together as a group and the many hours of joy and heartaches spent over work. I would also like to thank Dr Chia Ser Mien, Abhi, Narmada, Siow Thing, Inn Chuan and the other lab members of the Laboratory of Cellular and Tissue Engineering for engaging in discussions and for all support rendered. Finally I wish to thank NUS for use of its facilities and the Department of Physiology for supporting me financially as well providing an engaging environment to work through my graduate studies. i SUMMARY This study aims to investigate the biological events leading to the disruption of ER distribution during knock down of kinectin, an ER-bound protein, and its effect on cellular function and physiology. Kinectin has been proposed as the membrane anchor linking the ER to the motor protein kinesin and has a significant role in mediating ER distribution into the cellular lamella. By over-expression of the binding domains of both kinectin (KNT+) and kinesin (KHC+), we demonstrate that aberrant ER distribution during kinectin knockdown is due to a disruption of the kinectin-kinesin interaction. Furthermore, disruption of the kinectin-kinesin interaction, as with kinectin knockdown, reduces cellular migration and suggests that disruption to kinectin may inhibit cell migration. Indeed, confocal imaging reveals that reduction of cellular adhesions in the cellular lamella occurs when extension of ER via the kinectin-kinesin interaction is prevented and may be the cause for reduced cellular migration. To further elucidate the physiological relevance of ER in mediating focal adhesion formation, we tested the hypothesis that ER dynamics may be important in mediating cellular adhesion to certain ECM proteins. Fibronectin, collagen and laminin coated assay plates were used to study cellular attachment dynamics. Fibronectin and collagen both induced increased cellular attachment, while laminin did not improve attachment during the time frame of the experiments. Disruption of the kinectin-kinesin interaction reduced cellular attachment on fibronectin but not collagen surfaces, suggesting that fibronectin enhances cellular attachment in an ER dependant manner. We further used localized signals to study the ability of various ECM-coated beads to induce ER accumulation for study. In agreement with results from the ii adhesion plate assays, fibronectin-coated beads appear to significantly induce ER accumulation while neither collagen- nor laminin-coated beads managed significant improvements in ER accumulation despite high concentration of coating, A model emerges whereby ER accumulation, mediated by the kinectin-kinesin interaction, of sites of new focal adhesion formation is required for proper cellular attachment and migration. Our results further suggest that interaction of cells with fibronectin but not collagen or laminin is dependent on this mechanism for proper cellular attachment. iii LIST OF FIGURES Fig. 1: A schematic representation of kinectin. Fig. 2: Kinectin knockdown results in abnormal distribution of ER. Fig. 3: Kinectin knockdown results in aberrant cell and colony shape. Fig. 4: KNT+ and KHC+ are competitive inhibitors of kinectin and kinesin. Fig. 5: GFP-KNT+ over-expression results in retraction of ER from the cellular lamella. Fig. 6: GFP-KNT+ over-expression does not disrupt microtubule cytoskeleton. Fig. 7: GFP-KHC+ over-expression inhibits ER extension into cellular lamella. Fig. 8: ER retraction in GFP-KHC+ over-expressing cells is not due to microtubule collapse. Fig 9: Paxillin and vinculin plaque formation in cellular lamella is reduced in GFP-KNT+ overexpressing cells. Fig. 10: Quantification of paxillin and vincullin in cellular lamella. Fig. 11: Wound healing is reduced during disruption to the kinectin-kinesin interaction. Fig 12: Disruption to the kinectin-kinesin interaction reduces wound healing. Fig. 13: Schematic representation of the chemotaxis insert. Fig 14: Disruption to the kinectin-kinesin interaction reduces wound healing. Fig. 15: Fibronectin significantly enhances cellular attachment. Fig. 16: Kinectin knockdown results in reduced adhesion to fibronectin. Fig. 17: The kinectin-kinesin interaction is required for fast cellular attachment to fibronectin. Fig. 18: GFP-KHC+ but not Climp63 is required during cellular attachment. Fig. 19: XPS wide scanning spectrums of original beads’ surface and fibronectin coated beads’ surface. Fig. 20: Fibronectin-coated beads induce significant ER accumulation. Fig. 21: BSA-coated beads do not induce ER accumulation. Fig. 22: Collagen-coated beads induce ER accumulation at high coating concentrations. Fig. 23: Laminin-coated beads do not induce ER accumulation. Fig. 24: Comparison of ER accumulation by various ECM proteins. iv ACKNOWLEDGEMENTS………………………………………………………i SUMMARY …………………………………………………………………….…ii LIST OF FIGURES …………………………………………………….………..iv TABLE OF CONTENTS…………………………………………………………v LIST OF ABBREVIATIONS……………………………………………………vii CHAPTER 1: Introduction 1.1 Cell Migration: The four-step process of cell migration……………………1 1.2 Cell Matrix Interactions……………………………………………………..3 1.2.1 The Significance of the ECM on cell behaviour ...………………….3 1.2.2 Integrins are major mediators of cell-matrix interactions…………...4 1.2.3 Fibronectin-integrin interaction induces focal complex assembly and ER recruitment ………………………………………………6 1.3 Overview of the Endoplasmic Reticulum (ER) …………………………….7 1.3.1 Function of the ER and establishment of its position in a cell………7 1.3.2 Regulation of ER morphology: Dynamic and static anchors ……….8 1.3.3 Role of ER at localized sites of activity……………………………10 1.3.4 1.4 Kinectin: A multifunctional protein that interacts with kinesin …….11 Role of microtubule in cell adhesions ……………………………………….13 1.4.1 Microtubules are required for maintenance of proper cell shape & function…………………………………………………………….13 1.4.2 1.5 Adhesion dynamics are mediated by microtubules ………………...14 Significance and rationale of thesis research. ………………………………16 CHAPTER 2: Role of the Kinectin-kinesin interaction in mediating ER extension 2.1 ER-bound kinectin mediates ER extension…………………………………19 2.1.1 Kinectin knockdown disrupts kinectin-kinesin function and inhibits ER extension into the cellular lamella………………………....19 2.1.2 Kinectin knockdown alters cell shape while reducing cell migration ……………………………………………………………….20 2.2 The kinectin-kinesin interaction is important for ER extension into the cellular lamella…………………………………………………………23 v 2.2.1 GFP-KNT+ over-expression inhibits ER extension into the cellular lamella ……………………………………………………………24 2.2.2 ER extension into the cellular lamella is inhibited by KHC+ over-expression despite intact microtubule network structure…………….29 2.2.3 Focal adhesion formation is reduced in the cellular lamella of GFP-KNT+ overexpressing cells……………………………………….32 2.2.4 Cell migration is reduced in KNT+ and KHC+ over-expressing cells. ………………………………………………………………………36 2.3 Conclusion: Kinectin mediates ER extension into the cellular lamella via kinesin and is required during cellular attachment and migration…….40 CHAPTER 3: Fibronectin induces ER accumulation via kinectin to facilitate initial cellular attachment 3.1 Accelerated cell-matrix adhesion to fibronectin requires ER dynamics via kinectin…………………………………………………………….…43 3.1.1 Fibronectin surface modification significantly enhances cell attachment………………………………………………………….…….43 3.1.2 Kinectin is required for initial fast cell attachment to fibronectin………………………………………………………….…….44 3.1.3 ER extension via the kinectin-kinesin interaction is important for accelerated cell attachment to fibronectin…………………….……...47 3.2 Fibronectin- and collagen- coated beads induce localized ER accumulation………………………………………………….……...52 3.3 Conclusion: ER motility via kinectin is required for cellular attachment to fibronectin but not collagen and laminin………………………..……..62 CHAPTER 4: Conclusions and future prospects……………………..….……64 CHAPTER 5: Material & Methods……………………………………..….…..67 CHAPTER 6: References……………………………………………..…….…..70 vi List of Abbreviations a.u. Absorbance units BSA Bovine serum albumin Ca2+ Calcium ions ER Endoplasmic Reticulum FAK Focal adhesion kinase h Hour IAC Integrin-based adhesion complexes KNT+ Kinesin binding domain on kinectin KHC+ Kinectin binding domain on kinesin KD KNT HeLa cell expressing pSilencer vector against vd4 KNTVC HeLa cell expressing pSilencer vector only KNTWT Wild type HeLa cell o Degrees Celsius C kDa Kilo-Dalton Kb Kilobase-pairs M Molar mM Millimolar min Minutes µg Microgram µl Microlitre mg Milligram ml Millilitre ng Nanogram NUS National University of Singapore PBS Phosphate buffered saline pDsRed2-ER Vector with kanamycin/neomycin resistance cassette encoding ER localizing fluorescent protein DsRed2 PFA Paraformaldehyde PTP1B Protein tyrosine phosphatase 1B RNA Ribonucleic acid vii Stim1 Stromal interaction molecule 1 TBE Tris-Borate-EDTA Buffer vd3 Human kinectin insert 3 vd4 Human kinectin insert 4 v/v Volume by volume w/v Weight by volume viii CHAPTER 1: Introduction 1.1 Cell Migration: The four-step process of cell migration Cellular migration may be generalized to four main elements namely 1) polarized extension of cell membrane, 2) attachment to the substrate at the leading edge, 3) forward flow of the cytosol and 4) detachment of the rear end from the substratum. When exposed to migratory stimuli such chemokines, extracellular matrix (ECM) or growth factors, cells undergo signaling events mediated by regulators such as Rho family GTPases and actin modifying proteins (Cory and Ridley, 2002; Welch and Mullins, 2002), resulting in dynamic changes in the cytoskeletal structure of the cell, resulting in ruffling of lamellapodia and extension of numerous filopodia outward from the main cell body. (Cortese et al., 1989). Attachment to the substratum is achieved via interaction between a host of receptor and ligand interactions such as hetrodimeric integrins, receptor tyrosine kinases and phosphatases, immunoglobulin receptors and cellsurface proteoglycans (Huttenlocher et al., 1995; Firtel and Chung, 2000). Cell-ECM interactions will be described in greater detail later. Cell-matrix interactions also further induce signaling events which trigger actin polymerization mediated by Arp2/3 which associate with actin filaments to form a branching network of actin in the leading lamel la (Heath and Dunn, 1978; DeMali et al., 2002). Other regulators such as WAVE (Yamazaki et al., 2006), N-WASP (Ivanov et al., 2005) and cortactin (El Sayegh et al., 2004; Helwani et al., 2004) similarly induce polymerization or induce actin capping proteins to be dissociated, allowing for extension of existing actin tubules and leading to growth and outward extension of the leading lamella. Thus, interaction of the cell with the ECM allows perpetuation of membrane 1 ruffles toward the direction of stimuli (Nelson et al., 2005). Cell extensions that do not establish adhesive contacts with the ECM retract back rapidly into the cell body (Zhelev et al., 2004), ensuring a constant recycling of components necessary for extension. Newly formed adhesions serve as traction sites for the moving cell, stabilizing membrane extensions during migration or spreading and acting as links between the matrix and the cytoskeleton (Craig and Chen, 2003; Galbraith et al., 2002). Contractile forces exerted by acto-myosin II contractions enable the cell to pull itself forward (Heidmann and Buxbaum, 1998), simultaneously reinforcing adhesions and facilitating further actin polymerization (Giannone et al., Cell 2007; Kozlov and Bershadsky JCB). This actin-protein complex can apply forces transmitted to ECM via focal adhesion (Katoh et al., 2001; Matsumura et al., 1998). Contractile forces represent a significant element during cell migration since blocking of myosin light chain kinase results in inhibition of Rho, which leads to impairment of cell contraction, causing disassembly of adhesions in cells (Burridge et al., 1997). As the cell propels itself forward, adhesions at the rear of the cell are weakened and reduced to a low affinity binding state, allowing them to be dissociated (Hughes and Pfaff, 1998). Tyrosine phosphorylation-dependent events appear to control this process (Huttenlocher et al., 1995). Detachment of receptors from the ECM is followed by endocytosis, anterograde trafficking of vesicles containing integrin and fusion of these vesicles with the cell membrane for recycling of receptors to the leading edge (Bretscher, 1996). The above four steps perpetuate via a continuous cycle in which there is no distinct start or end. Focal adhesions, highly dynamic structures important for cell 2 adhesion, feature very significantly in the process of cell migration since their turnover rates determine the state of the cell. A slow turnover of established focal adhesions is found in relatively static processes including attachment and cell spreading (Smilenov et al., 1999), whereas a dynamic turnover of focal adhesions, then termed focal contacts, is observed during cell motility and in many transformed cells (Duband et al., 1986) 1.2 Cell-Matrix Interactions 1.2.1 Significance of the ECM on cell behaviour The ECM plays a role in cell-matrix interactions and a significant role in mediating cell behaviour and attachment (Geiger et al., 2001). Matrix proteins such as fibronectin have been shown to be essential for proper cell attachment. Others such as collagen and laminin are major players in mediating proper cell attachment and function. Thus, by better understanding the extra-cellular cues required to establish and maintain cell function and the signals arising from cellular interaction with the ECM, we can allow for greater cellular control and elicit specific cellular functions (Toh et al., 2006). This would be important during development of various bio-engineered materials. Many receptors exist to mediate interaction between cell and the ECM. A major class of proteins, the integrins, mediate numerous cell-ECM interactions and act as transmembrane signal transducers, channeling extracellular signals into the cell in an “outside-in” signaling (Hynes, 1992). Integrins are composed of α and a β subunits, which form a heterodimer and bind specific ECM proteins (Zhao et al., 2004). Integrin α5β1 and integrin α5β2 are receptors for fibronectin while integrin α1β1 and integrin α2β1 are well characterized receptors of collagen in the ECM (Heino 2000; Danen et 3 al.,2002). Other ECM proteins such as laminin compliment integrin interactions with other molecules such as with the heparan sulfate proteoglycan, perlecan and fibulin-1 (Ekblom et al., 2003; Timpl et al., 2003). Rigidity and 2 or 3-dimensional topographies can influence cell behaviours (Grinnel F, 2003). Cells cultured on 3D surfaces differ significantly from 2D matrix cultured cells, adopting a bipolar or stellate morphology markedly different from the spread sheet-like morphology of 2D cells, a result of differential signally from environmental cues (Beningo et al., 2004). The ECM similarly contributes cues via its rigidity by varying its composition of ECM proteins. In support of this malignant breast tissues have been demonstrated to be significantly associated with increased ECM rigidity and result in integrin clusters and increased focal adhesion assembly (Paszek et al., 2005). Similarly, it has been suggested that alternation of the ECM microenvironment may induce cells to accumulate metalloproteinases (MMP), a key invadopodial protease responsible for ECM degradation by carcinoma cells (Ayala et al., 2006.). Thus, ECM proteins can provide environmental cues to differentially induce attachment, migration, spreading and other behaviours in cells. 1.2.2 Integrins are major mediators of cell-matrix interactions Cell spreading (Ylanne et al., 1993) and migration (Kondo et al., 2000) assays have been used to evaluate the assembly and recruitment of focal complexes. There are different mechanisms that have been shown to regulate focal complexes and focal adhesions, such as mechanical forces from ECM (Galbraith et al., 2002), actin-myosin complex (Chrzanowska-Wodnicka and Burridge, 1996) and Rho-regulated microtubule 4 dynamics (Small and Kaverina, 2003). Cell-matrix interactions typically begin with ECM ligands in an extra-cellular environment binding to their receptors on cell membrane. Integrins form a large family of surface receptors that are localized on plasma membrane and share common features in their molecular structure and function (Berman et al., 2003). Integrin receptors mediate the formation of focal complexes and play important roles in signal transduction from the extra-cellular matrix (ECM) (Wayner et al., 1988). Ligand binding induces conformational changes in integrin which is required for its clustering and subsequent function in recruiting focal adhesion proteins (Miyamoto et al., 1995; Hato et al., 1998). Focal adhesions contain proteins that link integrin with cytoskeleton and recruit adhesion related proteins such as vinculin, paxillin and focal adhesion kinase (Zaidel-Bar et al., 2003). The integrin β subunit binds to talin and α-actinin which subsequently interacts with vinculin. Extra-cellular stimuli can thus induce cytoskeleton rearrangement and intra-cellular signals via integrin receptors (Berrier and Yamada, 2007; Lock et al., 2008). Various signalling regulators such as R-Ras can mediate integrin signals. R-Ras regulates integrin affinity and avidity, enhancing focal adhesion formation and promoting integrin clustering to collagen for greater cell adhesion (Zhang et al., 1996; Kwong et al., 2003). Integrins are thus important proteins responsible both for being an integral part of focal adhesions as well as being one of the earliest signaling events responsible for mediating the response between the cell and ECM. 5 1.2.3 Fibronectin-integrin interaction induces focal complex assembly and ER recruitment Fibronectin, a naturally occurring integrin ligand in the ECM, has been used widely in engineering surfaces to mimic the extra-cellular environment (Plopper and Ingber, 1993). Fibronectin forms a homodimer cross-linked by two disulfide bonds with an RGD domain essential for its recognition by integrin (Engel et al., 1981). Interaction between fibronectin and integrin α5β1 or α5β3 immobilizes the complex and a focal point for further recruitment of proteins required for strong adhesion of the cell to ECM. This interaction also establishes the linkage between extra-cellular fibronectins and intracellular actin filaments which is required for generating mechanical stress which regulates cell functions (Schatzmann et al., 2003; Schwartz, 2001). Focal adhesions develop from such sites, initially recruiting paxillin and α-actinin (Laukaitis et al., 2001). Talin, paxillin, vincullin and focal adhesion kinase (FAK) are subsequently recruited to reinforce the adhesion (Edlund et al., 2001). Subsequently late stage proteins zyxin and tensin are recruited to further stabilize the adhesion (Zaidel-Bar et al., 2003). Adhesion dynamics are regulated via both signaling and localization events. Phosphorylation events such as those by FAK, which is itself phosphorylated and activated by R-Ras, mediate the activity of proteins present in focal adhesions (Kwong et al,). Ras may also modulate adhesion dynamics via localization and has been observed to be enriched in the proximity of focal adhesions, thereby providing an alternative to phosphorylation for focal adhesion regulation (Furuhjelm and Peranen, 2003). Besides FAK, vinculin has been observed to localize to the lamellipodium of cells in a temporally 6 regulated fashion (DeMali et al., 2002), supporting localization of focal adhesion components as a method of regulation of adhesion dynamics after integrin recruitment. Beside proteins, mRNA and ribosomes have been demonstrated to be localized to focal complexes within 20min of interaction with fibronectin-coated beads (Chicurel et al., 1998). Furthermore, intra-cellular organelles have also been demonstrated to localize to integrin-based adhesion complexes (IACs). Fibronectin coated beads were used to isolate proteins in the IAC and a marked enrichment of ER-bound kinectin, RAP (lowdensity lipoprotein receptor-related protein (LRP) receptor-associated protein) and calreticulin were observed in the resulting pull-down (Tran et al., 2002), suggesting enrichment of the ER to IACs. Localization of protein translational machinery thus appears to play an integral part in adhesion dynamics. Using fibronectin modified beads as local signals to induce integrin clustering and focal complex assembly, mRNAs and ribosomes have been found recruited and accumulated to the focal complexes within 20 min upon the interaction of cells with beads (Chicurel et al., 1998). This redistribution of mRNA and ribosomes to the focal complexes may explain the rapid increase in protein synthesis that is observed before the detected changes in global transcription in response to substrate adhesion (Chicurel et al., 1998). 1.3 Overview of the Endoplasmic Reticulum (ER) 1.3.1 Function of the ER and establishment of its position in a cell The ER is a large and dynamic membranous organelle with numerous functions. Its network extends throughout the cell, originating from the nuclear envelope and 7 extending throughout the cytosol. The smooth ER is involved in the synthesis of lipids and membrane proteins, rough ER is important in the synthesis of other proteins while the transitional ER is where carrier vesicles are formed (Baumann and Walz, 2001; Krijnse-Locker et al., 1995). It is not a static but instead a highly dynamic organelle which undergoes constant rearrangement to facilitate its roles encompassing calcium regulation, lipid synthesis, protein synthesis and translocation of proteins into its lumen for modifications (Kaufman, 1999; Parikh et al., 2005). To mediate such processes throughout the cell, proper distribution and morphology of the ER is essential. ER morphology has been shown to be highly microtubule-dependant (Dailey and Bridgman, 1989; Vedrenne and Hauri, 2006; Waterman-Storer and Salmon, 1998). The ER has been observed, via electron microscopy, to be highly colocalized with the MT cytoskeleton, suggesting a role of the MT in mediating its distribution and proper morphology (Masurovsky et al., 1981; Tokunaka et al., 1983). ER tubules have been observed to both extend toward the lamella of migrating cells and growth cones of neurons to establish a network (Dailey and Bridgman, 1989; Waterman-Storer and Salmon, 1998), and to retract toward the cell nucleus (Waterman-Storer and Salmon, 1998), suggesting that the ER network is a dynamic organelle which may modify its distribution according to the needs of the cell. 1.3.2 Regulation of ER morphology: dynamic and static anchors Regulation of the ERs morphology occurs via a host of different interactions mediated by ER-bound proteins and can be generalized to static and dynamic anchors. Static anchors such as Climp63 and huntingtin contribute significantly to its morphology. 8 Expression of mutant CLIMP-63, which is unable to bind to microtubules, and siRNAmediated silencing of huntingtin, an inositol-trisphosphate receptor, results in abnormal distributions of ER despite normal microtubule networks (Tang et al., 2003; Omi et al., 2005). Other proteins such as p22 mediate fusion events in the ER and have been demonstrated to induce distinct vesiculation of the ER upon microjection of anti-p22 antibodies which is not apparent during conventional microtubule disruptions (Andrade et al., 2004). Dynamic anchorage of the ER is achieved via its interaction with the ends of polymerizing microtubule tips by Stromal interaction molecule 1 (Stim1) or by kinesin by kinectin for extension. Stim1 has been demonstrated to interact with end binding protein 1 (EB1) which is present on the plus end of microtubules (Grigoriev et al., 2008). Kinectin forms another key interaction which modulates ER extension along microtubules by interacting with kinesin (Toyoshima et al., 1992). The inhibition of kinesin with antisense oligonucleotides can cause retraction of ER from cell periphery (Feiguin et al., 1994), indicating that the kinesin-related mechanism is essential for ER extension along microtubule (Rodionov et al., 1993). Dyenin has also been demonstrated for ER retraction further suggesting a model of constant ER remodeling to suit the needs of the cell (Wozniak et al., 2009). Such models of ER extension highlight a particular dependence on microtubules to provide structural support for ER extension, or in the case of Stim1, the actual extension mechanism. Disruption of microtubules has been observed to affect both new tubule growth and causes retraction of ER tubules from the cell periphery (Terasaki et al., 1986), reinforcing the importance of the cytoskeleton for permitting ER dynamics in cells. 9 This observation is similarly repeated in plant and yeast cells which use the actin cytoskeleton as the structural basis for ER dynamics (Liebe and Menzel, 1995; Prinz et al., 2000). 1.3.3 Role of ER at localized sites of activity Due to its roles in protein synthesis and Ca2+ storage as well as presence of various proteins on its surface, ER can mediate adhesion dynamics in cells. The ER is a major protein synthesis mechanism and has been demonstrated to be capable of local synthesis of proteins in neuronal cells (Vale, 2003). Indeed, local translation appears to be the trend with a growing body of evidence of local synthesis of proteins such as RhoA and β-actin being demonstrated, further enforcing this model (Wu et al., 2005). Demonstrations of a direct link between intracellular mRNA localization and local protein synthesis of proteins such as integrin α3 and β-actin in non dendritic cells have also been shown via pulse-chase experiments (Adereth et al., 2005; Rodriguez et al., 2006). The mRNA of integrin α3 has an ER localization signal, highlighting the possibility that ER redistribution may be a possible prerequisite for such local synthesis to occur. The ER is also the predominant Ca2+ reservoir in the cell and actively releases or sequesters Ca2+ (Clapham, 2007). Its high affinity for Ca2+ in its lumen enables it to efficiently transport Ca2+ throughout the cell (Peterson and Verkhratsky, 2007). It can also mediate cytosolic Ca2+ levels via Stim1, an ER-bound protein which oligomerises upon ER calcium store depletion and translocates to the plasma membrane where it induces store operated calcium influx via interaction with Orai1 (Peinelt et al., 2006; 10 Zhang et al., 2005). Such localization of the ER may mediate local Ca2+ at specific sites such as focal adhesions, thereby influencing their turnover. Further more ER bound protein tyrosine phosphatase 1B (PTP1B) has been shown to be essential in mediating protein recruitment to focal complexes (Hernandez et al., 2006) while the enrichment of kinectin (Tran et al, 2002), a kinesin binding protein, suggest that the ER can localize and may actively modulate adhesion dynamics. Kinectin in particular is of significance as it utilizes the ATP-dependent kinesin motor protein, making its activity controllable. 1.3.4 Kinectin: A multifunctional protein that interacts with kinesin Kinectin, an integral membrane protein found in ER (Kumar et al., 1995; Toyoshima et al., 1992; Yu et al., 1995), serves as a receptor for the microtubule motor protein kinesin (Ong et al., 2000; Santama et al., 2004) Kinectin was initially discovered bound to conventional kinesin when microsomes from chick embryo brain cells were passed through a kinesin affinity column (Toyoshima et al., 1992). It has been discovered in a host of cells but is notably absent from Caenorhabditis or Drosophila genomes despite the presence of the conserved conventional kinesin heavy chain gene (Yang et al., 1989; Goldstein and Gunawardena, 2000). This suggests that kinectin may not be the main dynamic anchor present in certain cells or that alternative mechanisms may be compensating for kinectin’s absence. Two main kinectin isoforms of 160kDa and 120kDa exist (Leung et al., 1996). Kinectin exists as both a 160 kDa ER bound protein and a shorter 120 kDa kinectin, which lacks the N-terminus transmembrane domain (Futterer et al., 1995), localized in the mitochondria which can bind 160 kDa kinectin as a heterodimer via coiled-coil 11 domain interactions or via myristolation to the membrane (Toyoshima et al., 1992; Yu et al., 1995; Santama et al., 2004; Kumar et al., 1998). The C-terminus of kinectin contains variable domains that contribute to different novel isoforms via alternative splicing. The variable domains 3 (vd3, residues 1177-1200) and 4 (vd4, residues 12291256) have been mapped as the minimal kinesin-kinectin interaction domain on kinectin (Kumar et al., 1995; Ong et al., 2000). Knockdown of kinectin disrupts the kinectinkinesin interaction and leads to collapse of ER network from the cell periphery (Santama et al., 2004). Over-expression of vd4 domains (KNT+), or the kinectin binding domain (KHC+), resulted specifically in competition with endogenous kinectin or kinesin for binding with its partner. This disrupts the kinectin-kinesin interaction and similarly results in ER adopting a retracted morphology. A recent finding that over-expression of a kinesin-1 dominant negative mutant inhibited ER extension lends support to this hypothesis (Wozniak et al., 2009). The overexpression of kinesin binding fragment of kinectin resulted in perinuclear clustering of mitochondria, which is strikingly reminiscent of the phenotype in conventional kinesin heavy chain deficient cells (Tanaka et al., 1998). Further characterization of kinectin revealed that besides binding to kinesin, kinectin also interacted with Rho GTPase and was found to anchor the translation elongation factor-1 complex (Hotta et al., 1996; Alberts et al., 1998; Ong et al., 2003). This is illustrated schematically in Fig. 1. 12 Kinesin Elongation Factor-1 δ Rho GTPase Kinectin Fig. 1: A schematic representation of kinectin. The full length kinectin comprises an N-terminal transmembrane domain spanning the ER membrane and a large Cterminal coiled coil forming cytoplasmic part. Kinectin interacts with kinesin, EF-1δ and Rho GTPase via its cytoplasmic C-terminal (arrows). 1.4 Role of microtubule in cell adhesions 1.4.1 Microtubules are required for maintenance of proper cell shape & function Microtubules are hollow filamentous structures which form tubules of diameter 20-23 nm.They are composed of α and β tubulin subunits, 55 and 50 kDa respectively, which hetrodimerize to form a polarized structure with a plus and minus end (Nogales et al., 1999). Polymerization and deplolymerization of microtubules is dependent on the βtubulins which bind either GTP or GDP to facilitate polymerization and depolmerization respectively. Internal as well as external stimuli can also modulate the dynamics of cytoskeletal rearrangements by altering the binding state of β tubulin or via transmembrane receptors such as integrin (Berrier and Yamada, 2007; Lock et al., 2008). 13 The polarized arrays of microtubules provide tracks for the intra-cellular transport of membrane-organelles, vesicles and proteins. Microtubules radiate from an organizing center (MTOC) adjacent to the nucleus, and extend throughout the cytoplasm towards the cell periphery (Lane and Allan, 1998). Various motor proteins can mediate transport of cargo on microtubules specifically toward the plus or minus end, allowing for specific transport of material within the cell. The cytoskeleton network provides support and acts as a scaffold for which the motor proteins kinesin, myosin and dyenin function (Janmey, 1998). This bi-directional movement system along polarized microtubules plays important roles in the transportation of vesicles in neural axons and the establishment of polarity in epithelial cells (Hirokawa, 1998).The rearrangement of this network changes the cellular mechanical properties and the cell shape in cell adhesion, spreading and migration (Gumbiner, 1996; Ziegler et al., 2008). 1.4.2 Adhesion dynamics are mediated by microtubules The cytoskeleton can modulate adhesion dynamics. As previously described, actin filaments can serve to promote adhesion recruitment and formation by acting as the structural basis from which mechanosensing occurs. Myosin contraction propagates mechanical stresses which enhance adhesion formation. Unlike actin filaments however, microtubules function to modulate adhesion dynamics by inducing their dissociation and leads to adhesion dissociation upon contact (Small et al., 2002a). Experiments with microtubule inhibitors showed that adhesions initially stabilized nascent microtubules (Kaverina et al., 1998), however this is rapidly reversed and adhesions quickly dissociate 14 (Kaverina et al., 1999). It has been recently demonstrated that focal adhesions themselves may modulate microtubule dissociation via a biochemical mechanism involving paxillin and thus participate in a feedback loop which may regulate adhesion formation (Efimov & Kaverina, 2009). Disruption of microtubules by nocodazole treatment leads to enlargement of adhesion plaques in cells, demonstrating the significance of microtubules in regulating adhesion size (Kaverina et al., 1999). Many possible mechanisms such as dynamin-FAK dependent disassembly or Rho- mouse diaphanous (mDia) recruitment of c-Src have been used to explain disassembly of adhesions. (Yamana N, et al., 2006; Ezratty EJ, et al. 2005). Disruption of kinesin activity also prevents adhesion disassembly dispite normal extension of microtubules to adhesions, suggesting that microtubules might regulate adhesion disassembly via a kinesin-mediated cargo transport to adhesions (Krylyshkina et al., 2002). Together, all these observations converge in providing an explanation for the behaviour of adhesions when in contact with microtubules. An alternative role of microtubules as possible mediators of adhesion growth can be derived from ER-dependent means of mediating adhesion growth. ER-bound PTP1B, which is essential for adhesion growth, has been demonstrated to be localized to focal adhesions (Hernandez et al., 2006), suggesting that ER is transported to the sites of newly forming adhesions. Further evidence of a possible microtubule contribution can be drawn from the enrichment of kinectin to IACs fromed with fibronectin coated beads (Tran et al., 2002). Together, these data suggest that the ER is localized to adhesion sites and is important for mediating ER adhesion. Extension of the ER to adhesions likely occurs in a microtubule dependant manner with dynamic anchors of the ER, in particular those of 15 plus end directed transport, playing a significant role in mediating ER extension to adhesions. Of these dynamic anchors, the kinectin-kinesin interaction appears most prominent. Kinectin interacts with kinesin and the kinesin-binding domain enhances the microtubule-stimulated kinesin-ATPase activity (Ong et al., 2000), enhancing transport by kinesin to the adhesions. The dual function of microtubules in inducing assembly as well as disassembly suggests that adhesion regulation by microtubules is complicated and may be temporally or spatially regulated though the exact dynamics is yet unknown. The cargo transported to adhesions has also remained elusive though inhibition of kinesin results in enlarged adhesions, suggesting that kinesin appears to be a key carrier (Krylyshkina et al., 2002). 1.5 Significance and rationale of thesis research Cell adhesion to a substratum is dependent on the formation of focal adhesions and is essential for proper cellular spreading and migration. Adhesion plaque formation can be mediated by microtubule based activities as demonstrated by previous studies using nocadazole. However such treatment results in catastrophic disassembly of the microtubule network, limiting further understanding of the players involved. We thus sought to investigate the role of a single microtubule associated component, the ER, in mediating adhesion dynamics via microtubules and its effect on cellular adhesion to various substratum. In this study, we hypothesize that ER extension to focal adhesions is required for proper adhesion recruitment and cell function. We studied the role of the ER in mediating ER extension, cellular adhesion recruitment and migration by disrupting the kinectin- 16 kinesin interaction by over-expressing the kinectin (KNT+) or kinesin (KHC+) binding domains present on the respective proteins. This would allow for specific disruption of kinectin’s function via direct competitive inhibition. The requirement of the ER to effect attachment in cells induced by fibronectin, collagen and laminin during cell attachment was also investigated. This work will extend our understanding of ER-related events occurring in cells during attachment to ECM matrix proteins, paving the way for development of systems and methods leading to greater control over cell behaviour in tissue engineering. 17 CHAPTER 2: Role of the Kinectin-kinesin interaction in mediating ER extension There is evidence to suggest that the ER does play a role in maintenance or recruitment of proteins to focal adhesions. Numerous studies have demonstrated that ER bound proteins such as ER-bound PTP1B, calnexin and calreticulin can perform specific roles in mediating adhesion protein recruitment or inducing Ca2+ signaling (Hernandez et al., Wang et al., 2006), highlighting an important role played by the ER in adhesion regulation. It is unlikely that ER activity would occur simultaneously throughout the cells since this would not be efficient in ensuring specificity in activity. In dendrites, which often span thousands of micrometers in length, it has been established that the ER functions locally with proteins being locally synthesized at the distal neuronal end, as opposed to synthesis of proteins at the cell body and trafficking to the distal end (HKJ Vale, 2003). The high affinity of the ER lumen for Ca2+ also allows for a Ca2+ to be transported efficiently through its network compared to the cytosol as if it were in a tunnel (Petersen and Verkhratsky, 2007). This would allow for localized release at adhesion sites and thus allow for local modulation of adhesion turnover rates in cells (Conklin et al., 2005). Together, these studies have demonstrated a requirement of the ER to be localized to effect local signals in neuronal cells. Local ER activity appears also in non-dendritic systems, and is an increasingly important trend, with a growing body of evidence demonstrating local synthesis of proteins such as RhoA and β-actin being demonstrated (Wu et al., 2005). Signal transduction across the entire length of the cell would result in both non-specific signals and inefficient transmission of the signal. Demonstrations of a direct link between 18 intracellular mRNA localization and local protein synthesis of proteins such as integrin α3 and β-actin in non dendritic cells have also been shown via pulse-chase experiments (Adereth et al., 2005; Rodriguez et al., 2006). Furthermore, localization and enrichment of kinectin and other ER-bound proteins to fibronectin coated beads, and failure of PTP1B to localize to focal adhesions after treatment with nocodazole in mouse fibroblast cells result in aberrant focal adhesion formation (Hernandez et al., 2005). Together these findings support the hypothesis that ER localization is required for proper focal adhesion formation. We thus sought to investigate the role of ER localization in mediating cell function via kinectin-kinesin interaction. 2.1 ER-bound kinectin mediates ER extension 2.1.1 Kinectin knockdown disrupts kinectin-kinesin function and inhibits ER extension into the cellular lamella Kinectin has been demonstrated to be a key element in the maintenance of proper ER morphology in cells. Knockdown of kinectin has been previously shown to cause ER to be redistributed into a perinuclear morphology, highlighting its important role in specifically mediating ER extension toward the cell's leading edge (Santama et al., 2004). As observed in Fig. 2, distribution of ER in wild type (KNTWT) or vector control (KNTVC) cells is normal. ER appears as an evenly distributed network around the cell with even distribution of ER throughout the cell. In contrast, cells with knockdown of kinectin (KNTKD) however result in redistribution of ER away from periphery of the cell, resulting in the appearance of blue regions with low ER density at the cell edge. Kinectin thus is an important player in ensuring proper ER morphology. However it remains 19 unclear which of kinectin’s many interacting partners is responsible for the observed changes. B A B C Fig. 2: Kinectin knockdown results in abnormal distribution of ER. Distribution of ER is expressed in a pseudocolour intensity plot to demonstrate distribution of (A) wild type Hela cells (KNTWT), (B) vector control pSilencer transfected HeLa cells (KNTVC) and (C) kinectin knockdown HeLa cells (KNTKD).Red regions indicate regions of high ER density while blue regions indicate regions with low density of ER. KNTKD cells have distinct regions of retraction of eR from the cell periphery, as seen from the presence of blue regions at the cell edge. Scale bar: 10µm 2.1.2 Kinectin knockdown alters cell shape while reducing cell migration. To further understand the physiological impact of kinectin knockdown, KNTWT, KNTVC and KNTKD cells were seeded in 35mm culture dishes and imaged under a light microscope after 24h (Fig. 3a-c). Individual cells were observed under a 20X objective for morphological differences. Cell shape was similar in KNTWT and KNTVC cells with cells exhibiting a spread shape with visible extensions protruding outward from the cell center (Fig. 3a,b; White arrows). Cells also adopted a more spread and elongated morphology. KNTKD cells in contrast appeared to have a more rounded shape with significantly less cells having outward protrusions (Fig. 3c) Cells also displayed significantly less spread morphology, appearing different from the elongated KNTWT and KNTVC cells. 20 The cells were then allowed to continue in culture another 48h and thenobserved under a light microscope (Fig. 3d-f). Under a 10X objective, KNTWT and KNTVC cells had significant number of the cells displaying spread morphology, appearing to stretch out from the center of the colony body, resulting in the appearance of an irregular colony border (Fig. 3d,e). Individual cell shape from these colonies were observed to be similar to those cultured over 24h, appearing spread and with an elongated shape (Fig. 3d,e; white arrows). KNTKD cells in contrast did not spread away from the colony body (Fig. 3c). Morphologically, KNTKD colonies had distinctly rounded borders, likely due to the lack of cell protrusions from individual cells (Fig. 3f). Together, these results suggest that kinectin plays a significant role in mediating proper cell shape, likely by influencing the ability of cells to form outward protrusions. 21 24h 72h a d b e c f KNTWT KNTVC KNTKD Fig. 3: Kinectin knockdown results in aberrant cell and colony shape. KNTWT, KNTVC and KNTKD cells cultured sparsely over 24h and imaged under a light microscope with a 20X objective. (a-c) KNTWT and KNTVC cells exhibited spread morphology with distinct cellular protrusions while KNTKD cells exhibited significantly more rounded morphologies. (d-f) After 72h culture, colony morphologies of both KNTWT and KNTVC cells had cells extending outward from the colony center, resulting in an irregular colony border. KNTKD cells in contrast exhibited no cellular extensions from the colony body resulting in round, smooth bordered colonies. White arrows indicate cells with distinct cell protrusions. Scale bar (a-c): 30µm, (d-f): 60µm. 22 2.2 The kinectin-kinesin interaction is important for ER extension into the cellular lamella. Kinectin may interact with Rho GTPases, eukaryotic elongation factor 1δ (EEF1δ) and kinesin (Ong et al., 2003, Hotta et al., 1996, Vignal et al., 2001, providing an inconclusive picture as to the actual interactions involved in mediating ER retraction and abnormal cell and colony morphology. It has been demonstrated that over-expressing the kinectin fragment containing the EEF-1δ binding site disrupted the intracellular localization of EEF-1δ protein. This potentially highlights a possible role of kinectin in anchoring the elongation factor 1 complex, consisting of the 1α β γ δ subunits, to the ER and thus may be involved in regulating synthesis of proteins. Residues 630-935 of human kinectin can interact with GTP-bound forms but not GDP-bound forms of RhoA, Rac1 and Cdc42, suggesting that kinectin could be mediating Rho signaling events (Hotta et al., 1996; Alberts et al., 1998). These interactions could possibly mediate the reduced cell protrusions in KNTKD cells. They however do not provide clues as to how retraction of the ER observed in Fig. 2 occurs. Loss of interaction between kinectin with kinesin provides a more direct link to retraction of ER from the cell periphery. Kinectin constitutes the majority of kinesin binding sites on membrane organelles and is required for kinesin-driven motility (Toshiyama et al., 1992; Blocker et al., 1997; Ong et al., 2000). Knockdown of kinectin would ostensibly result in reduced interaction of kinectin with kinesin, resulting in the ER not being able to extend toward the cell periphery. This in turn could lead to events at the leading edge resulting ultimately in the reduction of cell protrusions. Thus, we overexpressed either the GFP tagged variable domain 4 of kinectin, which has been identified 23 to be the minimal kinesin binding domain (Ong et al., 2000; KNT+), or the GFP tagged minimal binding domain for kinectin on kinesin KHC (Santama et al., 2004; KNT+), to disrupt the kinectin-kinesin interaction and study its effect on cell behaviour (Fig. 4). Over-expression was verified by the expression of the GFP reporter which allowed for identification of successfully transfected cells. Since GFP signal was significantly higher than that of neighbouring cells, we deemed that these cells were over-expressing KHC or KNT+. Fig. 4: KNT+ and KHC+ are competitive inhibitors of kinectin and kinesin. Minimal binding domains of kinectin (KNT+) and kinesin (KHC+) to each other have been previously identified (Ong et al., 2000;Santama et al., 2004). Over-expression of either KNT+ or KHC+ results in competitive inhibition of the kinectin-kinesin interaction by blocking of the binding domain on kinectin or kinesin respectively. 2.2.1 GFP-KNT+ over-expression inhibits ER extension into the cellular lamella Kinectin’s interaction with kinesin via its variable domain 4 was previously confirmed using yeast two hybrid experiments (Ong et al., 2000). Over-expression of the GFP-tagged variable domain (GFP-KNT+) has been demonstrated to result in ER collapse of the ER in cells (Santama et al., 2004). While previous experiments focused on the total collapse of cellular ER, this current study instead focuses on the ER related events occurring at the cellular lamella to provide a more directed enquiry into the 24 interactions leading up to the disruption of protrusion formation in KNTKD cells. The cellular lamella is defined as the region at the cell leading edge that extends 1-3 µm back and are enriched with cortactin (Chhabra and Higgs, 2007; Kaksonen et al., 2000). To examine the ER motility within the cell, we generated a stable-cell line expressing the pDsRed2-ER vector. Briefly, HeLa cells were transfected with pDsRed2ER, a fluorescent tag coupled with the KDEL sequence, which targets and labels the ER specifically. Cells were subsequently selected using media containing the G418 antibiotic and screened via FACs resulting in a uniformed culture of cells with fluorescently labeled ER. Over-expression of GFP-KNT+ in DsRed2-ER stably expressing cells resulted in retraction of ER from the cellular lamella whereas over-expression of the GFP control vector (GFP+) resulted in no observable collapse of ER from the cellular lamella edge (Fig. 5). Cells were immuno-labelled using anti-cortactin antibodies to identify the leading lamellar region, which would have distinct accumulation of cortactin, an indicator of the leading edge of cells during formation of lamellipodia and other cell extensions (Kaksonen et al., 2000). Since enrichment of cortactin correlated highly with ruffling of the cell edge when viewed in the GFP channel, we thus used this method for identifying the leading lamella subsequent experiments. 25 GFP+ GFP-KNT+ DsRed2ER Cortactin Overlay 73.1±21.7% 21.1±6.5% Fig. 5: GFP-KNT+ over-expression results in retraction of ER from the cellular lamella. DsRed2-ER cells were transfected over-expressing either GFP+ or GFP-KNT+ (inset), immunolabelled with anti-cortactin antibodies (green) and imaged under a confocal microscope with a 60X objective. Cortactin is represented in green to enhance visual contrast. Accumulation of cortactin confirms the presence of the cellular lamella at the cell boundary. ER retraction from the cellular lamella was observed in GFP-KNT+ over-expressing cells. Values represent percentage of cells with ER extended to within 5µm of the cortactin border from 3 independent experiments. Scale bars: 5 µm. 26 Quantification of the number of GFP-KNT+ over-expressing cells with ER extended to within 5µm of the cortactin border reveal that only 21.1±6.5% of cells had ER extended to the cellular lamella. In comparison, 73.1±21.7% of GFP+ over-expressing cells had ER extended to within 5µm of the cortactin border (Fig. 4, numbers inset). Disruption of the kinectin-kinesin interaction thus appears to impede ER entry into the cellular lamella. ER tubular structure was also disrupted in GFP-KNT+ cells, appearing incapable of forming distinct terminal ER tubules. GFP+ over-expressing cells in contrast had an evenly distributed ER structure which had distinguishable ER tubule morphology at the terminal ends. Since ER morphology is highly dependant on the microtubule cytoskeleton, we sought to investigate the effect of GFP-KNT+ over-expression on the integrity of the microtubule cytoskeleton. DsRed2-ER labeled cells were transfected with either GFPKNT+ or GFP+, immuno-labelled with anti-tubulin antibodies and imaged under a confocal microscope (Fig. 6). Over-expression of GFP-KNT+ and GFP+ both did not result in any disruption of microtubule integrity despite retraction of ER from the cellular lamellar of GFP-KNT+ over-expressing cells. Over-expression of GFP-KNT+ fragment thus induces retraction of the ER from the cellular lamella despite an intact microtubule cytoskeleton network. 27 GFP+ GFP-KNT+ DsRed2ER α-Tubulin Overlay Fig. 6: GFP-KNT+ over-expression does not disrupt microtubule cytoskeleton. DsRed2-ER cells were transfected with GFP+ or GFP-KNT+ (inset) and immuno-labelled with anti-tubulin antibodies (Green). Cells were imaged using confocal microscopy with a 60X objective. Tubulin is represented in green to enhance visual contrast. Microtubule networks were observed to extend extensively throughout cells in both treatments. ER tubules however were retracted in GFP-KNT+ cells but not in GFP+ cells. Scale bars: 5 µm. Kinectin’s role in mediating ER extension thus appears to rely significantly on its interaction with kinesin since expression of exogenous GFP-KNT+ results in ER 28 retraction. Expression of GFP-KNT+ results in competition between endogenous kinectin for the kinectin-binding domain on kinesin, resulting in decreased dynamic linkage of the ER from the microtubule cytoskeleton (Ong et al., 2000). Expression of GFP-KNT+, which specifically binds kinesin, would also allow for specific disruption of ER distribution via the kinectin-kinesin interaction while minimizing disturbances to the other interactions of kinectin. 2.2.2 ER extension into the cellular lamella is inhibited by KHC+ over-expression despite intact microtubule network structure To corroborate our observations of GFP-KNT+ disruption of ER extension, we subsequently over-expressed a GFP-tagged kinectin binding fragment of kinesin (GFPKHC+) in DsRed2-ER cells to study its effect on ER morphology. Over-expression of GFP-KHC+ effectively reduces available endogenous kinectin interaction sites by competing with kinesin for binding, thus reducing the number of effective linkages between full length kinectin and kinesin. Using anti-cortactin antibodies to immuno-label cells for identification of the cellular lamella, over-expression of GFP-KHC+ in DsRed2-ER cells resulted in a great reduction in the number of cells having ER extending to within 5µm of the lamella boundary (Fig. 7). ER morphology was also significantly altered with terminal tubules becoming merged and convoluted, losing the distinct tubular morphology. This was similar to the ER morphology observed during GFP-KNT+ over-expression (Fig. 5) suggesting that both GFP-KNT+ and GFP-KHC+ disrupt a similar machinery. 29 GFP-KHC+ Cortactin DsRed2-ER Overlay Fig. 7: GFP-KHC+ over-expression inhibits ER extension into cellular lamella. DsRed2-ER cells over-expressing GFP-KHC+ were immuno-labelled with anti-cortactin antibodies and imaged via confocal imaging using a 60x objective. Cortactin is represented in green to enhance visual contrast. ER extension to the distinct cortactin boundary representing the cellular lamella is inhibited by disruption of the kinectin-kinesin interaction. Scale bar: 5 µm. To ensure that ER retraction after disruption of GFP-KHC+ was a result of disruption of the kinectin-kinesin interaction rather than non-specific disruption to ER dynamics, anti-tubulin antibodies were used to immuno-label GFP-KHC+ overexpressing DsRed2-ER cells to similaryly confirm that disruption did not occur for 30 microtubules (Fig. 8). Microtubule cytoskeleton network remained intact and unperturbed by over-expression of GFP-KHC+, maintaining a well distributed network throughout the cell despite retraction of ER from the cellular lamella. Over-expression of GFP+ similarly did not adversely affect microtubule structure. GFP-KHC+ α-Tubulin DsRed2-ER Overlay Fig. 8: ER retraction in GFP-KHC+ over-expressing cells is not due to microtubule collapse. DsRed2-ER cells were transfected with GFP-KHC+. GFP-KHC+ over-expression resulted in retraction of ER from the cellular lamella. Tubulin is represented in green to enhance visual contrast. Microtubule network distribution was normal and extended throughout the cell and up to the cell boundary. Scale bar: 5 µm. 31 Thus, over-expression of GFP-KHC+ effectively disrupts the kinectin-kinesin interaction resulting in disruption of ER extension to the cellular lamella, supporting earlier results from GFP-KNT+ over-expression. Kinectin thus appears to function to mediate the interaction with kinesin and facilitate ER extension into the cellular lamella. Knockdown of kinectin would thus result in disruption to this extension into the cellular lamella, resulting in abnormality in the cellular lamella. We therefore focused further experiments on the events occurring at the cellular lamella to provide further insight. 2.2.3 Focal adhesion formation is reduced in the cellular lamella of GFP-KNT+ over-expressing cells. Since ER extension is reduced in both kinectin knockdown cells and cells with disrupted kinectin-kinesin interactions (GFP-KNT+ or GFP-KHC+ over-expressing cells), ER activity at the cellular lamella would likely be reduced. Regulation of adhesion dynamics by PTP1B, Ca2+ or via direct synthesis of essential adhesion components may be affected at the cellular lamella during absence of the ER. This would result in disruption to formation of cell-substrate adhesions and could lead to the observed lack of distinct protrusions in KNTKD cells. We thus sought to investigate the effect of disrupting ER extension on adhesion formation in the cellular lamella. Adhesion plaques were identified by the accumulation of adhesion proteins such as vinculin and paxillin into distinct plaques (Craig and Chen, 2003; Zaidel-Bar et al, 2000). DsRed2-ER cells were transfected with either GFP+ or GFP-KNT+ and allowed to spread on a plate overnight. Cells were then fixed and immuno-labelled with antivinculin or anti-paxillin to identify adhesion plaques (Fig. 9, white arrows). Cells overexpressing GFP-KNT+ and GFP-KHC+ had notable reductions in number of both vinculin 32 and paxillin adhesion plaques along its cellular lamella. Comparatively, GFP+ overexpressing cells displayed evenly distributed adhesions with no distinct retraction from the cellular lamella. Quantification of both vinculin and paxillin plaques in GFP-KNT+ cells within 5µm of the cellular lamella boundary was done (Fig. 10). Briefly, images of the adhesion plaque channels (vinculin or paxillin repectively) were processed via ImageJ. A background filter was used to subtract any background signals leaving only the distinct plaques. A binary mask was then generated before plaques were theresholded for size and the plaques counted. These plaques were then normalized to the length of the visible membrane ruffles. Plaque in GFP-KNT+ cells were significantly reduced (peaks at 0.1-0.2 plaques/µm and 0-0.1 plaques/µm respectively for paxillin and vinculin) compared to GFP+ control cells (peaks at 0.2-0.3 plaques/µm for both paxillin and vinculin) after being normalized to the lamella perimeter. 33 Vinculin Paxillin GFP-KNT+ GFP-KHC+ GFP + Fig 9: Paxillin and vinculin plaque formation in cellular lamella is reduced in GFP-KNT+ over-expressing cells. DsRed2-ER cells were transfected with GFP-KNT+ or GFP-KHC+ and immuno-labelled for either vinculin or paxillin (green). Vinculin and paxillin are represented in green to enhance visual contrast. GFP-KHC+ over-expression resulted in retraction of ER from the cellular lamella and reduction of plaque formation (white arrows) in the leading lamella. Scale bar: 5 µm 34 GFP-KNT+ GFP+ 14 Number of cells 12 10 8 6 4 2 0 0-0.1 0.1-0.2 0.2-0.3 0.3-0.4 0.4-0.5 Paxillin plaques normalized to lamella perimeter GFP-KNT+ 18 GFP+ 16 Number of cells 14 12 10 8 6 4 2 0 0-0.1 0.1-0.2 0.2-0.3 0.3-0.4 0.4-0.5 0.5-0.6 0.6-0.7 Vinculin plaques normalized to lamella perimeter Fig 10: Quantification of paxillin and vincullin in cellular lamella. DsRed2-ER cells transfected with GFP-KNT+ had reduced adhesion plaques compared to GFP+ controls after being normalized to their lamella perimeters. 35 2.2.4 Cell migration is reduced in KNT+ and KHC+ over-expressing cells. Our results suggest that adhesion formation is affected when the kinectin-kinesin interaction is disrupted. We thus investigated the effect of this aberrant adhesion formation on migration of cells. To assess the rate of migration of cells, we utilized both a wound healing assay and a chemotaxis assay. In the wound healing assay, a confluent monolayer of cells is scratched, washed with PBS and allowed to recover for 24h. The wound is imaged at the time of wounding and again after 24h (Fig. 11). The percentage of wound recovery was subsequently quantified (Fig. 12). GFP-KNT+ over-expressing cells were subjected to wounding as described and allowed to heal overnight. GFP-KNT+ cells only had 13.00±5.38% wound closure after 24h incubation compared to 43.24±5.54% in GFP+ over-expressing cells. Disruption of the kinecin-kinesin interaction via GFP-KHC+ over-expression similarly resulted in reduced wound closure of only 13.00±3.11% in GFP-KHC+ over-expressing cells. 36 0h 24 h GFP+ GFPKNT+ GFPKHC+ Fig. 11: Wound healing is reduced during disruption to the kinectin-kinesin interaction. DsRed2-ER cells transfected with GFP+, GFP-KNT+ or GFP-KHC+ and seeded onto glass cover slips. A wound was subsequently created and the wound imaged at 0h and 24h post wounding. 37 Percentage of Wound Healed Wound Healing Assay 60 50 40 30 20 10 0 GFP KNT KHC Cell type Fig 12: Disruption to the kinectin-kinesin interaction reduces wound healing. DsRed2-ER cells transfected with GFP+, GFP-KNT+ or GFP-KHC+ were subjected a wound healing assay and their percentage of wound healing quantified after 24h. Values represent mean±s.e.m from 3 independent experiments. In the chemotaxis assay, chemotaxic inserts with 8µm pore size were used to probe the ability of cells to migrate across a membrane. Cells were trypsinized, resuspended in serum-free media and subsequently added to the insert. Serum containing FBS was used as an attractant outside the inset and the setup left for 4h. A diagrammatic representation of the assay setup is provided in Fig. 13. After 4h, cells were fixed with 3.7% PFA before cells remaining on the inside of the insert swabbed off gently with a cotton bud. Trypan blue was subsequently used to stain cells which migrated successfully across the membrane. Cells imaged and counted under a light microscope to quantify the extent of migration across the membrane (Fig. 13). 38 (A) Well Chemotaxis insert Serum free media Cells Complete media (B) Fig. 13: Schematic representation of the chemotaxis insert. (A) Cells were seeded into a chemotaxic insert and allowed to migrated from serum free media into complete media across an 8 µm membrane over 4 h before being stained. (B) The bottom of the membrane was imaged using a light microscope. Scale bar= 60µm. 39 GFP-KNT+ over-expressing cells were observed to have significant reductions in the number of cells which migrated across the membrane compared to GFP+ overexpressing cells. Only 22.78±3.73 of GFP-KNT+ cells compared to 40.3±11.33 GFP+ cells migrated across the membrane (Fig. 14). GFP-KHC+ over-expressing cells similarly had low numbers of cells migrating across compared to GFP tansfected cells (33±10.26). Chemotaxis-induced cell migration Number of migrated cels/mm2 50 45 40 35 30 25 20 15 10 5 0 GFP KNT+ KHC+ Fig 14: Disruption to the kinectin-kinesin interaction reduces wound healing. Quantification of the bottom membrane of chemoxis insert after DsRed2-ER cells transfected with GFP+, GFPKNT+ or GFP-KHC+ which were allowed to migrate for 4h. Values represent mean±s.e.m from 3 independent experiments. 2.3 Conclusion: Kinectin mediates ER extension into the cellular lamella via kinesin and is required during cellular attachment and migration. Taken together, our data demonstrates that the kinectin is required for ER extension and suggests that it functions mainly in a kinesin dependant manner since overexpression of both GFP-KNT+ and GFP-KHC+ result in absence of the ER in the leading 40 lamella. During disruption of this interaction, the absence of the ER has also been observed to be independent of any microtubule disturbance (Fig. 8), suggesting that its absence is due to kinectin interacting with kinesin (Ong et al., 2000). Furthermore, the kinectin kinesin interaction is required during cellular migration and is required to extend the ER into the cellular lamella, a leading site where new adhesions would be formed (Smilenov et al., 1999). ER extension into the cellular lamella appears to be required for mediating focal adhesion formation. Disruption of this interaction results in significant reduction in both focal adhesion formation and cellular migration. It also results in disruption to individual cell and colony shape. 41 CHAPTER 3: Fibronectin induces ER accumulation via kinectin to facilitate initial cellular attachment The kinectin-kinesin interaction has been demonstrated to be important in mediating ER extension into the cellular lamella. Disruption of ER extension subsequently results in reduction in focal adhesions in the cellular lamella and reduction of cellular migration. We thus hypothesized that ER dynamics is an important factor in mediating cellular adhesion to a surface. To dissect the role of the ER in mediating cellular adhesion, we sought to understand how ER dynamics mediate adhesion dynamics on various naturally occurring ECM substrates. Various ECM proteins interact with different receptors and may thus mediate different pathways during cell adhesion. We selected fibronectin, collagen and laminin as representative ECM proteins for study since each interacts with different receptors on the cell. Fibronectin has been demonstrated to interact with both integrinα5β1 and integrin-α5β3 (Zhang et al., 1995); collagen with integrin-α2β1 and integrinα1β1 (Heino, 2000); laminin binds various cell receptors, including integrins-α1β1, α2β2, α3β1, α6β1, α6β4, α7β1, α9β1, αvβ3 (Ekblom et al., 2003; Scheele et al., 2007; Suzuki et al., 2005), dystroglycans (Yamada et al., 1996) and syndecans (Carey, 1997). We thus sought to investigate the role of each of these ECM proteins in mediating cellular attachment dynamics. 42 3.1 Accelerated cell-matrix adhesion to fibronectin requires ER dynamics via kinectin 3.1.1 Fibronectin surface modification significantly enhances cell attachment To investigate the effect of surface modification on the cellular adhesion dynamics of HeLa cells, we used commercial fibronectin, collagen and laminin adhesion assay plates. Briefly, 1x106 cells were trypsinized, resuspended in serum-free media and re-plated in wells pre-coated with fibronectin, collagen, laminin or BSA. Plates were subsequently centrifuged for 30 sec at 50rpm to cause cells to settle onto the surface and were subsequently left for 10, 20, 30 or 60mins in a 37oC incubator. Cells were then washed gently with PBS to remove unattached cells and stained with the supplied cell dye for 10mins. After washing with PBS, the remaining stained cells were lysed and the lysis buffer read at 560nm to determine the amount of remaining attached cells. As seen in Fig. 15, both fibronectin and collagen induce higher cell attachment compared to the BSA control. Notably, cells in fibronectin-coated wells demonstrate a significantly distinct rise in attachment even after 10min incubation, which subsequently increases to about 2 absorbance units (a.u.) when given more time to adhere. In contrast, collagen appears to increase and plateau at 1 a.u. while laminin completely did not result in increased cellular attachment even after 60min incubation. Together, this data suggests that the fibronectin and collagen surface modifications do induce greater cellular attachment and in the case of fibronectin, distinctly accelerates early cellular adhesion to the matrix. 43 Fibronectin 560nm absorbance (a.u.) 2.5 Collagen Laminin BSA 2 1.5 1 0.5 0 10 20 Time (min) 30 60 Fig. 15: Fibronectin significantly enhances cellular attachment. HeLa cells were seeded onto fibronectin, collagen or laminin coated 48-well plates and briefly spun. Cells were allowed to adhere for the specified time frame and subsequently washed, stained and quantified. Fibronectin is able to induce strong cellular attachment even after only 10 min while collagen is able to induce some attachment. Laminin did not result in significant cellular attachment during the timeframe of the experiment. Values represent mean±s.e.m from 2 independent experiments. 3.1.2 Kinectin is required for initial fast cell attachment to fibronectin We have demonstrated earlier in this thesis that ER extension into the cellular lamella is dependent on kinectin. Disruption of kinectin’s function results in reduction to focal adhesion formation in the cellular lamella. Since attachment of cells to the ECMmodified surface would similarly require formation of focal adhesions, we sought to investigate how disruption of kinectin affects the dynamics of cellular adhesion. KNTKD, KNTVC and KNT WT cells were seeded onto fibronectin, collagen and laminin adhesion plates and subsequently assayed for cell attachment (Fig. 16). The fibronectin surface modification greatly facilitates cellular attachment across all time frames assayed. At 10mins, both KNTVC and KNT WT cells adhered better to the 44 fibronectin-coated surface compared to either collagen or laminin. After 60min, fibronectin surface resulted in a significantly larger absorbance of 2.22±0.12 a.u, compared to collagen (1.04±0.04 a.u.), laminin (0.28±0.05 a.u.) or BSA (0.15±0.02 a.u.). KNTKD cells however did not adhere as well to fibronectin as both KNTVC and KNT WT cells. At both 10 and 20min, KNTKD absorbance was comparable to that of collagen levels and only after 30mins did it result in a jump above collagen in absorbance, suggesting that disruption of kinectin results in slower attachment response of cells to fibronectin. The 60min absorbance reading of KNTKD cells is also reduced at 1.64±0.05 a.u. Adhesion to collagen or laminin were not significantly affected by kinectin disruption as observed by the similar levels of absorbance. Taken together, our data suggests that fibronectin can enhance early cellular adhesion and that kinectin is mediates this process. 45 20 min 10min Collagen Laminin 2.5 2.5 2 2 560nm absorbance (a.u) 560nm absorbance (a.u) Fibronectin 1.5 1 0.5 Fibronectin Laminin 1.5 1 0.5 0 KNTKD KNTVC 0 KNDWT KNTKD KNTVC Cell type Fibronectin 2.5 Collagen KNDWT Cell type 30min 60min Fibronectin Laminin Collagen Laminin 2.5 2 560nm absorbance (a.u) 2 560nm absorbance (a.u) Collagen 1.5 1 0.5 1.5 ` 1 0.5 0 0 KNTKD KNTVC Cell Type KNDWT KNTKD KNTVC KNDWT Cell Type Fig. 16: Kinectin knockdown results in reduced adhesion to fibronectin. KNTWT, KNTVC and KNTKD cells were seeded onto fibronectin, collagen and laminin coated surfaces for various times. Knockdown of kinectin resulted in reduced significantly reduced cellular attachment to fibronectin especially within the first 20min but did not significantly alter attachment dynamics of both collagen and laminin. Values represent mean±s.e.m from 2 independent experiments. 46 3.1.3 ER extension via the kinectin-kinesin interaction is important for accelerated cell attachment to fibronectin To probe kinectin’s role in greater detail, we further assayed the effect on cellular attachment by specifically disrupting the kinectin-kinesin interaction via over-expression of GFP-KNT+ in wild-type HeLa cells over 60min. GFP+ over-expressing cells were used as a control to this experiment (Fig. 17). Consistent with previous observations, both KNT-GFP+ and GFP+ cells attached best to fibronectin-coated surfaces across all time frames. Fibronectin also significantly induces more cellular attachment compared collagen or laminin even after only 10min of attachment. Thus our findings are in agreement with reports from other groups that fibronectin functions within 15min to strengthen adhesion, in a cytoskeleton dependant manner, by more than an order of magnitude (Lotz et al., 1989). Disruption of the kinectin-kinesin interaction however reduced the ability of GFP-KNT+ cells to adhere to the fibronectin-coated surface compared to GFP+ over-expressing cells, suggesting that ER extension, mediated by kinectin on microtubules, may be required for proper cellular attachment. The reduction in cellular attachment by GFP-KNT+ disruption is notably less than that of kinectin knockdown, suggesting that kinectin may have other functions besides mediating interaction with kinesin. It is possible that interactions with other binding partners such as Rho proteins or elongation factor-1δ are not disrupted by GFP-KNT+, thus reducing the disruption to cellular activity. Comparatively, collagen coated-surfaces resulted in less cells adhering to the surface compared to fibronectin-coated surfaces although this was still significantly greater than that induced by laminin-coated surfaces (Fig. 17). Collagen, in contrast to 47 fibronectin, appears to elicit different, possibly ER independent, mechanisms during cell attachment. This is apparent since over-expression of GFP-KNT+ did not significantly abolish cellular attachment. Lastly, cell attachment to laminin-coated surfaces remained low even after 60min and similarly remained unresponsive to disruption of the kinectinkinesin interaction, suggesting that laminin does not function significantly to mediate initial ([...]... enrichment of kinectin (Tran et al, 2002), a kinesin binding protein, suggest that the ER can localize and may actively modulate adhesion dynamics Kinectin in particular is of significance as it utilizes the ATP-dependent kinesin motor protein, making its activity controllable 1.3.4 Kinectin: A multifunctional protein that interacts with kinesin Kinectin, an integral membrane protein found in ER (Kumar... over-expressing KHC or KNT+ Fig 4: KNT+ and KHC+ are competitive inhibitors of kinectin and kinesin Minimal binding domains of kinectin (KNT+) and kinesin (KHC+) to each other have been previously identified (Ong et al., 2000;Santama et al., 2004) Over-expression of either KNT+ or KHC+ results in competitive inhibition of the kinectin- kinesin interaction by blocking of the binding domain on kinectin or kinesin... schematically in Fig 1 12 Kinesin Elongation Factor-1 δ Rho GTPase Kinectin Fig 1: A schematic representation of kinectin The full length kinectin comprises an N-terminal transmembrane domain spanning the ER membrane and a large Cterminal coiled coil forming cytoplasmic part Kinectin interacts with kinesin, EF-1δ and Rho GTPase via its cytoplasmic C-terminal (arrows) 1.4 Role of microtubule in cell adhesions... anchors, the kinectin- kinesin interaction appears most prominent Kinectin interacts with kinesin and the kinesin-binding domain enhances the microtubule-stimulated kinesin-ATPase activity (Ong et al., 2000), enhancing transport by kinesin to the adhesions The dual function of microtubules in inducing assembly as well as disassembly suggests that adhesion regulation by microtubules is complicated and may... attachment in cells induced by fibronectin, collagen and laminin during cell attachment was also investigated This work will extend our understanding of ER-related events occurring in cells during attachment to ECM matrix proteins, paving the way for development of systems and methods leading to greater control over cell behaviour in tissue engineering 17 CHAPTER 2: Role of the Kinectin- kinesin interaction in. .. et al., 2009) The overexpression of kinesin binding fragment of kinectin resulted in perinuclear clustering of mitochondria, which is strikingly reminiscent of the phenotype in conventional kinesin heavy chain deficient cells (Tanaka et al., 1998) Further characterization of kinectin revealed that besides binding to kinesin, kinectin also interacted with Rho GTPase and was found to anchor the translation... kinesin heavy chain gene (Yang et al., 1989; Goldstein and Gunawardena, 2000) This suggests that kinectin may not be the main dynamic anchor present in certain cells or that alternative mechanisms may be compensating for kinectin s absence Two main kinectin isoforms of 160kDa and 120kDa exist (Leung et al., 1996) Kinectin exists as both a 160 kDa ER bound protein and a shorter 120 kDa kinectin, which lacks... splicing The variable domains 3 (vd3, residues 1177-1200) and 4 (vd4, residues 12291256) have been mapped as the minimal kinesin -kinectin interaction domain on kinectin (Kumar et al., 1995; Ong et al., 2000) Knockdown of kinectin disrupts the kinectinkinesin interaction and leads to collapse of ER network from the cell periphery (Santama et al., 2004) Over-expression of vd4 domains (KNT+), or the kinectin. .. Over-expression of vd4 domains (KNT+), or the kinectin binding domain (KHC+), resulted specifically in competition with endogenous kinectin or kinesin for binding with its partner This disrupts the kinectin- kinesin interaction and similarly results in ER adopting a retracted morphology A recent finding that over-expression of a kinesin-1 dominant negative mutant inhibited ER extension lends support to this... binding induces conformational changes in integrin which is required for its clustering and subsequent function in recruiting focal adhesion proteins (Miyamoto et al., 1995; Hato et al., 1998) Focal adhesions contain proteins that link integrin with cytoskeleton and recruit adhesion related proteins such as vinculin, paxillin and focal adhesion kinase (Zaidel-Bar et al., 2003) The integrin β subunit binds ... ATP-dependent kinesin motor protein, making its activity controllable 1.3.4 Kinectin: A multifunctional protein that interacts with kinesin Kinectin, an integral membrane protein found in ER (Kumar... dynamic anchors, the kinectin- kinesin interaction appears most prominent Kinectin interacts with kinesin and the kinesin-binding domain enhances the microtubule-stimulated kinesin-ATPase activity... to investigate the role of ER localization in mediating cell function via kinectin- kinesin interaction 2.1 ER-bound kinectin mediates ER extension 2.1.1 Kinectin knockdown disrupts kinectin- kinesin