Investigate the function of kinectin isoforms in endoplasmic reticulum dynamics

In order to explore the function of different kinectin isoforms, a series of EGFP-tagged chimera proteins were constructed including KNT1 no insert, KNT15 Int3 only, KNT2 Int4 only, KNT1

THE FUNCTION OF KINECTIN ISOFORMS IN

ENDOPLASMIC RETICULUM DYNAMICS

YAJUAN ZHU

(B Eng., Zhejiang Univ., China)

A THESIS SUBMITTED FOR THE DEGREE OF

MASTER OF SCIENCE

GRADUATE PROGRAMME IN BIOENGINEERING

NATIONAL UNIVERSITY OF SINGAPORE

2006

First of all, I would like to express my deepest gratitude to my supervisor, Associated Professor Hanry Yu, not only for his technical direction in my research work, but also for his philosophical inspiration that would be helpful throughout my life Many

of the original ideas in my research came from his inspirational suggestions and fruitful discussion

I would also like to express my appreciation to Dr Lee Lee Ong, who patiently helped me through all the techniques and troubleshooting process, as well as to Mrs Pao Chun Lin and Mrs Xin Zhang for their cooperation and all beneficial suggestions on this project

I am also grateful to Mr Jun Ni, Mr Xiaotao Pan, Mr Kong Heng Lee and Mr Wangxin Sun for their assistance on cellular imaging and image processing In addition, I would like to thank Dr Ser Mien Chia, Mrs Lijuan He, Mrs San San Susanne Ng and all the other members in Prof Yu’s lab for creating a friendly and happy environment for my research

My deepest gratitude goes to my parents and my husband Zhiling for their love, understanding and sacrifice Their support is an indispensable source of my strength and confidence to overcome any barrier

Extended appreciation goes to NUS for supporting me financially and providing

me the opportunity with the research facilities during my course of research for my master degree

SUMMARY

In this project, a basic cell biology problem, the function of kinectin isoforms during ER dynamics, was explored using multiple molecular and cellular imaging techniques Kinectin has been proposed as a membrane anchor for kinesin on intracellular organelles Both the 160 kDa kinectin, an integral transmembrane protein found mainly in the ER, and the 120 kDa form, a truncated version without the N-terminal transmembrane domain, have been reported In addition, there are at least five small inserts (23-33 residues, corresponding to different exons) scattered throughout the C-terminus of kinectin sequence, which also contribute to variable isoforms by alternative splicing In order to explore the function of different kinectin isoforms, a series of EGFP-tagged chimera proteins were constructed including KNT1 (no insert), KNT15 (Int3 only), KNT2 (Int4 only), KNT11 (Int5), KNT9 (Int 3, 4 and 5) Both the 160 kDa and 120 kDa forms have been cloned A HeLa cell line constitutively expressing a DsRed-ER marker (Clontech) was also established (DsRed-ER HeLa) and EGFP-tagged kinectin isoforms were microinjected into this cell line to investigate their subcellular localization relative

to ER dynamics under confocal microscopy All five 120 kDa isoforms showed diffused distribution pattern in cytosol with some organelle-like staining in perinuclear areas and obvious accumulation at lamellipodia But no co-localization with ER was observed On the contrary, all five 160 kDa isoforms revealed significant co-localization with ER throughout the cell body Interestingly, two of them, KNT2 and KNT9, which contained Int4, also co-accumulated with ER in the tips of lamellipodium-like structures Live imaging confirmed these were leading edges of migrating cells Since Int4 has been mapped into the minimal kinesin-interacting domain, these observations suggested kinesin mediated transportation along microtubules might contribute to ER dynamics

during cell migration For the other three isoforms without Int4, neither kinectin themselves nor ER accumulated at the cell periphery

To further investigate ER dynamics at the leading edge, a small kinectin fragment harboring Int4, was overexpressed in DsRed-ER HeLa cells The overexpression disrupted the endogenous kinesin-kinectin interaction and inhibited ER extension into the cell periphery The migration speed of these cells also decreased in wound healing assays and the level of decrease was similar to that in kinectin knockdown cells using RNA interference (RNAi) These results suggested kinectin isoforms with Int4 contributed to cell migration by transporting ER to the leading edge along microtubules Furthermore, the importance of Int4 in ER dynamics during cell division was also investigated Z-stack time-lapse imaging of migrating cells revealed intact ER structure with well-regulated dynamics and the strong association with the mitotic spindle throughout mitosis However, when cells were treated with morpholinos which specifically knocked down Int4 containing kinectin isoforms, neither the overall ER dynamics nor the association with the spindle was affected, indicating ER dynamics during mitosis might be kinectin- independent

In summary, these results from multiple imaging techniques demonstrated the importance of kinectin isoforms with Int4 in modulating ER dynamics of interphase cells These findings would serve as a platform for more detailed studies of kinectin isoforms The opportunities opened up by quantitative bioimaging might help us to identify subtle differences in their functions in the future

LIST OF FIGURES Fig 1.1 The structure of kinectin

Fig 1.2 Novel kinectin isoforms

Fig 1.3 The current model for the maintenance of ER dynamics in mammalian cells

Fig 2.1 The cloning strategy of EGFP-tagged kinectin isoforms

Fig 2.2 RT-PCR products of the first and second cDNA fragments of 160 kDa kinectin

(160Knt1 and Knt2)

Fig 2.3 PCR products of the first cDNA fragment of 120 kDa kinectin (120Knt1)

Fig 2.4 PCR products of the third cDNA fragments of different kinectin isoforms (Knt3) Fig 2.5 Ligation products of EGFP-tagged kinectin isoforms examined by restriction

mapping with EcoR I/Xba I/BamH I

Fig 2.6 Ligation products of EGFP-tagged kinectin isoforms estimated by restriction

mapping with Hind III

Fig 2.7 The survival curve of wild type HeLa cells under G418 selection

Fig 2.8 The screening of single clones stably expressing the DsRed-ER marker

Fig 2.9 The subcellular localization of EGFP-120 kDa kinectin isoforms and soluble

EGFP in DsRed-ER HeLa cells

Fig 2.10 The subcellular localization of EGFP-160 kDa kinectin isoforms in DsRed-ER

HeLa cells

Fig 2.11 The co-accumulation of the ER and 160 kDa kinectin isoforms of Int4 at the

leading edge of migrating cells

Fig 3.1 The effect of stable expression of pSilencer/KNT RNAi on the endogenous

kinectin protein level

Fig 3.2 Kinectin knockdown using RNAi inhibited the migration of HeLa cells in wound

healing assays

Fig 3.3 The disruption of kinesin-kinectin interaction through Int4 overexpression

resulted in the retraction of ER from the cell migration leading edge

Fig 3.4 The overexpression of Int4 fragment inhibited HeLa cell migration in wounding

healing assays

Fig 3.5 The proposed model for the role of Int4 containing-kinectin isoforms in ER

dynamics along microtubules

Fig 3.6 Four dimensional (3D plus time) images of the mitotic ER dynamics

Fig 3.7 Splicing of kinectin pre-mRNA Exon 40 and the knockdown of Int4 in the

presence of morpholinos in DsRed-ER HeLa cells

Fig 3.8 Mitotic ER dynamics after morpholino treatment

TABLE OF CONTENTS

ACKNOWLEDGEMENT i

SUMMARY ii

LIST OF FIGURES iv

TABLE OF CONTENTS vi

Chapter 1 Introduction 1

1.1 Cytoskeleton and molecular motors 1

1.2 Kinectin 5

1.2.1 An overview of motor protein receptors 5

1.2.2 Kinectin: a transmembrane receptor for kinesin 6

1.2.3 Kinectin isoforms 8

1.2.4 Kinectin in organelle motility 11

1.2.5 Other roles of kinectin and its clinical implication 13

1.3 ER dynamics 15

1.3.1 The establishment and maintenance of organelle positions inside cells 15

1.3.2 Microtubule-dependent ER dynamics 15

1.3.3 Actin in the ER movement 18

1.4 Cell migration 19

1.4.1 The four-step concept of cell migration 19

1.4.2 The MT based membrane transport in cell migration 22

1.5 Cell division 25

1.5.1 Membrane partitioning during cell division 25

1.5.2 ER partitioning during mitosis and the role of microtubules 27

1.6 The purpose and rationale of the thesis work 29

Chapter 2 The subcellular localization of kinectin isoforms in DsRed-ER HeLa 31

2.1 The construction of EGFP-tagged kinectin isoforms 32

2.1.1 The isolation of first two fragments of full length kinectin 34

2.1.2 The isolation of variable C-terminus of different kinectin isoforms 39

2.1.3 The construction of EGFP-tagged full-length kinectin isoforms 42

2.2 The establishment of the DsRed-ER HeLa stable cell line 46

2.3 The subcellular localization of kinectin isoforms in DsRed-ER HeLa 51

2.3.1 The distribution of kinectin isoforms in fixed DsRed-ER HeLa 51

2.3.2 The co-accumulation of the ER and 160 kDa kinectin of Int4 at the leading edge of migrating cells 60

Chapter 3 Roles of kinectin Int4 in ER dynamics during cell migration and division 63

3.1 Roles of kinectin Int4 in ER dynamics during cell migration 63

3.1.1 The HeLa cell migration is inhibited by kinectin knockdown using RNAi .64

3.1.2 Int4 overexpression affected ER extension into the leading edge and thus inhibited the HeLa cell migration 70

3.2 Roles of kinectin Int4 in ER dynamics during cell division 77

3.2.1 ER dynamics during cell division 78

3.2.2 The effect of Int4 knockdown on mitotic ER dynamics 83

Chapter 4 Conclusions and future prospects 87

Chapter 5 Materials and Methods 90

5.1 Isolation of kinectin cDNA 90

5.1.1 Isolation of total RNA from Swiss-3T3 cells 90

5.1.3 Reverse Transcription (RT) 91

5.1.4 Polymerase chain reaction (PCR) amplification 91

5.1.5 Agarose gel electrophoresis 94

5.1.6 Extraction of DNA from agarose gel 95

5.1.7 TOPO TA cloning 95

5.1.8 Transformation 96

5.1.9 Small-scale plasmid preparation 96

5.1.10 Restriction endonuclease digestion 97

5.1.11 DNA sequencing 97

5.2 Construction of EGFP-tagged full length kinectin isoforms 98

5.2.1 Restriction endonuclease digestion for ligation 98

5.2.2 Ligation of the four DNA fragments 98

5.2.3 Preparation of competent cells 99

5.2.4 Transformation 99

5.2.5 Medium-scale plasmid preparation 100

5.2.6 Quantification of DNA 100

5.3 Mammalian cell culture 101

5.4 Stable HeLa cell lines expressing the DsRed-ER marker 101

5.4.1 Optimization of G418 concentration 101

5.4.2 Transfection 101

5.4.3 Selection of stably transfected cell lines 102

5.5 Subcellular localization studies of kinectin isoforms 102

5.5.1 Microinjection of EGFP fused kinectin isoforms into DsRed-ER HeLa102 5.5.2 Confocal laser scanning microscopy for fixed cell and live imaging 103

5.6 Stable kinectin knockdown cells lines using pSilencer vectors 103

5.6.1 Establishment of stable lines 103

5.6.2 Estimating the knockdown efficiency by immunoblotting 104

5.7 Wound healing assay 106

5.8 Overexpression studies 106

5.8.1 Plasmid construction 106

5.8.2 Overexpression of the kinectin fragment 106

5.9 Transient knockdown of Exon40 of kinectin using morpholinos 107

5.9.1 Morpholino design 107

5.9.2 Delivery of morpholinos into cells 107

5.9.3 Estimating the knock-down efficiency at mRNA level and protein level .108

REFERENCES 110

Chapter 1 Introduction

1.1 Cytoskeleton and molecular motors

The entire cytoplasm of eukaryotes is filled with a dynamic three-dimensional structure, the cytoskeleton, whose importance has been appreciated since the earliest days

of cell biology The cyto-architecture contributes not only to the diversity in shapes and sizes of cells but also to the multifaceted function of variable cell types (Fuchs and Karakeisogou, 2001) In eukaryotic cells, cytoskeleton is composed of three distinct elements: microtubules which form a polarized network allowing organelle and protein movement, actin microfilaments which provide protrusive and contractile forces, and intermediate filaments which are most rigid component responsible for the maintenance

of the overall cell shape

Microtubules are tubular and hollow filamentous structures with a constant diameter of 20-25 nm They are composed of globular tubulin, which exists predominantly as heterodimer of α and β subunits of about 55 and 50 kDa respectively (Alberts et al., 2002) Tubulin dimmers align within microtubules leaving α-tubulins exposed to one end and β-tubulins to the other, an arrangement which confers intrinsic structural polarity to microtubules (Nogales et al., 1999) The α-tubulin is bound with a GTP that is never hydrolyzed or exchanged, whereas the nucleotide bound to β-tubulins could be either GTP or GDP In addition, microtubules also display kinetic polarity in that the polymerization rate of the plus end (β-tubulin exposed) is faster, while the one of the minus end (α-tubulin exposed) slower The polarized arrays of microtubules provide tracks for the intracellular transport of membrane-enclosed organelles In most cells, microtubules radiate from an organizing center (MTOC) adjacent to the nucleus, and extend throughout the cytoplasm towards the cell periphery (Lane and Allan, 1998)

Since the minus ends of microtubules are close to the MTOC in such cells, any organelles

or vesicles traveling to this end will accumulate at the cell center, whilst organelles moving to plus ends will be transported to the cell periphery The bi-directional movements along polarized microtubules play important roles in the transportation in neural axons and the establishment of polarity in epithelial cells (Hirokawa, 1998)

Actin microfilaments are originally discovered as contracting bundles in muscle contraction They are fine thread-like protein fibres, 3-6 nm in diameter (Alberts et al., 2002) A 43-45 kDa globular protein, actin, forms the basic unit of actin filaments A protofilament with polarity is formed by actin assembling end-to-end, and then two protofilaments of the same polarity wrap in a helix to form a microfilament The assembly of actin monomer at two ends is of different dynamics, with one end designated the barbed end and the other pointed end The assembly and function of actin filament is extensively regulated by numerous capping proteins and actin monomer binding proteins, both of which are sensitive to intracellular signalling (Schafer and Cooper, 1995; Sun et al., 1995; Zigmond, 1996)

Intermediate filaments are discovered recently as rope-like fibres with average diameter of 10 nm which is ‘intermediate’ between the size of actin filaments and the one

of microtubules (Alberts et al., 2002) At least 67 human genes encode functional intermediate-filament proteins (Hermann et al., 2003) The individual monomers are elongated molecules with an extended central α-helical rod domain that form a parallel coiled-coil with another monomer A pair of dimmers then associates in an anti-parallel fashion to form a staggered tetramer The tetramers finally assemble together laterally to form filament without a structural polarity (Chang and Goldman, 2004) In most cells, intermediate filaments extend radially in all directions and form cage-like networks In

though cross-bridge molecules which range from molecular motors to multidomain proteins such as plectin (Helfand et al., 2002; Yabe et al., 2000) Thus intermediate filaments may have a central role in coordinating cytoskeletal interactions Further more, many kinases and cofactor are found to bind with phosphorylated intermediate filaments, placing them in a favorable position for mediating molecular crosstalks among signalling events of the other cytoskeletal systems (Meyer and Feldman, 2002; Tzivion et al., 2000)

Two types of cytoskeleton-mediated intracellular transport have been discovered, namely microtubule-based and actin-based transportation (Langford, 1995) No organelles have yet been found to move on intermediate filaments Intracellular organelle transport is fundamental to cellular functions such as endocytosis, secretion, phagocytosis

in macrophages, fast axonal transport in neurons and antigen presentation in lymphocytes Abnormalities of these cellular phenomena have been correlated with various diseases

Previous efforts to understand intracellular transportation have led to the partial identification of some components of this system These components include tracks (microtubules or actin filaments) on which organelles are transported, motor proteins which drive the organelles, activators that regulate the activities of motor proteins, and motor receptors that anchor the organelles to motor proteins

Molecular motors transform chemical energy derived from ATP hydrolysis into mechanical force to drive cellular motility (Howard and Hyman, 2003) They bind to polarized cytoskeletal filaments using the ‘head’ region (motor domain) and ‘walk’ along the filaments though cycles of nucleotide hydrolysis and conformational change Molecular motors differ in the type of filaments (either microtubules or actin filaments) they bind to, the direction in which they move, and the ‘cargo’ they carry Most motors carry membrane-enclosed organelles such as ER, mitochondria, Golgi or secretory vesicles and transport them to their desired destination along cytoskeletal tracks (Cole

and Lippincott-Schartz, 1995) Besides this, motor proteins can also cause cytoskeletal filaments to slide against each other and generate the force to drive muscle contraction, ciliary beating and cell division (Albert et al., 2002) Basically, there are three classes of molecular motors They are myosin which moves along actin filaments, kinesin and dynein which bind to microtubules and responsible for plus-end- and minus-end-directed movement respectively The focus of my project concerns kinesin

Plus end directed motility along microtubules is driven by conventional kinesin and a number of kinesin related proteins Conventional kinesin is composed of two 120 kDa heavy chains (KHC) and two 64 kDa light chains (Hirokawa et al., 1989) It has a rod-like structure with two globular heads, an α-helical coiled-coil stalk and a fan-like end More detailed analysis of structure and sequence suggests that the globular head, also called the motor domain which generates force, is composed of a catalytic domain which hydrolyzes ATP and interacts with microtubule and a short neck domain important for the control of direction (Vale and Fetterick, 1997; Romberg et al., 1998) Conventional kinesin has punctuated distribution in cells as imaged by immunofluorescence, and associates with a variety of organelles (Pfister et al., 1989; Henson et al., 1992; Schmitz et al., 1994; Lippincott-Schwartz et al., 1995) Genetic

analyses in C elegans and Drosophila melanogaster have shown that kinesin mutations

cause neuronal defects of varying severity including paralysis and death (Gho et al., 1992; Hurtley and Helenius, 1989)

Besides conventional kinesin, several subfamilies of kinesin related proteins could also server as molecular motors which mediate either plus end- or minus end- directed movement along microtubules Through the combination of molecular biological approaches together with BLAST search in genome databases, a total of 45 kinesin

al., 2001) Recently a new kinesin tree including 155 proteins from eleven species was reported (Dagenbach and Endow, 2004) KIFs can be divided into three classes based on the position of the motor domain, which could be located either N-terminally (N-kinesins), C-terminally (C-kinesins) or internally (M-kinesins) (Miki et al., 2001) N-kinesins could

be subdivided into 11 classes, among which KIF1, KIF3, KIF4, KIF5 (conventional kinesin), and KIF17 are major members M-kinesins consist of only KIF2 C-kinesins are composed of KIFC1 and KIFC2/C3 In general, N-kinesins and M-kinesins move towards the plus ends, while C-kinesins move towards the minus ends Besides, M-kinesin also has a unique microtubule-depolymerising activity Although the motor domains of different KIFs are highly conserved, other regions are quite divergent and may serve as cargo-binding domains This diversity explains why KIFs is capable of transporting a wide variety of cargoes, including membranous organelles, macromolecular complexes and mRNAs (Aizawa et al., 1992; Yonekawa et al., 1998; Zhao et al., 2001)

1.2 Kinectin

1.2.1 An overview of motor protein receptors

The binding of cargos to motor protein is mediated by motor protein receptors The existence of many motor protein receptors seems essential given the plethora of motors and cargoes in cells So far, identified kinesin receptors include coat proteins, scaffold proteins and transmembrane proteins These motor receptors are important for recruiting specific cargoes to kinesins

The first step for the transport of cargo proteins between membrane-bound organelles is self-assembly of coat proteins, such as clathrin, on the cytoplasmic side of the membrane The AP-1 clathrin-associated adaptor complex, which mediates the transport of clathrin-coated vesicles from the trans-Golgi network to plasma membrane,

has been reported to bind to KIF13A (Nakagawa et al., 2000) Spectin, another type of coat protein with different isoforms, is also found to interact with motor proteins For example, fodrin, a neuronal isoform of spectrin, was reported to bind to KAP3, the accessory subunit of KIF3 which has been suspected to be involved in organelle binding (Takeda et al., 2000)

The tail domains of several KIFs contain protein-protein interaction motifs, such

as SH2, SH3 and PDZ domains, which may associate with scaffold proteins Thus components of signalling cascades could be recruited to motor proteins through these scaffold proteins to drive the regulated movement of attached cargo (Klopfenstein et al., 2000) This hypothesis is supported by the observation that conventional kinesin light chain binds to JIP-1 and JIP-2, scaffolding proteins of the c-Jun N-terminal kinase (JNK) signalling pathway (Verhey et al., 2001) The iterations of kinesin and JIP proteins might

be essential for specific localization of JIPs to the nerve terminal as well as the spatial organization of the signalling pathway within the cell

Several transmembrane proteins have been identified as receptor for kinesin, such

as amyloid precursor protein (APP) and Sunday driver (SYD) APP was found to form a complex with kinesin by directly binding to the TPR domain of KLC (Kamal et al., 2000; Kamal et al., 2001) SYD, which was discovered in a genetic screen for axonal transport

mutants in Drosophila, also bind to TRP domain of KLC (Bowman et al., 2000) SYD

likely mediates functional interaction of KIF5 with axonally transported post-Golgi vesicles

1.2.2 Kinectin: a transmembrane receptor for kinesin

Kinectin was initially discovered as a protein bound to conventional kinesin when the detergent solubilized chick embryo brain microsomes preparation was passed through

cDNA was cloned by immunoscreening the embryonic chick brain cDNA library with three different anti-kinectin monoclonal antibodies (Yu et al., 1995) A human kinectin homologue was also identified from human lymphoid cell cDNA library using a similar approach Chicken and human kinectin protein sequences share 61-70% identity (Futterer

et al., 1995) Subsequently, kinectin was isolated from mouse and fox, whose sequences show 83% and 89% homology with human kinectin respectively (Leung et al., 1996; Xu

et al., 2002) However, a kinectin-like sequence was not found in C elegans or

Drosophila genomes even though they harbor conventional kinesin heavy chain genes

(Goldstein and Gunawardena, 2000; Rubin et al., 2000)

Kinectin is thought to be an integral membrane protein since it is resistant to alkaline extraction Primary sequence analysis reveals that kinectin is a highly hydrophilic α-helical coiled-coil protein with a short hydrophobic region at N-terminus (Fig 1.1) The hydrophobic region is likely a transmembrane domain since the transient expression of kinectin lacking the first 50 residues in CV1 cells leaded to an immunolocalization pattern distinctly different from the ER distribution of endogenous

kinectin (Toyoshima et al., 1992; Yu et al., 1995) In addition, in vitro translation of a 30

kDa N-terminal fragment in the presence of microsomes shows the association of this fragment with membranes The presence of high proline content is conserved in the N-terminus of kinectins from different species, which may result in an extreme rigid structure of this transmembrane domain (Futterer et al., 1995) Multiple heptad repeats were found between residues 318 and 1330, which are predicted to form α-helical coiled-coil, implying that kinectin may form dimmers (Leung et al., 1996; Yu et al., 1995; Kumar et al., 1998) These observations also suggest that kinectin is predominantly a rod-shaped protein, a notion which has been confirmed by electron microscopy analysis (Kumar et al., 1998)

Northern blot analysis showed that mouse kinectin transcripts were expressed in tissues such as 12-day embryo, adult heart, brain, ovary, kidney, lung, colon, small intestine, spleen, thymus and pancreas (Leung et al., 1996), while fox kinectin was detected in adult fox brain, heart, kidney, liver, lung, muscle, spleen and testis (Xu et al., 2002) The subcellular distribution of kinectin was revealed by immunostaining with anti-kinectin antibodies Kinectin was found primarily in the ER but no colocalization with Golgi and endosomes was observed (Futterer et al., 1995; Yu et al., 1995) In motor neurons, the cell bodies and dendrites were punctuately stained with anti-kinectin antibodies However, kinectin was not found in the axons where kinesin was detected (Hollenbeck, 1989; Niclas et al., 1994)

Fig 1.1: The structure of kinectin The full length kinectin comprises an N-terminal

transmembrane domain spanning the ER membrane and a large C-terminal coiled forming cytoplasmic part A shorter 120-kDa kinectin, lacking the initial 232 N-terminal residues, may bind to 160-kDa kinectin as a heterodimer Either 160 kDa or 120 kDa kinectin alone may also form homodimers There are at least five short inserts scattered throughout the C-terminus, whose alternative splicing contributes to variable isoforms

coil-1.2.3 Kinectin isoforms

Two kinectin isoforms of 160 kDa and 120 kDa were reported in human and chicken, but only a single 150 kDa form was found in mouse (Leung et al., 1996) Antibodies against the C-terminus of kinectin recognize both forms, whereas the one against N-terminus reacts only with the 160 kDa protein indicating the transmembrane domain is missing in 120 kDa kinectin (Futterer et al., 1995) N-terminal sequencing

amino acids from the N-terminus of full-length kinectin Actually both isoforms are encoded by the same mRNA (Kumar et al., 1998), but the 120 kDa kinectin may result from an alternate translation start site (unpublished observation from McGoldrick C and Sheetz M.) or proteolytic processing In chick embryo fibroblasts, both forms were identified in high density fractions whereas only 120 kDa form was found in the low density vesicles (Kumar et al., 1998) By forming a heterodimer with 160 kDa kinectin which has the transmembrane domain, the truncated 120 kDa form could be anchored to organelles (Fig 1.1) However, 120 kDa kinectin homodimers lacking this strong membrane attachment may attach to low density membranes through myristoylation or remain soluble in the cytosol Recently, it is reported that the treatment of cells with diverse apoptotic stimuli leads to rapid proteolytic cleavage of the 160 kDa kinectin to a

120 kDa form (Machleidt et al., 1998) Caspase 7 mediated this cleavage during apoptosis Thus cleavage may results in disruption of the highly regulated membrane trafficking pathways by disconnecting membrane vesicles from the microtubule-based transport system

The C-terminus of kinectin can also contribute to variable isoforms though alternative splicing (Fig 1.1) There are at least five small inserts (23-33 residues) scattered throughout the C-terminus of mouse kinectin sequence (Leung et al., 1996), while six inserts in human (Futterer et al., 1995) or chicken kinectin sequence (Yu et al., 1995) Tow different splice variants have been reported for human kinectin (Futterer et al., 1995), but more variants with different splicing pattern were found in human hepatocellular carcinoma (Wang et al., 2004) Recently, fifteen novel kinectin isoforms have been isolated from mouse nerve tissue such as embryonic/adult hippocampus and cultured astrocytes (Santama et al., 2004) This brings the total known number of kinectin isoforms in mouse to 16 (Fig 1.2) The identification of total of 16 kinectin isoforms

reveals the long-suspected existence of a family of kinectin isoforms Since the C-domain

is exposed for interaction with motor proteins or other molecules whereas N-terminus is embedded in the membrane, the existence of alternative C-terminal ends may have functional relevance such as determining the directionality of transport or modulating motility of motor-membrane complexes The minimal kinesin-interacting domain of kinectin has been mapped between residue 1188-1288, which consists of the part of insert

3 (Int3), the whole of Int4 and the constant domain linking Int3 and Int4 (Ong et al., 2000) This region enhances the microtubule stimulated ATPase activity of kinesin and inhibits the kinesin-dependent lysosome motility Isoforms containing Int2 are overexpressed in human hepatocellular carcinoma cancerous tissues suggesting that this domain may be involved in tumor cell proliferation, motility or anti-apoptotic ability (Wang et al., 2004)

Fig 1.2: Novel kinectin isoforms These isoforms contain combination of different

inserts at C-terminus and are found in distinct cell types and developmental stages in mouse hippocampus (Santama et al., 2004) Isoforms marked with a star were constructed for our work

E15 embryonic hippocampus

Adult hippocampus

Glia (Astrocyte)

1.2.4 Kinectin in organelle motility

Consistent with its predominant expression in ER, kinectin is reported to contribute to ER dynamics through anchoring ER to kinesin motor proteins The antibody against kinectin inhibits conventional kinesin binding to microsomes and reduces

organelle motility in vitro (Kumar et al., 1995) The binding of cytoplasmic dynein to

microsomes and dynein-dependent organelle motility are also reduced in the same experiments This raises the question of whether kinectin could also be a receptor for dynein However, it was recently reported that ER branching was remarkably reduced and collapsed to the cell center upon kinectin knockdown using RNAi, indicating the involvement of kinectin in plus-end directed transport of ER (Santama et al., 2004) These data also suggest that the inhibition on dynein-dependent motility by kinectin antibody might be a result of steric effect since conventional kinesin and cytoplasmic dynein may exist in close spatial apposition (Brady et al., 1990)

Although earlier study using antibody recognizing the N-terminal end of kinectin failed to reveal the localization of kinectin to mitochondria, a recent study reported the specific accumulation of 120 kDa kinectins but not 160 kDa ones in the mitochondrial-enriched fraction (Santama et al., 2004) 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) The author concluded from these observation that the interaction of 120 kDa kinectin with kinesin influence mitochondrial dynamics

Besides ER and mitochondrial, kinectin was also found on isolated phagosomes (Blocker et al., 1997) The function-blocking kinectin antibodies have the capacity to

inhibit phagosome motility in both directions in vitro In addition, kinectin was reported

to associates with melanosomes in human melanocytes under immunoelectron

microscopy (Vancoillie et al., 2000) However, the double immuno-staining shows kinesin and kinectin only co-localize with melanosomes in the perinuclear area, which suggest that kinesin-kinectin interaction based motor transport is involved in some stage

of melanosome movement

A recent study showed that the overexpression of kinesin binding fragments of kinectin in COS7 cells inhibited the kinesin-mediated redistribution of lysosomes to the cell periphery upon acetate treatment, which indicates a role of kinectin in the motility of lysosomes (Ong et al., 2000) In the same study, it was also reported that kinectin can bind to the neuronal kinesin heavy chain subunits, KIF5A and KIF5C, as well as the ubiquitously expressed kinesin heavy chain KIF5B The interaction was proved by co-

immunoprecipitation, yeast two-hybrid assay and direct in vitro binding assay These

results together with another recent report, which revealed distinct labeling of the like processes of differentiated PC12 cell with pronounced immunoreactivity concentrating at the tips using anti-Insert2 or anti-Insert3 kinectin antibody (Santama et al., 2004), highlight the potential roles of kinectin in kinesin-dependent neuronal transport

neurite-However, kinectin gene knockout by homologous recombination results in kinectin null mice viable and fertile (Plitz and Pfeffer, 2001) No gross abnormalities were observed up to one year of age The steady-state distribution and movement of ER, mitochondria and lysosomes was unaffected, and phagocytes internalized and cleared bacteria normally These data seem to contradict the previous studies, especially the RNAi results from Santama et al (2004) It is possible that kinectin functions were compensated by other molecules in the knock-out mouse model just like many other

essential proteins (reviewed by Pearson, 2002) However, in in vitro studies the

compensatory mechanisms might have not worked in the cell type used The functional

1.2.5 Other roles of kinectin and its clinical implication

Besides the interaction with kinesin, the evidence that kinectin functions in some other specialized pathways is emerging Hotta et al (1996) found the interaction of kinectin with Rho family GTPases during a yeast two-hybrid screen for RhoA-interacting proteins The Rho family consists of the Rho, Rac and Cdc42 subfamilies in mammals and has been implicated in regulating various cell functions, such as cell motility, cytokinesis and smooth muscle contraction in various cell types (Hall, 1994) 630-935 residues of human kinectin can interact with GTP-bound forms but not GDP-bound forms

of RhoA, Rac1 and Cdc42 (Hotta et al., 1996) Another distinct region of kinectin, residues 1053-1327, was isolated in another screen, thus suggesting kinectin contains multiple Rho-binding elements (Alberts et al., 1998) The biological significance of these interactions, however, remains unclear

GTP-bound form RhoG was also found to interact in vivo with the central part of

kinectin, which is distinct from previous Rho family binding sites (Vignal et al., 2001) RhoG is another Rho family protein that can activate both Rac1 and Cdc42 RhoG, kinectin and kinesin show highly similar subcellular distribution, mainly in ER but also in lysosomes In addition, the endogenous kinectin distribution was found to depend on RhoG activity It’s possible that RhoG acts on kinectin to facilitate kinesin and microtubule-dependent transport, leading to the delivery of products to the cell periphery and activating pathways controlled by Rac1 and Cdc42

The translation elongation factor-1 δ (EF-1δ) is another interesting interaction partner of kinectin (Ong et al., 2003) Their interaction was shown by yeast two-hybrid

assay and a number of in vitro and in vivo studies Overexpression of the kinectin

fragment containing EF-1δ binding site disrupted the intracellular localization of EF-1δ proteins Since elongation factors exist as a complex consisting EF-1α β γ δ subunits, one

could speculate that kinectin is responsible for anchoring the whole complex to ER to facilitate the synthesis of membranous protein

In another study, kinectin is reported to be remarkably enriched in the based adhesion complexes (IAC) induced by fibronectin-coated beads (Tran et al., 2002) Two other ER-resident proteins, RAP and calreticulin were also found to be clustered at IAC, whereas kinesin was not Therefore kinectin accumulation at IAC seems to be unrelated to the conventional kinectin-kinesin dependent motility but may contribute to some other specific pathways

integrin-Recently, kinectin was found to associate with a number of diseases Kinectin is a frequently found antigen in human hepatocellular carcinoma (HCC) patients (Wang et al., 2004) From these patients, eight theoretical forms of kinectin isoforms were identified by serological analysis of recombinant cDNA expression library and RT-PCR These isoforms contain four putative inserts and only two of them are located within the open reading frame which correspond to Int2 and Int4 The variants containing Int2 were overexpressed in tumor tissues and this alteration may contribute to tumor cell proliferation, motility or anti-apoptotic ability In addition, kinectin may also associate with aplastic anemia (AA), a T cell-mediate autoimmune disease The pathogenic immuno response in AA patients includes both antigen-specific T-cell and autoantibody production Antibodies against kinectin have been frequently detected in patients with

AA but not in healthy people (Hirano et al., 2003; Hirano et al., 2005) The epitope mapping of immunoglobulin of autoantibodies against kinectin revealed that several epitopes were shared by different AA patients These results suggest that kinectin may be the candidate autoantigen involved in the pathophysiology of AA

1.3 ER dynamics

1.3.1 The establishment and maintenance of organelle positions inside cells

Both microtubule and actin filament are responsible for establishing and maintaining organelle positions in cells, but microtubule is a main player It has been shown that ER (Terasaki et al., 1986), Golgi elements (Cooper et al., 1988), early endosomes (Hopkins et al., 1990) and lysosomes (Swanson et al., 1987) form extended tubular processes through moving along microtubules Golgi membranes are normally localized at the minus ends of microtubules in the perinuclear region In the presence of brefeldin A, a drug interfering with protein transport, extensive Golgi tubules are formed, extending along microtubules towards the plus ends (Lippincott-Schwartz et al., 1991) Late endosomes and lysosomes redistribute towards the plus ends of microtubules upon cytosolic acidification, whereas alkalinization causes a shift in distribution towards the minus ends (Parton et al., 1991; Heuser, 1989)

Another important player responsible for the position of organelle inside cells is the family of motor proteins Through moving along microtubules, motor proteins contribute to the positioning of organelles at particular locations within cells and the maintaining of spatial distribution among the organelles (Land and Allan, 1998) Most organelles have the capacity to bind to both plus and minus end-directed motors which support the concept that regulated motor complex consisting of motor proteins and shared activator molecules may control the direction of organelle movement (Schroer and Sheetz, 1991; Sheetz and Yu, 1996)

1.3.2 Microtubule-dependent ER dynamics

The ER is an extensive network emanating from the outer leaflet of the nuclear envelope and spreading throughout the cytosol During interphase, it is a polygonal

network of interconnected tubules and cisternae In spite of its continuous appearance, the

ER is organized into functionally and morphologically distinct domains The smooth ER

is involved in the synthesis of lipids and membrane proteins; the 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) The extensive distribution maximizes the surface area of ER which facilitates its various roles such as calcium regulation, lipid synthesis and translocation of newly synthesized proteins into the ER lumen for folding and other modifications In addition, the fine structure, spatial distribution and abundance of ER vary among cell types, which may reflect specialized functional requirements

ER undergoes constant rearrangement of it fine structure in virtually all cell types, which is believed to be crucial to the maintenance of its characteristic structure Single

ER tubules extend into areas of cellular expansion, such as leading edges of migrating cells and growth cones of neurons, to establishing a reticular ER network (Dailey and Bridgeman, 1989) Retraction of ER towards the cell center also occurs frequently (Terasaki and Reese, 1994) Thus the outward and inward movements generate a balance

of opposing forces that control the distribution of ER

The dynamics of ER network depends on microtubules (Terasaki et al., 1986; Lee and Chen, 1988) Numerous studies implicate microtubules as tracks for ER extension and ER motility in interphase cells The good correlation between the distribution of ER cisternae and microtubules has been observed by electron microscopy in different animal cells (Masurosky et al., 1981; Tokunaga et al., 1983) Microtubule-based ER dynamics was studied with time-lapse microscopy and appeared to be based on three different mechanisms (Waterman-Storer and Salmon, 1998) Firstly, new ER tubules can be pulled

tubules may be dragged along by the tips of polymerizing microtubules Finally, ER tubules may associate with the sides of microtubules, via motor proteins, as they slide along other microtubules Each of these mechanisms can lead to tubule extension and, when tubules intersect, they fuse and create three-way junctions Interestingly, microtubules-dependent motility exclusively controls the rapid extension of ER to the cell periphery, whereas a slower movement of ER tubules towards the cell center is independent of microtubules (Terasaki and Reese, 1994) Indeed, when microtubules are depolymerized by drugs ER retracts toward the cell center, and it is capable of re-tending outwards along newly formed microtubules when the drugs are washed out (Lee et al., 1989)

A variety of studies revealed that the plus end-directed motor, kinesin, is implicated in the ER expansion toward the cell periphery along microtubules For instance, immunofluorescence study using anti-kinesin antibodies generated a punctuated staining pattern closely associated with ER membranes (Hollenbeck, 1989) The antibodies to kinectin, a putative kinesin receptor, gave a reticular staining pattern characteristic of ER in several cell types (Toyoshima et al., 1992) Further more, depleting kinesin heavy chain (KHC) in astrocytes using antisense oligonucleotides or kinectin in HeLa cells using RNAi caused the ER to retract away from the cell periphery and to be incapable of centrifugal extension (Feiguin et al., 1994; Santama et al., 2004), which is consistent with the inhibition of ER motility by anti-kinesin or anti-kinectin

antibodies in in vitro assays (Land and Allan, 1999; Kumar et al., 1995) There are contradicting evidences of the roles of cytoplasmic dynein in ER dynamics in vivo (Lane

and Allan, 1999)

Fig 1.3: The current model for the maintenance of ER dynamics in mammalian cells

Kinesin mediated transport along microtubules is responsible for ER extension to cell periphery ER contraction to cell centre might depend on dynein or other motor proteins

of actin filaments

1.3.3 Actin in the ER movement

In contract to animal cells, plant and budding yeast cells use actin as tracks for ER movement In plant cells, ultrastructural studies have shown the close association between

ER tubules and actin bundles, and video microscopy revealed the sliding of ER tubules along stationary actin bundles (Lichtscheidl et al., 1990; Kachar and Reese, 1988) In yeast, although actin filaments do not co-align with most ER tubules, their disruption results in a rapid and dramatic decrease in ER dynamics (Prinz et al., 2000) Recently, ER tubules were also reported to move along actin filaments in animal cells, especially in regions devoid of microtubules In locust photoreceptor cells, smooth ER was showed to move on actin filaments from the cell center to microvilli (Stürmer et al., 1995) In Purkinje cells lacking myosin Va, the absence of smooth ER in dendritic spines also suggested the role of actin in transporting ER (Bridgman, 1999) Both the photoreceptor microvilli region and the Purkinje dendritic spines contain abundant actin cytoskeleton but lack microtubules In addition, even in the regions where microtubules are known to control ER extension, the retrograde movement of ER towards cell center is perturbed when actin assembly or myosin motor is disrupted (Terasaki and Reese, 1994; Waterman-

affected in these cells These observations indicate that a complex combination of different cytoskeletons and opposing motors might contribute to the organization of the

ER and its dynamics in animal cells

1.4 Cell migration

1.4.1 The four-step concept of cell migration

Cell migration involves at least four coordinated functional elements, namely extension of the cell membrane, attachment to the substrate at the leading edge, forward flow of cytosol, and detachment of the rear end This is the four-step concept of cell migration

In the initial step, ruffling of filopodia and protrusion of lamellipodia is accompanied by the polymerization and subsequent cross-linking of actin filaments, which results in a pushing action to extend the membrane (Cortese et al., 1989) Although initial membrane protrusion is independent of integrins or the extracellular matrix (ECM) (Zhelev et al., 2004) and does not guide the cell body translocation (Condeelis, 1993), attachment of the protruding pseudopodium to the substratum at the leading edge stabilizes the pseudopodium and provides positive feedback signals to maintain actin polymerization toward specific direction At the same time, pseudopodia that do not establish adhesive contacts to the ECM retract back rapidly into the cell body (Zhelev et al., 2004.)

Adhesion signals at the leading edge are provided by chemokines, ECM or growth factors such as vascular endothelial growth factor (VEGF), epidermal growth factor (EGF), platelet-derived growth factor (PDGF) and insulin-like growth factor (Ferrara et al., 2003; Brooks, 1997) Transduction of these signals into the cell body is mediated by multiple receptor families, including the heterodimeric integrins, receptor tyrosine kinases

and phophatases, immunoglobulin receptors, and cell-surface proteoglycans (Huttenlocher et al., 1995; Firtel and Chung, 2000) Most of those receptors are trans-membrane molecules containing both extracellular and intracellular domains The extracellular domain binds ECM or growth factors and the intracellular domain interacts with cytoskeletal components such as talin and α-actinin, and signalling molecules such

as FAK and integrin-linked kinase (Yamada and Geiger, 1997) Take the integrin receptor

as an example, the interaction between ECM and integrin triggers integrin dimerization which results in the assembly of focal adhesion and the recruitment of intracellular signalling and adapter proteins (Burridge et al., 1997; Jockusch et al., 1995) The focal adhesions are mainly composed of densely packed transmembrane integrin receptors attached to extracellular substrate and intracellular proteins linking the cytoskeleton to matrix The assembly of focal adhesions leads to phosphorylation of protein on tyrosine residues and rearrangement of the cytoskeleton (Yamada and Geiger, 1997; Clark and Brugge, 1995; Longhurst and Jennings, 1998) The signalling pathways triggered include the MAP kinase cascade, protein kinase C and Rho proteins Different signal transduction cascades resulting from these intracellular signals can change the affinity of integrin for their extracellular ligands and further influence cell attachment and migration

After cells attach to the substrate at the leading edge, a gradient of both binding and traction forces is established from the front to the rear end, leading to its forward movement Myosin II is involved in producing contractile force to pull cells forward by interacting and linking adhesion complexes with actin filaments (Heidmann and Buxbaum, 1998) Besides cell displacement, cell contraction also affects the strength of focal adhesions Previous research showed that blocking myosin light chain kinase activity or myosin-actin interactions lead to the inhibition of Rho and impairment of cell

adhesions (Burridge et al., 1997) This result is in consistent with the concept that the strengthening of integrin-cytoskeleton interactions is controlled by biomechanical traction, substrate rigidity and the force provided by cell contraction (Sheetz et al., 1998)

Focal contacts are resolved at the rear end after cell body contraction ligand interactions are weakened to a low-affinity binding state, making it possible for integrin to detach from the substrate (Hughes and Pfaff, 1998) The release and deposition

Integrin-of integrins on the substrate are likely controlled by tyrosine phosphorylation-dependent events (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)

In summery, cell migration is formed by the cycling of these four steps without obvious start and end points In this process, focal adhesions, highly dynamic structures, are specifically important and are required for cell attachment, spreading, as well as forming ‘hot spots’ for cell signalling Varying degrees of focal adhesion assembly and cytoskeleton organization result from the resolution of cell-matrix interactions and determine cellular behavior For example, a slow turn-over 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 seen during cell motility and in many transformed cells (Duband et al., 1986; Friedl and Brocker et al., 1998) In deed, the weakening of focal adhesions is an early step during the onset of migration, and focal contacts found in migrating cells are frequently of smaller size, unstable and incompletely assembled (Friedl and Entschladen et al., 1998) Generally, increased migration dynamics is associated with decreased focal adhesion

assembly and vice versa (Dunlevy et al., 1995; Ilic et al., 1995)

1.4.2 The MT based membrane transport in cell migration

Although actin dynamics has been considered the driving force for cell motility, recent studies have highlighted the role of microtubules in this dynamic process Microtubules contribute to cell motility in a cell type-dependent manner Although disruption of microtubules alters migration in most cell types, migration of fish kerotocytes is not affected by microtubule disassembly (Euteneuer and Schliwa, 1984) and neurophil motility is even increased in the absence of microtubules (Keller et al., 1984) Like actin, microtubules polarize during cell migration with different dynamics at the protruding front and the retracting end The most striking polarization phenomenon of microtubules in many migrating cells is the orientation of microtubule organization center (MTOC) toward the direction of migration As a result of this, microtubules themselves, especially stabilized detyrosinated microtubules, are polarized and tend to be aligned along the cell migration axis with their plus ends facing the leading edge (Gundersen and Bulinski, 1988) In addition to their polarized organization as a whole, microtubule polymerization dynamics is polarized in a migrating cell During cell migration, microtubule plus ends are often found close to the leading edge even though microtubules

in the lamella are moved backwards by the actin retrograde flow (Yvon and Wadsworth, 2000) Thus we would expect that microtubule growth is biased towards the leading edge Indeed, in protruding lamellipodia, ‘pioneer’ microtubules appear to spend more time growing than those in quiescent cell edges (Wadsworth, 1999) The molecular mechanism underlying the polarization of microtubules during cell migration is largely unclear It is suggested that Rho GTPases, the pivotal regulators of actin dynamics in cell migration, might also influence the organization and dynamics of microtubules (Watanabe et al., 2005)

Polarized microtubules could serve as tracks for anterograde transport of membranes and organelles towards the leading edge of migrating cells This transportation is dispensable for providing building materials for the protruding lamellipodia (Nabi, 1999) In fact, it was originally speculated that the reorientation of MTOC towards the leading edge may contribute to the orientation of secretory apparatus, and this hypothesis is supported by the observation of the secretion which preferentially polarizes towards the leading edge of migrating fibroblasts (Bergmann et al., 1983; Hopkins et al., 1994) The requirement of microtubule-based transport during cell migration was also demonstrated by microinjection of kinesin-specific antibodies, which inhibited cell locomotion in a way similar to microtubule disruption (Rodionov et al., 1993)

In addition to membranes and organelles, polarized microtubules also transport signalling molecules to the leading edge, where they locally modulate the activity of Rho GTPases or other signalling proteins Recently, p190RhoGEF, a RhoA-specific exchange factor, was reported to partly colocalize with microtubules in cells and might interact with the microtubule motor kinesin through JIP scaffolding protein (van Horck et al., 2001; Verhey et al., 2001) In addition, microtubule depolymerisation could induce the release

of a microtubule-bound RhoA activator from the microtubule lattice (Enomoto, 1996) Thus it’s hypothesized that kinesin-dependent transport together with regional microtubule depolymerization-dependent release of p190RhoGEF might result in a local activation of RhoA In another study, both RhoG and its exchange factor TrioGFE1 were reported to lose their localization at the cell periphery upon microtubule depolymerisation (reviewed in Wittmann and Waterman-Storer, 2001), which also suggest a microtubule-dependent transport process This could be mediated by kinesin, since RhoG was showed

to interact with kinectin (Vignal et al., 2001) In addition, both kinesin and kinectin

appear to be required for RhoG induced modification of the actin filaments (Vignal et al., 2001)

However, if transport of membrane compartments is the only role of microtubules

in cell motility, it would be expected that partial inhibition of microtubule assembly dynamics, which do not affect the overall organization and thus does not inhibit microtubule-dependent transport, should not alter cell motility However, even low concentration of nocodazole significantly reduces the speed of protrusion of fibroblasts into a wound (Liao et al., 1995) Clearly other mechanisms may contribute to the microtubule-dependent migration in which growing ‘pioneer’ microtubules could directly promote lamellipodial protrusion and thus be required for selective stabilization of one particular leading edge to maintain the direction of cell movement Consistent with this hypothesis, membrane ruffling and protrusive activity are reduced upon microtubule depolymerization, whereas microtubule regrowth after the removal of the microtubule-depolymerising drug induces the formation of ruffling lamellipodia (Waterman-Storer et al., 1999) Besides regulating protrusion, microtubules may also be responsible for the local regulation of contraction Indeed microtubule depolymerisation also induces increased contractility through the formation of focal adhesions and actin stress fibers (Danowski, 1989) By imaging microtubules and focal adhesions simultaneously in living cells, it has been showed that focal adhesions targeted by microtubule plus ends undergo dynamic instability and that repeated targeting leads to focal adhesion disassembly (Kaverina et al., 1998; Kaverina et al., 1999) The consistent targeting of microtubule was further confirmed by images from total internal reflection fluorescence microscopy (TIRFM) which demonstrated the plus ends of growing microtubules track close to the dorsal surface on a nanometer scale (Krylyshkina et al., 2003) Thus adhesions in the

tailing part of migrating cell are released by microtubule targeting to allow detachment from the substrate

How microtubules fulfil their multiple roles in cell migration is still unclear Again Rho GTPases might play pivotal roles in these processes By regulating the activity

of these signalling molecules, microtubules can cooperate with actin filament to promote the directional movement (Wittmann and Waterman-Storer, 2001)

1.5 Cell division

1.5.1 Membrane partitioning during cell division

Dividing cells are biosynthetically quiescent Very little new proteins or RNA and

no DNA are synthesized (Prescott and Bender, 1962); transport of vesicles between organelles ceases (Warren, 1989); and the cell does not respond to external stimuli (Volpi and Berlin, 1988) To segregate the chromosomes and partition other cellular components are the sole tasks of dividing cells The research of organelles partitioning progressed much slowly compared with the better studied area of chromosome segregation probably due to the complexity of the process itself and the variable partitioning mechanisms adopted by different organelles A number of organelles fragment during mitosis and reassembly occurs when cytokinesis begins to separate the two daughter cells Fragmentation and vesiculation is thought to aid the partitioning of organelles during mitosis However this is not a general rule and many organelles do not break down at all

or only break down to what can be regarded as units of growth and division

Partition of nuclear envelope (NE) is one of the most well studied examples of membrane partitioning during cell division NE exists as a single copy throughout interphase and breaks down during mitosis in mammalian cells The breakdown of NE is considered the hallmark of mitosis in most cells The mechanism dividing NE between

two daughter cells involves fragmentation, dispersal and subsequent reassembly in each daughter cell (Helper and Wolniak, 1984) Electron microscopy indicates this is a progressive process Fragmentation of NE begins at prometaphase after condensation of chromatin Initially large smooth flattened cisternae appear surrounding the chromatin, and as fragmentation continues some fragments line up with the mitotic spindle while others move to the spindle poles (Tamaki and Yamashina, 1991) Fragmentation is accompanied by the disassembly of the nuclear pores Both membrane-bound and soluble components of the pore complex were observed under immunofluorescence microscopy during mitosis (Snow et al., 1987) The dismantling of the nuclear pores leaves holes in

NE and the holes grow in size while NE fragments grow smaller until they are eventually indistinguishable from other membrane fragments (Helper and Wolniak, 1984) Fragmentation of NE is also accompanied by breakdown of the nuclear lamina (Gerace et al., 1978), which consists of a meshwork of intermediate filaments underlying NE Reassembly of NE occurs during telophase NE fragments are observed on lateral chromatin surfaces and lateral fusion reconstitutes the NE (Zeligs and Wollman, 1979) In most cases NE bind to chromatin via the nuclear lamina directly or indirectly, and often

in association with the nuclear pores ER also breaks down during mitosis but to a variable extent in different cell types Fragmentation is often thought to be minimal in cultured cells (Koch et al., 1987), whereas cells in tissues undergo almost complete breakdown, yielding vesicles of diameter from 0.3 to 1 µm (Zeligs and Wollman, 1979; Tamaki and Yamashina, 1991)

For those organelles existing in multiple copies during interphase, their distribution is often unchanged in mitosis Although in some cases, lysosomes, secretory granules and endosomes were observed to cluster at the mitotic spindle poles (Zeligs and

and Thyberg, 1990) The general observation is that a large number of copies of an organelle would facilitate nearly equal partitioning by a stochastic process

As mentioned before, membrane trafficking between organelles is inhibited during mitosis This inhibition is almost entirely restricted to stopping vesicle-mediated transport between membrane compartments The compartments continue to function if they can Therefore, although the protein synthesis in ER is low because there is little mRNA to translate during mitosis, the synthesis rate can be increased even as high as in interphase cells if mRNA is supplied (Warren et al., 1983)

There are two strategies that ensure partitioning accuracy between daughter cells First, organelles may exploit the properties of mitotic spindle, which is used primarily to segregate chromosomes Alternatively, partitioning could be a stochastic process, the accuracy of which depends on the copy number of organelles or organelle fragments during mitosis

1.5.2 ER partitioning during mitosis and the role of microtubules

It has been believed for long time that ER, together with NE, fragments to form numerous discrete vesicles during mitosis to facilitate the partitioning of ER between daughter cells, a view termed the vesiculation model (Warren, 1993) However, recent studies have provided evidence for an alternative model, in which the ER network appears to remain its integrity during mitosis and show no significant fragmentation (Waterman-Storer et al., 1993; Ioshii et al., 1995; Ellenberg et al., 1997; Terasaki, 2000)

The strongest supports for maintenance of ER continuity come from FLIP (Fluorescence Loss in Photobleaching) and FRAP (Fluorescence Recovery after Photobleaching) experiments demonstrating that ER markers retain inter-phase patterns

of motility during mitosis (Ellenberg et al., 1997) It also reported that ER networks can

be clearly visualized during cell division under both light and electron microscopy

(Terasaki et al., 1986; Yang et al., 1997; Terasaki, 2000) In addition, some GFP-tagged inner nuclear membrane proteins or nuclear pore proteins became highly mobile throughout the ER network during mitosis (Ellenberg et al., 1997; Yang et al., 1997), which prompted the proposal that NE is absorbed into the ER network and its components are partitioned together with ER membranes However the details of the ER rearrangement in mitosis have yet to be elucidated The accumulation of peripheral ER membranes at the mitotic poles in embryonic systems indicates that microtubules and microtubule motors are likely to be involved (Terasaki, 2000; Bobinnec et al., 2003)

The role of microtubules in membrane partitioning has been questioned since the primary function of microtubules is to form the mitotic spindle which segregates the

chromosomes In an early in vitro study using Xenopus egg extracts, it was shown that the

cell cycle state of the extracts determines the extent of microtubule-based membrane movement (Allan and Vale, 1991) In contract to interphase extract, metaphase extracts inhibit both the plus end- and minus end- directed movement of vesicles along microtubules as well as the formation of tubular membrane networks These observations suggest that the dispersal of vesicles and membrane fragments during the early stages of mitosis may take place without the aid of microtubules However, in another study using live cells labeled with the vital dye DiOC6(3), X-Z series images from the confocal microscopy revealed an intact ER structure with well-regulated dynamics and the strong association of ER membrane with microtubules throughout the mitosis (Waterman-Storer

et al., 1993) In addition, taxol treatment induced a dense and extensive collection of small vesicles at the spindle poles, and nocodazole treatment resulted in a distinct loss of organization of the membranous components associated with the spindles Together these results suggest that microtubules are responsible to organize the membrane distribution in

mitotic cells However this organization may vary in different cell types depending on the quantity of microtubules within the spindle

Whether other cytoskeletal elements are used instead during the dispersal of ER or other organelles in mitosis remains unclear Studies in budding yeast indicate that ER inheritance involves a type V myosin-mediated polarized transport of cytoplasmic ER tubules to newly formed buds along actin filaments (reviewed in Du et al., 2004) The ER tubules are then anchored to the bud tip and propagated along the cortex of the bud to yield the cortical ER structure in daughter cells It is not clear yet to what extent components controlling ER distribution in yeast might be conserved in animal cells

Another role of microtubules during mitosis, which is less controversial, is to serve as tracks for membrane trafficking during cytokinesis Membrane traffic is required

to delivery membranes to the surface of diving cells during cleavage furrow ingression, and thus to provide the increased surface area necessary to form two new daughter cells Once the cleavage furrow has completed its ingression, two daughter cells remain connected by a narrow intracellular bridge (the mid-body) and membrane dynamics is also required here to separate and seal the newly formed cells (Matheson et al., 2005) It has been reported that during cytokinesis vesicles derived from recycling endosomes traffic to the furrow and mid-body along microtubules (reviewed in Matheson et al.,

2005) In addition, comprehensive RNAi analysis in Drosophila S2 cell line revealed that

microtubules-based motor proteins contribute to cytokinesis in animal cells (Goshima and Vale, 2003)

1.6 The purpose and rationale of the thesis work

The overall objective of this thesis work is to understand the roles of different kinectin isoforms in ER dynamics Multiple cellular and molecular imaging techniques were coupled with basic biology research approaches in this work to achieve this purpose

A series of EGFP-tagged kinectin isoforms was constructed and their subcellular localizations were studied under a confocal microscopy Both the static and the dynamic distribution of these chimera proteins relative to ER dynamics were investigated in fixed and live DsRed-ER HeLa cells respectively Our preliminary results revealed the importance of Int4 in kinesin-dependent ER extension along microtubules Thus this domain was deleted in follow-up experiments and the effect of Int4 deletion on ER extension as well as cell migration was studied These works extended our understanding

of kinectin functions and would lay the foundation for future studies of the mechanism of microtubule-dependent ER dynamics