The Role of

Phosphatidylinositol Transfer Proteins

in Phosphoinositide Signalling

submitted by

Siow Khoon Tan

University College London

as part of the requirements for the degree of

Doctor of Philosophy

University of London ProQuest Number: U644336

All rights reserved

INFORMATION TO ALL USERS The quality of this reproduction is dependent upon the quality of the copy submitted.

In the unlikely event that the author did not send a complete manuscript and there are missing pages, these will be noted. Also, if material had to be removed, a note will indicate the deletion. uest.

ProQuest U644336

Published by ProQuest LLC(2016). Copyright of the Dissertation is held by the Author.

All rights reserved. This work is protected against unauthorized copying under Title 17, United States Code. Microform Edition © ProQuest LLC.

ProQuest LLC 789 East Eisenhower Parkway P.O. Box 1346 Ann Arbor, Ml 48106-1346 Abstract

The classical model of the phosphoinositide (PI) signalling pathway postulates that the substrate lipid, phosphatidylinositol (Ptdlns), is sequentially phosphorylated by PI kinases to form Ptdlns 4,5-bisphosphate (Ptdlns? 2), before being cleaved by phospholipase C (PLC) to generate two second messengers, inositol 1,4,5-trisphosphate and diacylglycerol (DAG). Ptdlns is thought to be embedded in the membrane lipid bilayer throughout this process. However, several reports have strongly suggested that a Ptdlns transfer protein (PtdlnsTP) is not only essential but may also be a co-factor in this pathway. The experiments described in this thesis address a key aspect of this suggestion.

The specific hypothesis to be tested is that the mammalian PtdlnsTPs PITPa and PITPp bind one or more Ptdlns kinases in order to present bound Ptdlns for phosphorylation.

The results presented in this thesis provide evidence that Ptdlns bound to PITP is a direct substrate for PI kinases. Recombinant PITP was found to co-purify with PI kinases, which recognised and phosphorylated PITP-bound Ptdlns to produce Ptdlns 4-phosphate and PtdInsP 2 - Studies of a mutant PITP provided additional evidence for the above conclusion and also revealed that the co-purification of PITP and PI kinases was not dependent on the ability of PITP to bind Ptdlns. Liquid chromatographic analyses showed that PITP and PI kinases behaved as a complex. Furthermore, PLC and DAG kinase activities were also detected in the PITP-PI kinase complex. These studies provide strong support for the hypothesis that PITP is not only a component of a multi-enzyme signalling complex but that it also presents bound Ptdlns sequentially to the enzymes in a signalling cascade.

Comparative studies were carried out on the two (alpha and beta) isoforms of PITP. Despite their different subcellular distribution, both isoforms were able to (i) transfer Ptdlns in vitro, (ii) reconstitute PLC signalling in permeabilised cells, and (iii) bind and present substrate to PI kinases in vitro. The biological significance of these results is discussed in the light of apparently complementary and conflicting studies by other groups. Acknowledgement

Firstly, I would like to express my gratitude and sincere thanks to my thesis supervisor, Dr. Justin Hsuan, for the opportunity to work on this exciting project in his laboratory. He is a supportive and understanding mentor, who readily shares his knowledge and know­ how and is generous with his ideas, time and resources. I had a truly memorable time working with him.

I would like to thank present and former members of the Hsuan laboratory, especially Shane Minogue, Damian Holliman, Maria dos Santos, Yvonne Fullwood, Mark Waugh, Nick Totty and Oanh Truong, for their help, advice, encouragement and generosity in sharing research materials.

I also thank the staff and members of the Ludwig Institute for Cancer Research (LICR) for their help, support, advice, criticism, encouragement, patience, generosity and friendship during my stay at the institute.

Many thanks to Prof. Shamshad Cockcroft (Dept, of Physiology, UCL) and her laboratory members (particularly, Geraint Thomas, Emer Cunningham and Andrew Ball) for their time, help, advice and generosity.

I would also like to thank Drs. Peter Parker and Frank Cooke of the former Imperial Cancer Research Fund for access to his laboratory and for performing the lipid HPLC analyses, respectively.

Last but not least, I would like to thank the LICR for its generous financial support. Table of Contents

Abstract

Acknowledgement 3

List of Tables 9

Abbreviations 10

Chapter 1 Introduction

1.1 Signal Transduction 12 1.2 Receptors 12

1.2.1 Epidermal Growth Factor Receptor 13 1.3 Signalling Pathways 18 1.4 Homology Domains 19 1.4.1 SH2 and Phosphotyrosine-binding Domains 19 1.4.2 SH3 and WW Domains 20 1.4.3 Pleckstrin Homology Domain 21 1.4.4 C2 Domain 22 1.4.5 PDZ Domain 22 1.4.6 FYVE Domain 22 1.4.7 ENTH Domain 23 1.4.8 PX Domain 23 1.5 Phosphoinositide Signalling Pathways 23 1.5.1 Phospholipase C 25 1.5.1.1 PLCÔ 27 1.5.1.2 PLCP 28 1.5.1.3 PLCy 29 1.5.1.4 PECs 33 1.5.2 Functional Roles of Phosphatidylinositol 4,5-bisphosphate 34 1.5.3 Phosphatidylinositol 4-Kinase 37 1.5.3.1 Structures and Characteristics 37 1.5.3.2 Regulations and Functions 42 1.5.4 Phosphatidylinositol Phosphate Kinase 45 1.5.4.1 Structures and Characteristics 45 1.5.4.2 Regulations and Functions 49 1.5.5 Phosphatidylinositol Transfer Protein 54 1.5.5.1 PITP 55 1.5.5.2 RdgB 59 1.5.5.3 Sec 14 61 1.5.5.4 Functions 62 1.6 Compartmentation of Signalling 68

Chapter 2 Materials and Methods 2.1 Molecular Biology 74 2.1.1 DNA Biochemistry 74 2.1.1.1 Polymerase Chain Reaction 74 2.1.1.1.1 Polymerase Chain Reaction Primers 75 2.1.1.2 Bacterial Transformation 77 2.1.1.3 Plasmid Preparation 78 2.1.1.4 DNA Digestion with Restriction Enzymes 79 2.1.1.5 Agarose Gel Electrophoresis of DNA 79 2.1.1.6 DNA Ligation 79 2.1.1.7 Automated DNA Sequencing 80 2.1.2 Protein Biochemistry 80 2.1.2.1 Polyacrylamide Gel Electrophoresis 80 2.1.2.1.1 Sodium Dodecyl Sulphate-PAGE 80 2.1.2.1.2 Native PAGE 80 2.1.2.1.3 Iso-Electric Focusing 81 2.1.2.2 Staining of Protein Gels 81 2.1.2.2.1 Coomassie Staining 81 2.1.2.2.2 Silver Staining 81 2.1.2.3 Semi-dry Electrophoretic Transfer of Protein 82 2.1.2.4 Western Blotting 82 2.1.2.5 Bradford Assay 82 2.1.3 Lipid Biochemistry 83 2.1.3.1 Separation of Inositol Phosphates 83 2.1.3.2 Identification of Inositol Phospholipids 83 2.1.3.2.1 Thin-layer Chromatography 83 2.1.3.2.2 Lipid Déacylation 84

2.2 Cell Culture 84 2.2.1 Insect Cell Culture 84 2.2.1.1 Routine Maintenance 84 2.2.1.2 Generation of Baculovirus Stock 85 2.2.1.2.1 Transfection of Sf9 cells 85 2.2.1.2.2 Plaque Assay 85 2.2.1.2.3 Virus Amplification 86 2.2.2 Mammalian Cell Culture 86 2.2.2.1 Routine Maintenance 86 2.2.2.2 In situ Immunofluorescence and Confocal Microscopy 87 2.3 Protein Expression and Purification 87 2.3.1 Expression System in Bacteria 88 2.3.1.1 Expression of Recombinant Proteins in co/z 88 2.3.1.2 Purification of Recombinant PITP 88 2.3.2 Expression System in Insect Cells 88 2.3.2.1 Expression of Recombinant Proteins in Sf9 Cells 88 2.3.2.2 Purification of PITP 89 2.3.3 Expression System in Mammalian Cells 89 2.3.3.1 Transfection of Cells 89 2.3.3.2 Immuno-affmity Purification of PITP 89 2.4 Assays for PITP, PLC and Lipid Kinases 90 2.4.1 PITP Assays 90 2.4.1.1 /« vzYro Ptdlns Transfer 90 2.4.1.1.1 Preparations of Reagents 90 2.4.1.1.2 Ptdlns Transfer Assay 91 2.4.1.2 Reconstitution of G-protein-regulated PLC Signalling 91 2.4.1.3 Reconstitution of EGF-dependent Phosphoinositide Kinases 92 2.4.2 Ptdlns 4-Kinase Assay 92 2.4.3 PLC Assay 92

2.4.3.1 PLC Assay using Exogenous Ptdlns ? 2 93

2.4.3.2 PLC Assay using Endogenous, Labelled Ptdlns ? 2 93 2.4.4 DAG Kinase Assay 93

Chapter 3 Reconstitution of Polyphosphoinositide Signalling by Both Isoforms of PITP 3.1 Introduction 94 3.2 Cloning and Expression of PITP 95 3.2.1 Cloning 96 3.2.1.1 Cloning Human PITPa 96 6 3.2.1.2 Cloning Rat PITPp 97 3.2.2 Expression and Purification of PITP 98 3.3 Functional Comparison of PITPa and PITPp 104 3.3.1 In vitro Ptdlns Transfer 104 3.3.2 Reconstitution of PI Signalling 106 3.3.2.1 Reconstitution in A431 Cells 106 3.3.2.2 Reconstitution in HL-60 Cells 109 3.4 Discussion 109 ipter 4 Presentation of Phosphatidylinositol to Lipid Kinases am Lipases by Phosphatidylinositol Transfer Proteins 4.1 Introduction 117 4.2 In vitro Interactions between PITP and Ptdlns 4-kinase 117 4.2.1 Interaction with Unpurified Ptdlns 4-kinase 117 4.2.2 Interaction with a Partially Purified Ptdlns 4-kinase 119 4.3 Co-purification of PI Kinases with PITP 120 4.3.1 Co-purification of Ptdlns 4-kinase Activity 120 4.3.2 Co-purification of PtdlnsP 5-kinase Activity 123 4.3.3 Confirmation of Lipid Identity 123 4.3.4 PITP-PI Kinase Interactions are Specific 126 4.4 PITP-bound Ptdlns as a Substrate 126 4.4.1 Phosphorylation by PI Kinases 126 4.4.2 Interactions are Independent of Lipid Type 131 4.5 Co-purification of DAG Kinase 131 4.6 Detection of Phospholipase Activity 133 4.7 Sf9 Cells as a Model of PITP Function 137 4.8 Expression of Recombinant PITPs in Cultured Mammalian Cells 137 4.8.1 Generation of New cDNAs and Preparations of Pure Plasmids 138 4.8.2 Mammalian Cell Transfection 138 4.8.2.1 COS-7 and A431 Cells 138 4.S.2.2 Human Embryonic Kidney Cells 139 4.8.3 Immunoprécipitation and PI Kinase Assays 141 4.8.4 Immunofluorescence Microscopy 144 4.8.4.1 Wildtype PITPa 144 4.8.4.2 Mutant PITPa 146 4.S.4.3 PITPp 146 4.8A 4 A33 148 4.9 Discussion 148

Chapter 5 Chromatographic and Electrophoretic Analyses of PITP and Co-purifying Enzymes 5.1 Introduction 156 5.2 Production of Polyphosphoinositides on PITP 156 5.2.1 Resolution by Gel Filtration Chromatography 156 5.2.2 Resolution by Iso-Electric Focusing 158 5.2.2.1 Analysis of PITP - Before Incubation with ATP 158 5.2.2.1.1 PITPa 158 5.2.2.1.2 PITPp 160 5.2.2.2 Analysis of PITP - Post Incubation with ATP 160 5.2.2.2.1 Western Blots 161 5.2.2.2.2 SDS-PAGE 161 5.2.2.2.3 Lipid Characterization 164 5.2.2.2.4 PI Kinase Assay 168 5.3 Characterization of the Putative Complexes 168 5.3.1 Gel Filtration Chromatography of PITP Preparations 170 5.3.1.1 PITPa 170 5.3.1.2 PITPp 170 5.3.2 Distribution of Co-purified Kinase Activities 173 5.3.2.1 PITPa 173 5.3.2.2 PITPp 175 5.4 Discussion 175

Chapter 6 Discussion

6.1 Proposed Models of PITP Function 184 6.2 Interaction with PI Signalling Enzymes 185 6.3 Viability of PITP-bound Ptdlns as a Substrate 187 6.4 Mechanism of Interaction 189 6.5 Concluding Remarks 192

References 195 List of Tables

Table 1 Apparent Molecular Masses of PITP page 181 In Various Gel Filtration Media Abbreviations

ARF ADP-ribosylation factor CCTase choline-phosphate cytidylyltransferase cDNA complementary DNA CDP cytosine 5’-diphosphate CTR C-terminal region DAG diacylglycerol DAGK diacylglycerol kinase EGF epidermal growth factor EGFR epidermal growth factor receptor

EGTA ethylene glycol-bis(p-aminoethyl ether)-#, A,# % A -tetraacetic acid

ER endoplasmic reticulum FCS foetal calf serum fMLP formylmethionyl-leucyl-phenylalanine G protein GTP phosphodiesterase GAP GTPase-activating protein GEF guanine nucleotide exchange factor GPCR G-protein-coupled receptor HB-EGF heparin-binding epidermal growth factor HEK human embryonic kidney lEF iso-electric focusing InsPs inositol-1,4,5-trisphosphate

IPTG isopropylthio-p-D-galactoside kDa kilodalton LKH lipid kinase homology LKU lipid kinase unique mAChR muscarinic acetylcholine receptor MAPK mitogen-activated protein kinase NGF nerve growth factor ORF open reading frame

PAGE polyacrylamide gel electrophoresis PCR polymerase chain reaction PDGF platelet-derived growth factor PDGFR platelet-derived growth factor receptor PH pleckstrin homology PI phosphoinositide PI3K phosphoinositide 3-kinase

PKC protein kinase C

10 PLA phospholipase A PLC phospholipase C PLD phospholipase D PLTP phospholipid transfer protein PPI polyphosphoinositide PTB phosphotyrosine-binding PtdCho phosphatidylcholine PtdEtn phosphatidylethanolamine Ptdlns phosphatidyl inosito PtdlnsSP phosphatidylinosito -3-phosphate PtdIns3P5K phosphatidylinosito -3-phosphate 5-kinase PtdIns4K phosphatidylinosito 4-kinase PtdIns4P phosphatidylinosito -4-phosphate PtdIns4P5K phosphatidylinosito -4-phosphate 5-kinase PtdlnsSP phosphatidylinosito -5-phosphate PtdIns5P4K phosphatidylinosito -5-phosphate 4-kinase PtdlnsP phosphatidylinosito -4-phosphate or phosphatidylinosito monophosphate (wherever ambiguity may occur)

Ptdlnsf] phosphatidylinosito -4,5-bisphosphate or phosphatidylinosito bisphosphate (wherever ambiguity may occur) PtdlnsPs phosphatidylinosito -3,4,5-trisphosphate PtdlnsPK phosphatidylinosito phosphate kinase

PtdlnsTP phosphatidylinosito transfer protein PtdOH phosphatidic acid PTK protein tyrosine kinase pTyr phosphotyrosine RTK receptor tyrosine kinase SDS sodium dodecyl sulphate SH2 src-homology-2 SH3 src-homology-3 SLO streptolysin O TON trans-Go\g\ network TXlOO Triton X -100 Tyr tyrosine WT wortmannin

11 Chapter 1

Introduction

1.1 Signal Transduction

Multicellular organisms coordinate their many physiological activities through complex bioehemical systems. Chemical messages are carried throughout the organism in the form of growth factors, hormones and cytokines, and regulate or control such diverse cellular functions as proliferation, migration, differentiation, metabolism and apoptosis. These messages bind to specific receptors in target cells and trigger characteristic cascades of chemical reactions within the cells to bring about the desired response. Each cellular response to a message is defined and restricted by the number and type of cellular receptors and the combination of intracellular signalling molecules. Thus different cell types will respond to each stimulus differently.

1.2 Receptors

A cell is able to recognise an extracellular chemical messenger only if it expresses cellular receptors that bind the chemical. Most receptors are found on the surface of cells, but the structure of different receptors and the mechanisms by which they transmit messages across the plasma membrane vary widely.

For example, a variety of ligand-gated ion channels are concentrated at nerve synapses and, upon binding neurotransmitters, change their conformation to permit or inhibit ion flux through its intrinsic ion channel (Bertaccini and Trudell, 2001; Cockcroft et a l, 1990). This changes the ion permeability of the plasma membrane and thereby the polarisation of the postsynaptic cell.

Another category of surface receptor contains a diverse range of heterotrimerie GTP phosphodiesterase (GTPase or G protein)-coupled receptors (GPCRs) (Pierce et al., 2002), which interact with and control the activities of cytosolic and/or neighbouring membrane proteins (including ion channels), through intermediary G proteins (Hamm and Gilchrist, 1996). These G proteins relay messages by acting as molecular switches: they bind GTP (“ON” state) and then hydrolyse GTP to produce GDP (“OFF” state). Activation of GPCRs triggers associated G proteins to bind GTP in place of GDP, and bring about dissociation into Ga(GTP) and GPy subunits, both of which can interact with and alter the activities of target proteins in the cell. The eventual hydrolysis of the GTP to

12 GDP by the a-subunit will lead to re-association of the Ga and G(3y subunits in the inactive state and terminate this message relay. GPCRs are integral membrane proteins with seven trans-membrane segments and transduce signals in response to neurotransmitters and many hormones, as well as odourants and light.

As a final example of cell surface receptors, the protein kinase linked receptors comprise receptor serine/threonine kinases, receptor tyrosine kinases (RTKs), and non-enzymatic receptors associated with cytosolic kinases. When activated, they phosphorylate distinctive sets of intracellular proteins on either tyrosine (Tyr) or serine and threonine residues. RTKs share a common molecular topology: an extracellular, glycosylated, ligand-binding domain, a single hydrophobic transmembrane region, and a cytoplasmic region containing a protein tyrosine kinase (PTK) catalytic domain (Ullrich and Schlessinger, 1990).

1.2.1 Epidermal Growth Factor Receptor

The epidermal growth factor receptor (EGFR) is the first RTK to be purified and cloned, and comes from the ErbB receptor family, which derived its name from the avian erythroblastosis retroviral oncogenes (v-erbB). The family has four closely-related receptors: the EGFR/c-erbBl, HER2/neu/c-erbB2, HER3/c-erbB3 and HER4/c-erbB4 (Olayioye et al., 2000). The EGFR is found in increased abundance and modified forms in many human cancer cells, and has been extensively studied to help in the understanding of human oncology (Yarden and Sliwkowski, 2001). In addition to epidermal growth factor (EGF), EGFR also binds transforming growth factor alpha, vaccinia virus growth factor, heparin-binding EGF (HB-EGF), betacellulin and amphiregulin (Barnard et al., 1994). It appears that most of these ligands share a 50- amino-acid homologous domain with a defined three-loop secondary structure (Riese and Stem, 1998). This domain is both required and sufficient for binding and activating ErbB family receptors.

The EGFR is a 170-kDa glycoprotein with a single transmembrane domain. The extracellular moiety has two cysteine-rich domains while the cytoplasmic half consists of the juxtamembrane region and the PTK, hinge and autophosphorylation domains (Fig. 1.1). The mechanism of EGFR activation by EGF is still not fully understood. A widely held model proposes that binding of EGF to the extracellular domains induces conformational changes and causes the receptors to dimerise, forming a tetramer complex of two ligands and two receptors (Greenfield et al., 1989; Ullrich and Schlessinger, 13 Plasma membrane

CPtdIns4K) CPtdInsP5K) Juxtamembrane region

Kinase domain

( ü ) <1^

967/971 992 (PLCy^ 1002 © Hinge domain

1068

Autophosphorylation ^ domain

1148

1173

Fig. 1.1 Intracellular Domain of the EGFR A schematic representation of the four component segments of the EGFR intracellular domain. Serine (S), threonine (T) and tyrosine (Y) phosphorylation sites are indicated. The four tyrosine phosphorylation sites outside the autophosphorylation domain are phosphorylated by Src. Binding proteins ascribed to the various segments or phosphorylated sites are selectively shown to the right (Hsuan and Tan, 1997; Olayioye et aL, 2000, and references therein). 14 1990). Subsequent work further suggested that bivalent ligands, with two receptor binding sites, bind to one receptor with a high-affmity binding site and to another receptor with a second, low-affinity site (Tzahar et aL, 1997).

This issue is further complicated by the ability of the four EGFR family members to heterodimerise and the different signalling characteristics induced by each ligand and dimer (Lenferink et aL, 1998; Olayioye et aL, 2000; Riese and Stem, 1998). The various heterodimers have distinct sets of autophosphorylation sites, which may produce different cellular responses (Riese and Stem, 1998). For example, HB-EGF activates the EGFR to induce cell proliferation and chemotaxis whereas HER4 activation by the same ligand induces only chemotaxis (Elenius et aL, 1997). Even with the same heterodimer, different ligands induce distinct signalling properties. In an ErbBl-ErbB4 heterodimer, EGF treatment results in more tyrosine phosphorylation and less threonine phosphorylation of ErbBl than NGF treatment (Olayioye et aL, 1998).

More recently, an alternative model postulates that a major driving force for receptor dimérisation is the interactions between the transmembrane domains (Tanner and Kyte, 1999). Ligand binding will induce conformational changes in the extracellular domains and hence relieve the steric hindrance to transmembrane domain-mediated receptor dimérisation. This proposal is supported by a recent study, which shows that the single transmembrane domains of ErbB receptors are able to self-associate strongly in the absence of extracellular domains (Mendrola et aL, 2002). Such transmembrane domain- driven oligomerisation has also been observed in the erythropoietin receptor (Constantinescu et aL, 2001; Kubatzky et aL, 2001).

In addition to direct receptor stimulation by extracellular ligands, it was recently proposed that activated ErbB receptors can dissociate from dimers and subsequently activate other ErbB receptors in a rapid, ligand-independent, lateral propagation of receptor activation in the plasma membrane (Verveer et aL, 2000). This would suggest that the dimérisation of activated ErbB receptors is a transient rather than a stable feature.

Furthermore, the EGFR can be activated by other stimuli, such as other membrane receptors (e.g., cytokine and adhesion receptors), membrane depolarization (e.g., the influx of extracellular Ca^^ into PCI2 cells), and environmental stressors (e.g., arsenite, UV and gamma radiation, and oxidants) (see other studies cited in Carpenter, 1999; Jones et aL, 1997; Miyamoto et aL, 1996; Rosen and Greenberg, 1996; Yamauchi et aL, 1997; Zwick et aL, 1997). In certain instances, EGFR tyrosine kinase activity is necessary to 15 initiate signalling, while in others, it may act as a structural adaptor. For example, UV radiation appeared to provoke EGFR activation that mimicked addition of a cognate ligand: receptor dimérisation, phosphorylation, aggregation and internalization, the binding of signalling molecules and phosphorylation of EGFR substrates. However, instead of directly activating EGFR kinase activity, it was suggested that UV radiation induced EGFR phosphorylation through inactivation of phosphotyrosine phosphatase activity or production of reactive oxygen species (Carpenter, 1999).

In addition, the EGFR can be ‘transactivated’ by GPCRs for mitogenic signalling (Daub et aL, 1996). There may be several mechanisms by which GPCRs transactivate RTKs, and they may differ between different GPCRs and different cellular contexts. Protein kinase C (PKC) activity is required for EGFR transactivation in HEK-293 cells but not in smooth muscle cells (Eguchi et aL, 1998; Tsai et aL, 1997). The non-receptor tyrosine kinase, c-Src, may be playing a central role by associating with and phosphorylating the EGFR, as reviewed in Thomas and Brugge (1997). For example, overexpression of Csk (a regulatory kinase that inhibits Src function) or a dominant-negative Src blocked EGFR phosphorylation in COS-7 cells when LPA or a2 adrenergic receptors were activated (Luttrell et aL, 1997). Src activation was also found to precede EGFR transactivation in vascular smooth muscle cells (Eguchi et aL, 1998). Two alternative mechanisms were proposed to describe Src activation by GPCRs (Carpenter, 1999) - via (i) phospholipase C-P (PLCp)-mediated inositol-1,4,5-trisphosphate (Ins?]) production and subsequent activation of Src by Ca^^-activated Pyk2 (a cytoplasmic tyrosine kinase), or (ii) in the case of P 2 adrenergic receptors, binding of the adaptin-type molecule p-arrestin to phosphotyrosine sites on the receptor and activation of Src through interaction with the receptor-arrestin complex.

Ullrich and colleagues subsequently proposed a radically different mechanism for GPCR- mediated EGFR transactivation. They showed that GPCR stimulation rapidly induced an extracellular metalloproteinase activity, which then cleaved proHB-EGF (Prenzel et aL, 1999). Subsequent binding of HB-EGF to the extracellular domain of the EGFR activated the latter, while inhibition of proHB-EGF cleavage or of HB-EGF mitogenic activity completely blocked GPCR-induced EGFR transactivation and downstream effectors.

Following receptor activation and dimérisation, the intracellular kinase domains, now in juxtaposition, cross-phosphorylate Tyr residues in trans. The intrinsic protein tyrosine kinase activity of EGFR is essential for initiation of many downstream signalling

16 pathways (Chen et a l, 1987; Honegger et a l, 1987a; Honegger et al., 1987b; Livneh et al., 1986; Moolenaar et al., 1988; Prywes et al., 1986). For example, a point mutation within the ATP-binding site of EGFR abolishes its protein tyrosine kinase capability and its ability to stimulate phosphatidylinositol-4,5-bisphosphate (Ptdlnsf], or PtdIns(4,5)P2 wherever ambiguity may occur) hydrolysis by PLC (Moolenaar et al., 1988).

At least five autophosphorylation sites have been identified: three major (Tyr-1068, 1148, and 1173) and two minor (Tyr- 992 and 1086) (Downward et al., 1984; Hsuan et al., 1989; Margolis et al., 1989b; Walton et al., 1990). Phosphorylation of these residues does not appear to increase the activity of the kinase domains but instead serves to create binding sites for cytosolic proteins with SH2 and PTB domains (see Section 1.4.1). Some of these proteins are enzymes that subsequently became tyrosine-phosphorylated by the receptor and thereby activated, and include Src, PLC and phosphoinositide 3-kinase (PI3K). Others, such as She, Grb2 and Nek, act as adapter molecules to link RTKs to downstream signalling pathways.

In receptors such as those for platelet-derived growth factor (PDGF) and nerve growth factor (NGF), all of their autophosphorylation sites are stringently required for binding specific downstream signalling molecules (Fantl et al., 1992; Kashishianet al., 1992; Kazlauskas et al., 1992; Lechleider et al., 1993; Nishimuraet al., 1993; Obermeier et al., 1993a; Obermeier et al., 1993b; Ronnstrand et al., 1992; Valius et al., 1993). In others such as the EGFR, none of the autophosphorylation sites are individually essential (i) for mediating association with at least four SH2-containing signalling molecules (Soler et a l, 1994) and (ii) for a mitogenic response (Honegger et al., 1988). Truncation of the C- terminus of the EGFR, including the five autophosphorylation sites, does not abolish EGF-induced proliferation and growth. Nevertheless, simultaneous deletion or mutation (to phenylalanine residues) of all five autophosphorylation sites in the EGFR reduces receptor kinase activity, abolishes formation of association complexes, and delays receptor internalisation and generation of a proliferative signal (Helin and Beguinot, 1991; Sorkin et al., 1991).

Stover et al. (1995) attempted to explain these seemingly contradictory observations by proposing that Src can phosphorylate three Tyr residues at the C-terminus of the EGFR kinase domain, which is sufficient to mediate several EGF-stimulated responses. In an unstimulated EGFR, the C-terminal tail is proposed to fold over and block these residues (Stover et al., 1995). If the autophosphorylation sites are mutated, the tail remains

17 unphosphorylated and the Src-targeted residues are blocked. Removal of the tail by truncation is proposed to expose these residues for phosphorylation by Src and subsequent interaction with SH2-containing signalling proteins.

A similar mechanism probably operates in the generation of hydrogen peroxide (H 2O2) following EGF-treatment of A431 cells. H 2 O2 is the predominant reactive oxygen species produced downstream of the EGFR and production is abolished by a kinase-inactive EGFR, but not by a C-terminally-truncated EGFR (Bae et a l, 1997). It was also suggested that H 2O2 inhibits protein tyrosine phosphatase activity so that EGF-induced protein tyrosine phosphorylation is enhanced (Bae et al., 1997).

At least ten other residues on the EGFR are also phosphorylated by heterotypic kinases, such as Src, PKC and CAM kinase II. For example, Tyr-891 and 920 are phosphorylated by Src and may serve as binding sites for Src itself and the p85a regulatory subunit of PI3K, respectively (Songyang et a l, 1993; Stover et a l, 1995).

Interestingly, two recent reports proposed new mechanisms of ErbB-mediated signalling. After stimulation by EGF and heregulin, respectively, full-length EGFR and the intracellular domain of ErbB4 (following cleavage by y-secretase) are able to translocate to the nucleus to promote cell proliferation and differentiation (Lin et al., 2001; Ni et al., 2001). In particular, EGFR may be functioning as a transcription factor.

Thus, the multiplicity of ligands, the ability to form EGFR heterodimers, and different ligand-receptor affinities combine to generate variation in interactions with intracellular signal transducers, in the duration and intensity of signalling, and in proliferative responses.

1.3 Signalling Pathways

The specific set or combination of signalling molecules that bind to a receptor will determine which signal transduction cascade will be initiated. This in turn may depend on the repertoire of downstream signalling molecules that is expressed by the cell in which the receptor is resident. In addition, cross-talk between various stimulated pathways serves to further fine-tune their signals and thus the eventual cellular response to the original extracellular stimulus. Consequently, the responses of different cell types to the same receptor-binding ligand can be quite disparate.

18 A key feature of intracellular signal transduction is the binding of signalling molecules within a cascade to the receptor or among themselves via specialized protein modules or “homology domains”, leading to specific associations and creating multimeric complexes. Signalling molecules that bind RTKs can be divided into three main classes, depending on the mechanisms by which they are regulated by the receptor in order to relay a message. First are those whose enzymatic activities are modulated either through a conformational change or by phosphorylation on specific Tyr residues by the receptor. Second are cytosolic enzymes that are physically juxtaposed with their upstream activators, downstream effectors and/or higher local concentrations of substrates as a result of translocation to the receptor at the plasma membrane. Thirdly, some molecules have no known enzymatic properties and they act instead as adapter proteins or docking proteins that can bind multiple protein ligands. These proteins use homology domains to link receptors with downstream enzymatic effectors to form large signalling complexes.

1.4 Homology Domains

The transduction of chemical messages from the activated receptor to the cell interior, including the nucleus, typically requires the assembly and functional interaction of signalling molecules in order to relay the signal. Specificity in these interactions is often achieved by several classes of polypeptide domains which bind cognate amino acid sequences in target proteins and/or lipids and lipid derivatives. Some signalling enzymes have more than one copy or class of these homology domains, while adapter proteins are comprised entirely of such domains, acting to bring several proteins together. Furthermore, some domains, such as the SH2 and SH3 domains of Src family kinases, can bind either to a motif within the same molecule or on other proteins via intermolecular interactions. Thus, modular domains may localize proteins to specific subcellular sites, regulate enzyme activity, facilitate formation of multi-protein complexes and transduce signals directly (Pawson and Scott, 1997).

1.4.1 SH2 and Phosphotyrosine-binding (PTB) Domains

One of the first types of binding domains to be identified was the src-homology-2 (SH2) domain, whose amino acid sequences have an average length of about 100 residues and are all similar to a conserved, non-catalytic region of the non-receptor tyrosine kinase ppbO^'^"^^ family (Pawson, 1995). While all SH2 domains bind phosphotyrosine (pTyr) residues, they differ in their specificity for the five amino acid residues lying C-terminal to the pTyr residue. For a RTK, the sequence surrounding its autophosphorylation sites 19 will dictate which SH2 domain-containing proteins are able to associate with it and therefore determine the signalling pathways that it can activate. Thus, SH2 domains mediate docking of signalling molecules at specific sites in response to Tyr residue phosphorylation which, in turn, may facilitate substrate phosphorylation, regulate enzymatic activity or alter cellular localization (Moran et a l, 1991; Pronk et al., 1993; Sugimoto et a l, 1994; Uchida et a l, 1994; Vega et a l, 1992).

For example, the phosphorylation of Tyr residues in the cytoplasmic half of the EGFR induces the direct binding of several SH2 domain-containing proteins, such as PLCy, Grb2 (Lowenstein et a l, 1992) and ras-GAP (Serth et al., 1992). Interestingly, SH2 domains have also been found to bind phosphatidylinositol-3,4,5-trisphosphate (PtdlnsPg) competitively with pTyr residues in vitro (Rameh et al., 1995). This suggests that proteins with SH2 domains may also be targeted to PtdlnsPs-containing membranes, and that SH2 domain interactions may be regulated by Ptdlns?].

The PTB domain also binds pTyr residues but, unlike the SH2 domain, uses hydrophobic amino acid residues that lie 5-8 residues N-terminal to the pTyr site to determine specificity (Trub et al., 1995; van der Geer et al., 1996). There are also PTB domains that can bind non-phosphorylated peptide motifs, suggesting that PTB domains are more general peptide recognition modules (Borg et al., 1996; Li et al., 1997). The PTB domains are comparatively less widespread than SH2 domains, structurally related to PH domains (Eck et al., 1996; Zhou et al., 1995)(see below), and also bind phosphoinositides

(Pis)' such as Ptdlns?] and Ptdlns ? 2 (Rameh et al., 1997a).

1.4.2 SH3 and WW Domains

SH3 domains, as the name implies, are also homologous in sequence to a conserved, non- catalytic part of ppbO^'^"^^, and have been found in many intracellular signalling molecules, some of which bind the EGFR. SH3 domains are smaller in size (50-75 residues) than SH2 domains and bind proline-rich sequences with the consensus VXX? that form a left- handed polyproline type II helix (Feng et al., 1994). Like SH2 domains, SH3 domain binding specificity is defined by residues adjacent to the proline-rich region, and phosphorylation of serine or threonine residues within this region may influence SH3 domain binding (Chen et a l, 1996). SH3 domains can form functional oligomeric

' Phosphoinositide (PI) is a generic term which includes phosphatidylinositol (Ptdlns) and its various phosphorylated forms or PPIs. 20 complexes at defined subcellular sites, often in conjunction with other modular domains (Pawson, 1995).

In addition, SH3 domains appear to have regulatory properties: for example, isolated SH3 domains from pp60^'^"^^ and Grb2 can stimulate the large G protein dynamin (Gout et al., 1993), while the SH3 domain of PLCyl was shown to exhibit guanine nucleotide exchange factor (GEF) activity towards a nuclear GTPase, PI3K enhancer (Ye et al., 2002). In addition, mutations of SH3 domains encoded within c-abl and c-src proto­ oncogenes can enhance their oncogenic activity (Hirai and Varmus, 1990; Jackson and Baltimore, 1989; Seidel-Dugan e/a/., 1992).

The WW (or WWP) domains derive their name from conserved tryptophan residues and have average size of about 40 residues (Chan et al., 1996). They also bind proline-rich sequences, but with the consensus PPYY or PPLP (Macias et al., 1996). More recently, the WW domain of the peptidyl-prolyl isomerase Pinl and the ubiquitin ligase Nedd4 were shown to bind phosphoserine- and phosphothreonine-containing sequences, and in the case of Pinl, this binding is essential for its function in vivo (Lu et al., 1999b). WW domains may also regulate enzymatic activity contained within the same molecule, for example when the WW domain forms part of the ligand-binding surface and contributes to substrate recognition (Ranganathan et al., 1997).

1.4.3 Pleckstrin Homology (PH) Domain

PH domains vary widely in sequence and a consensus can be only loosely defined. Their functions are also poorly understood in many proteins due to their ability to bind phospholipids, inositol phosphates and Py subunits of G proteins (Harlan et al., 1994; Irvine and Cullen, 1996; Rameh et al., 1997a; Touhara et al., 1994). Hence, it was postulated that PH domains may regulate the subcellular targeting of signalling proteins to specific regions of the plasma membrane for interactions with regulators or targets (Harlan et al., 1994; Lemmon et al., 1995). It is perhaps not surprising then that PH domains are found in PI kinases, inositol phosphatases and phospholipases (Toker and Cantley, 1997). Perhaps the best-characterised PH domain is that of PLCôl which can bind either Ptdlnsf] or InsPs. This enables the PH domain of PLCôl to target the protein to the plasma membrane (Paterson et al., 1995; Yagisawa et al., 1998) and to act as both a mediator of processive catalysis (when binding PtdlnsPi or PtdIns? 2 -containing

21 membrane) and a feedback inhibitor (when binding Insfg) (Cifuentes et al., 1993; Garcia et a l, 1995; Lemmon et al., 1995; Lomasney et al., 1996).

1.4.4 C2 Domain

Like PH domains, C2 domains are found in many proteins involved in signal transduction and membrane interaction (Ponting and Parker, 1996), but their functions are poorly defined, due to their ability to bind different phospholipids, inositol phosphates, and proteins. In some cases such as syntaxins, binding is dependent on Ca^^ ions. Also known as CalB domains, C2 domains come in a wide variety of sizes (the smallest being 100 residues long) and can be found either singly (as the extended Y box of PLCy) or in pairs (as the C2A and C2B domains of synaptotagmin) (Cho, 2001; Rizo and Sudhof, 1998).

1.4.5 PDZ Domain

Most PDZ (post-synaptic density protein, disc-large, zo-1) domains recognise short binding sites (S/TYV) at the C-terminus of target proteins with a C-terminal hydrophobic residue and a free carboxyl group. Specificity is defined by the residues at the -2 to -4 positions relative to the C-terminus and binding may be regulated by phosphorylation. The common structure of PDZ domains comprises six p strands (PA- pp) and two a helices (aA and aB), which fold into a six-stranded P sandwich (Hung and Sheng, 2002). Many PDZ-domain-containing proteins have multiple copies of these domains and may form oligomers. Individual PDZ domains may have specific target peptides, as shown for the 95-kDa post-synaptic density protein PSD-95, and may associate through intermolecular disulphide bonds (Hsueh et al., 1997). As many proteins interacting with or containing PDZ domains are localized at the plasma membrane, PDZ domains are thought to provide a framework or scaffold for the recruitment and assembly of target molecules into membrane-bound macromolecular signalling complexes (Ponting et al., 1997). Such interactions may be affected by changes in PI levels in membranes, as PDZ domains were recently shown to bind Ptdlnsf] (Zimmermann et al., 2002). In addition, PDZ domains were shown to modulate the function of their associated proteins (Raghuram et a l, 2001 ; Wang et al., 2000).

1.4.6 FYVE Domain

The FYVE domain, named after the first letters of four proteins that were initially found to contain this domain, is found in several peripheral membrane proteins, such as the

22 human early endosome autoantigen 1 (EEAl) protein and Fab Ip, a yeast lipid kinase (see Section 1,5.4.1) (Stenmark et al., 1996). It binds two Zn^^ ions and specifically binds PtdlnsSP in vitro (Burd and Emr, 1998; Gaullier et al., 1998; Kutateladze et al., 1999; Patki et al., 1998). The crystal structure of a FYVE domain revealed a shallow, basic groove that is too small to accommodate Ptdlns bis- and trisphosphate and that is also incompatible with binding PtdIns4P (Misra and Hurley, 1999). Thus, the FYVE domain appears to target effector proteins to membrane compartment(s) that contain Ptdlns3f (Gillooly et al., 2000).

1.4.7 ENTH Domain

The epsin NH2-terminal homology (ENTH) domain, which binds PtdInsP 2 , is found primarily in endocytic proteins such as epsin, adaptor protein 180 (API80) and huntingtin interacting protein HIPl (de Camilli et al., 2002). Many of these proteins have additional roles in signalling and actin regulation. Although ENTH domains from different proteins share low sequence similarity, they are very similar structurally. Its compact, globular

structure is a superhelix of a-helices, but the number of a-helices and the Ptdlns? 2- binding site can differ among various ENTH domains. For example, a sub-group of

ENTH domains (e.g., in the API80 homologue CALM) bind Ptdlns ? 2 via a lysine-rich motif (Ford et al., 2001), while the rest (e.g., in epsin) contain a pocket of basic residues, which is conserved in all ENTH domains (Itoh et a l, 2001).

1.4.8 PX Domain

The PX (Phox homology) domain is another Pl-binding structural motif that has been found in diverse proteins involved in various different cellular processes (see recent review by Xu et al., 2001b). PX domains from various proteins have different specificity for Pis. For example, the PX domain of sorting nexin 3 can interact with Ptdlns3f but not other Pis (Xu et al., 2001a), while that of class II PI3K interacts specifically with Ptdlns(4,5)?2 (Song et a l, 2001b). Nevertheless, recent studies of PX domain structures indicate a common flat, compact conformation consisting of three p-strands followed by three a-helices (see, for example, Hiroaki et al., 2001). Furthermore, a conserved proline- rich motif that resembles a SH3 domain-binding site has been found in PX domains (Hiroaki et al., 2001; Takemoto et al., 1999). Thus, it appears that PX domains not only target proteins to membrane but may also participate in protein-protein interactions.

1.5 Phosphoinositide Signalling Pathways 23 Ptdlns is a eukaryotic lipid occurring widely as a minor component of many cellular membranes. It contains a diacylglycerol (DAG) group, usually l-stearyl-2-arachidonyl- glycerol in mammals, linked via a phosphodiester bond to a D-myo-inositol headgroup which can be phosphorylated at a number of hydroxyl groups. PI metabolism is widely initiated by growth factors and hormones to transmit mitogenic, developmental and other signals within the cell. Although polyphosphoinositides (PPIs) are minor lipids in eukaryotic cells, they play important roles in many cellular processes (Simonsen et al., 2001; Toker, 2002). PPIs have been found to regulate the activities and/or distribution of proteins that contain PPI-binding domains (see Section 1.4) and which play major roles in signal transduction, cytoskeletal rearrangement or membrane tracfficking (see, for example. Section 1.5.2).

Ptdlnsfi also serves as precursor of second messengers. In the classical model of this pathway, the primary substrate, Ptdlns, is sequentially phosphorylated at the D-4 and D-5 positions of its inositol ring to generate phosphatidylinositol-4-phosphate (Ptdlnsf, or

PtdIns4P wherever ambiguity may occur) and then Ptdlns? 2 , respectively (Fig. 1.2). Ptdlnsf] is then cleaved by PLC at its phosphodiester bond to generate two second messengers: DAG, which activates PKC, and Insfg, which triggers Ca^^ ion release from the endoplasmic reticulum (ER) (Berridge, 1993; Nishizuka, 1988). These ultimately lead to changes in many cellular processes such as vesicle trafficking and secretion, cytoskeletal association, and apoptosis inhibition.

PI signalling pathways have also been detected in the cell nucleus and are suggested to play roles in the cell cycle, chromatin arrangement, mRNA processing, and cell differentiation (Cocco et a l, 2001; Irvine, 2002). It was further proposed that PI lipids in the nucleus may be residing not in the nuclear envelope or a bilayer membrane but instead forming proteolipid complexes with proteins (see discussion in Irvine, 2002; Maraldi et a l, 2000). In addition, PI kinases and lipases can be translocated to the nucleus. For example, the amount of phospholipase C increased in the nucleus within minutes after treatment of human osteosarcoma Saos-2 cells with interleukin la (Zini et a/., 1996).

The fact that DAG and Insfg act to initiate other cascades of phosphorylation places Ptdlnsf] hydrolysis in a central position in signal transduction. Signalling by DAG and Insf] stops when DAG is phosphorylated to phosphatidic acid (PtdOH) and Insf] converted into various inositol polyphosphates (Irvine and Schell, 2001). No known

24 biological function has been established for most inositol polyphosphates, although inositol-1,3,4,5-tetrakisphosphate (InsP^) is known to be involved in Ca^^ ion flux and the regulation of small G proteins (Irvine and Cullen, 1996). A recent report suggested that

Ins ? 4 acts as a bi-modal regulator of cellular sensitivity to Insf] - facilitating InsP^- induced Ca^^ influx at low concentrations and inhibiting Insfg receptors at high concentrations (Hermosura et ah, 2000). In the nucleus, inositol polyphosphates are apparently involved in mRNA transport from the nucleus (Odom et al., 2000; Salardi et al., 2000; York et a l, 1999). PtdOH and the various inositol polyphosphates are sequentially modified in a series of enzymatic reactions and eventually re-form Ptdlns at the ER. This is commonly known as the Ptdlns cycle (Fig. 1.2).

Some studies suggested that this cycle comprises a closed ‘pool’ of metabolites, whose turnover is dependent on extracellular agonists such as growth factors (and is hence termed the agonist-sensitive pool). This is in contrast to the agonist-insensitive pool, which is thought to have housekeeping functions. However, this model of discrete pools or compartments has yet to be accepted due to conflicting results from other studies (Monaco and Gershengom, 1992).

1.5.1 Phospholipase C

In Pl-dependent PEG signalling pathways, the decrease in the level of Ptdlnsfz in the membrane caused by hydrolysis may act as a signal itself because the activities of several proteins are directly modulated by this phospholipid (see below). This further emphasizes the need for the stringent regulation of PEG activity.

At least 11 PEG isoforms have been cloned in mammals. They are divided into four classes (namely, P, y, Ô and s), each with an apparently different mode of regulation. Alternative mRNA splicing gives rise to additional diversity in structure and function. All mammalian Pl-specific PEGs (hereafter termed “PEGs” for simplicity) identified to date are single polypeptides and contain a single catalytic domain (which can be separated into two regions, termed X and Y), and a G2 domain (Eee and Rhee, 1995; Parker et al., 1994; Ponting and Parker, 1996). The X and Y regions are both required for enzymatic action; in the case of PEGÔ1, the X domain is responsible for the catalytic hydrolysis of PPIs while the Y domain is involved in substrate binding (Gheng et al., 1995). In addition, a PH domain and an EF-hand domain are also found at the N-terminus of each p, y and ô isozymes.

25 C H j - O - C - R , type I

R? — C — O ~ CH ADP PtdlnsPK NH; PlKfyve CMP ATP CH;—(P)— O —(P)— CH; AI Ptdlns Ptdlns(3,5)P2 HO OH OH ADP ATP Ptdlns synthase CDP-DAG

R j - C - O - C H C H j - O - C - R ,

R j - C - O - C H

Ptdlns3K OH Ptdlns3P HO

OH P P i

OH CDP-DAG ATP Inositol PtdlnsP4-Pase synthase OH PlKfyve " ATP Pi CTP Ptdlns4K ADP type II PtdlnsPK p85/p110PIK

o CHj— O — C — R, ADP

R ; - C - 0 - C H

C H j - O - C - R , C H j - ( p ) - 0 OH R j - C - O - C H PI3K C H ; - 0 - C - R , PtdlnsSP Ptdlns4P R ;— c — O — CH OH HO C H ; - 0 ATP PtdOH type I PtdlnsPK

ADP ATP A D P - ^ PtdOH type II PtdlnsPK phospho- ADP DAG kinase hydrolyase type I ADP “0 ATP InsPa Pi PtdlnsPK R2 -C- 0 -CH © ATP ^ ATP ADP

CHj— O — C — R,

R j - C - O - C H PI3K

O C H ; - 0 - C - R , II I R;— 0 — O — CH I CH;-OH

Ptdlns(3,4,5)P3 DAG PLCs can also hydrolyse Ptdlns and PtdIns4P in vitro but PI 3-phosphates appear to be refractory to such activity. Substrate preference is influenced by Ca^^ concentration:

Ptdlns4f and Ptdlns ? 2 are preferred at micromolar Ca^^ concentrations while Ptdlns is cleaved at higher, non-physiological concentrations (Ryu et aL, 1987).

1.5.1.1 PLCÔ

The PLCÔ sub-class, with four known members (01-54), is weakly activated by Gpy subunits in vitro but has not been shown to be directly regulated by receptors (Lee and Rhee, 1995). Other studies have shown that it can be immunoprecipitated with and stimulated by pl22-RhoGAP, which activates the GTPase activity of Rho (Homma and Emori, 1995). Since Rho regulates the assembly of focal adhesions and actin stress fibres (Ridley and Hall, 1992) and Ptdlnsf] can bind to and regulate the activity of several actin-binding proteins (see below), a link between Rho and Ptdlnsf] via PLC5 and pl22- RhoGAP was suggested. Exchange of GDP on Rho for GTP may induce association between pl22Rho-GAP and PLC5, leading to PLC5 activation, Ptdlnsf] hydrolysis and changes in cytoskeletal organization (Homma and Emori, 1995). PLC5 has also been found to be activated by and to bind Gh protein a-subunits, which are downstream targets of stimulated ai-adrenergic receptors (Nakaoka et a l, 1994). More recently, PLC51 (but not PLCyl and PLCpl) was shown to be stimulated by increases in Ca^^ concentration within the physiological range (from 100 nM to 1-10 pM), and full activation required an intact PH domain (Allen et a l, 1997). Thus, it is possible that PLC51 contributes to the Ca^^-stimulated PEG activity observed in many cellular systems.

The Saccharomyces cerevisiae gene PLCl, which is closely related to the mammalian PLCÔ isoforms, is not an essential gene. Its deletion resulted in a number of phenotypes, such as slower growth, defective utilisation of carbon sources other than glucose, altered cell morphology, inability to complete cytokinesis and increased frequency of chromosome mis-segregation (Flick and Thomer, 1993; Payne and Fitzgerald-Hayes, 1993). Specifically, Pic Ip interacts with the Ptdlns kinase homologue Tor2p, which is involved in the regulation of protein synthesis, cell cycle progression and organization of the actin cytoskeleton (Lin et a l, 1998). This led to the suggestion that Tor2p, as a putative Ptdlns kinase, phosphorylates Ptdlns, which is then hydrolysed by the associated Plclp. In mice, disruption of the PLCS4 gene resulted in males being sterile or producing

27 very few progeny. Sperm from PLCô4'^‘ mice was unable to interact appropriately with the egg coat, the zona pellucida, to initiate the acrosome reaction (Fukami et aL, 2001).

The crystal structure of PLCôl without its N-terminal PH domain has been determined (Essen et aL, 1996). The catalytic domain is described as a “distorted but closed” triosephosphate isomerase-like barrel, with the active site at the C-terminal end. The C2 domain is structurally similar to that from synaptotagmin I: an 8 -stranded, anti-parallel p- sandwich (Sutton et aL, 1995). In addition to the Ca^^ ion required by the C2 domain, another Ca^^ ion is found at the bottom of the active site, and together with residues His- 311 and His-356 in the catalytic domain, are essential for Ptdlnsf] hydrolysis. A two-step catalytic mechanism, the so-called “tether and fix” model, has been proposed, in which the PH domain attaches, or tethers, PLC5 to the membrane by binding Ptdlnsf], while the EF-hand domain acts as a flexible link between the PH domain and the rest of the enzyme. The C2 domain then fixes and orientates the catalytic domain towards the membrane, enabling processive substrate catalysis. Since all eukaryotic PLCs require Ca^^ for activity and the two aforementioned histidine residues are highly conserved among all PEG isoforms, it is likely that the same catalysis mechanism is used by other isoforms.

1.5.1.2 PLCP

PLCp isozymes are regulated by GPCRs, but different isozymes are sensitive to different heterotrimeric G proteins. For example, PLCp2 and PLCp3 are most sensitive to the Py- subunits of Gj/o proteins (p3 ~ p2 » pi » p4) while Gq a-subunits have a preference for PLCpl (pi > P3 > p4 > p2) (Lee and Rhee, 1995). This variation in sensitivity probably reflects the minimal sequence homology of PLCp isozymes both at their N- terminal PH domains, which interact with Gj/o Py-subunits, and at their C-terminal regions, which interact with Gq a-subunits (Lee et aL, 1993; Rhee and Bae, 1997). In particular, charge interactions have been suggested to be crucial for both general protein- protein interactions and signal transfer interactions between Gp-subunits and PLC-P2 (Buck et aL, 1999), while post-translational prenyl modification at the C-terminus of the y-subunit plays important roles in interacting with and modulating the function of PLCp

(Akgoz et aL, 2002; Fogg et aL, 2001). PtdOH also increased stimulation of PLCpl activity by GPCR (Litosch, 2000). In addition, it may be possible that Gqa and Gj/o py- subunits modulate a single PLCp molecule independently and concurrently by interacting

28 with its C- and N-terminal regions, respectively. This may allow signalling by various PLCp isozymes to discriminate between different types of GPCRs and hence different agonists, leading to cellular responses with varying intensity and duration.

Interestingly, PLCpl can act as a GTPase-activating protein (GAP) for a purified mixture of closely related Gq and Gn proteins (Berstein et aL, 1992) and thus rapidly terminate its own activation via a negative feedback loop. This may provide the type of fast response required for such biological processes as nerve signal transduction and G-protein- stimulated, PLCp-mediated phototransduction (for example, in the Drosophila visual system) (Stark et aL, 1983; see also below).

In mouse neutrophils, PLCpi and PLCp3 are the only PLC isoforms that are activated by chemoattractants and play an important role in formylmethionyl-leucyl-phenylalanine (fMLP)-mediated production of superoxide, autophosphorylation of PKC and activaton of Jun N-terminal kinases and mitogen-activated protein kinases (MAPKs) (Li et aL, 2000c). Interestingly, different methods of targeted disruption of PLCP3 expression in mice produced different phenotypic manifestations. Wu and co-workers showed that mice lacking PLC(33 had increased behavioural and cellular responses to p opioids (Xie et aL,

1999) and started to develop spontaneous, multifocal skin ulcers at 6 months of age (Li et aL, 2000c), while others reported embryonic lethality (Wang et aL, 1998). In the brain, the prevalent forms of PLCp isozymes (pi, P3 and p4) are each uniquely distributed and interact with different receptors. PLCpl is highly expressed in the cerebral cortex and hippocampus and is coupled predominantly with the muscarinic acetylcholine receptor. Transgenic mice with a null mutation for PLCpi developed epilepsy (Kim et aL, 1997).

PLCp4 is mostly found in the cerebellum and acts downstream of the metabotropic glutamate receptor; PLCp4'^' mice showed a motor defect (ataxia) and retarded development of their cerebellum (Kim et aL, 1997). PLCp3 is found in trace amounts throughout the brain (Kim et aL, 1997; Ross et aL, 1989; Tanaka and Kondo, 1994).

Thus, different PLCp isozymes are likely to perform distinct functions, at least in the brain.

1.5.1.3 PLCy

The PLCy class has two isozymes, types 1 and 2, and is primarily regulated by both receptor type and non-receptor type PTKs. In addition, PLCy has a SH3 and two SH2 domains, and a second, split PH domain lying between the X and Y regions of the 29 catalytic domain (Parker et aL, 1994). It has been proposed that this central, SH domain region may play a role in autoregulation of PLCy activity, by functioning as an intramolecular “cap” or “lid” to occlude the active site (Carpenter and Ji, 1999; Kamat and Carpenter, 1997). PLCy binds via its SH2 domains to pTyr residues of activated PTKs and is subsequently phosphorylated either by the receptor’s intrinsic protein kinase or associated protein kinases. For example, treating the EGFR-overexpressing, A431 human epidermoid carcinoma cell line with EOF (i) increased both association of PLCy with activated EGFRs and PI turnover in parallel, and (ii) increased the catalytic activity and tyrosine phosphorylation of PLCy in immunoprecipitates (Margolis et aL, 1989a; Nishibe et aL, 1990).

In the case of the PDGF receptor (PDGFR), fibroblast growth factor receptor and NGF receptor, specific autophosphorylated Tyr residues serve as the sites of association with PLCy. Mutation of these residues to phenylalanine caused a substantial decrease in the association with PLCy (Mohammadi et aL, 1991; Obermeier et aL, 1993a). In contrast, none of the five autophosphorylation sites in the EGFR are stringently required for association with PLCyl (Soler et aL, 1994; Sorkin et aL, 1991) or for EGF-stimulated cell growth and transformation. Instead, it was suggested that the N-terminal SH2 domain is the primary contributor to PLCyl association with the EGFR by binding Tyr-1173, while the association of the C-terminal SH2 domain to the Tyr-992 autophosphorylation site is a secondary event (Chattopadhyay et al., 1999). Nevertheless, loss of more than one autophosphorylation site in the EGFR was accompanied by a progressive reduction in association with PLCyl (Soler et aL, 1994; Sorkin et aL, 1991). This suggests a cumulative effect in which all of the EGFR autophosphorylation sites are required to allow maximal association of PLCy.

For some receptors such as the PDGFR, but not others such as the EGFR, association with PLCy depends on continued occupancy of the agonist-binding site on the receptor.

PLCy can associate with the EGFR in untreated cells (Margolis et aL, 1989a) but association with the PDGFR in NIH-3T3 cells is detected only following treatment with PDGF.

After binding the activated receptor, PLCy is phosphorylated on different Tyr residues, depending on the type of receptor. The EGFR phosphorylates PLCyl at Tyr-771, 783, and 1254 and, to a lesser extent, Tyr-472 (Kim et aL, 1990; Wahl et aL, 1990). In PDGF-

30 treated NIH-3T3 cells, phosphorylation of Tyr 783 and 1254 but not 771 is essential for full activation of PLCy (Kim et aL, 1991). Nevertheless, the significance of tyrosine phosphorylation on PLCy remains unclear. While immunoprecipitated, tyrosine- phosphorylated PLCyl from A431 cells has 3 to 6 -fold higher activity following EOF treatment (Nishibe et aL, 1990), others have not found any effect of tyrosine phosphorylation on PLC activity towards PtdInsP 2 , its inhibition by other lipids (e.g., PtdCho) or its sensitivity to Ca^^ (Goldschmidt-Clermont et aL, 1991; Kim et aL, 1990). Furthermore, phosphorylation of PLCy may not increase its activity in vitro, as PLCy isolated from unstimulated cells (and therefore unphosphorylated on its Tyr residues) appeared to be already fully active (Rhee et aL, 1989). Indeed, mutating tyrosine phosphorylation sites did not appear to affect the specific activity of PLCy in vitro or its association with the PDGFR, although the mutation of only Tyr-783 to phenylalanine abolished PDGF-stimulated inositol phosphate generation in NIH-3T3 cells (Kim et aL, 1991). Thus, tyrosine phosphorylation does not directly affect PLCyl activity. Instead, both the binding to autophosphorylated EGFRs and the phosphorylation of PLCy are necessary for full activation of PLC activity and Ptdlnsf] hydrolysis in vivo (Vega et aL, 1992).

Non-enzymatic receptors, such as the T-cell antigen and the az-macroglobulin receptors, which do not possess protein tyrosine kinase activity, can also induce phosphorylation and activation of PLCy, by activating non-receptor PTKs of the Src, Syk and Jak/Tyk families (Dasgupta et aL, 1992).

In addition to regulation by tyrosine phosphorylation, PLCy appears to be regulated by heterotrimeric G proteins. Tyrosine phosphorylation of PLCy was observed in response to stimulation of a GPCR, the m5 muscarinic acetylcholine receptor (Noh et aL, 1995), and PLCy can co-immunoprecipitate with G| a-subunits (Yang et aL, 1991). PLCy activity may therefore be influenced by Ga subunits and there may consequently be cross-talk between G-protein- and RTK-dependent PI signalling (for example, via transactivation of a RTK by a GPCR). PLCy does not appear to be regulated by Py subunits or InsPs in vitro, despite having two putative PH domains (Park et aL, 1993).

After phosphorylation of PLCy, the SH3 domain targets the enzyme to the actin cytoskeleton, which is thought to bring PLCy into contact with its substrate or another activator (Bar-Sagi et aL, 1993; Yang et aL, 1994). It may also be possible that the N-

31 terminal PH domain (Falasca et aL, 1998) and/or the C-terminal SH2 domain (Rameh et al., 1998) bind PtdInsPs and translocate PLCy to membranes, thus enhancing Ptdlns ? 2 substrate availability (Bae et aL, 1998). Pretreatment with pertussis toxin, which ADP- ribosylates and inactivates Gj/o a-subunits, inhibited Ptdlns ? 2 hydrolysis in EGF- stimulated hepatocytes and renal epithelial cells but not other cell types (Johnson et aL, 1986; Liang and Garrison, 1992; Teitelbaum et aL, 1990; Yang et aL, 1991). However, neither translocation to the cytoskeleton nor tyrosine phosphorylation of PLCyl were affected in hepatocytes. Taken together, these results support an earlier conclusion that tyrosine phosphorylation of PLCyl is not sufficient to trigger hydrolysis of Ptdlns ? 2 (Liang and Garrison, 1992; Yang et aL, 1993), and suggest that, at least in some cell types, an interaction between PLCyl and Gj proteins occurs downstream of both PLC tyrosine phosphorylation and its translocation to the cytoskeleton, but before any increase in PtdInsP 2 hydrolysis (Yang et aL, 1994). The above conclusions are consistent with the recent observation that activated, internalized EGFRs were unable to stimulate Ptdlnsf 2 hydrolysis and that the point of regulation lies downstream of PLCyl tyrosine phosphorylation (Haugh et aL, 1999).

A potential substrate for PLCy at the cytoskeleton is PtdInsP 2 bound to cytoskeletal proteins including profilin (Goldschmidt-Clermont et aL, 1991). In human platelets, profilin binds four to five molecules of Ptdlnsf 2 (Goldschmidt-Clermont et aL, 1990; Machesky et aL, 1990), which are accessible to PLCy when it is phosphorylated

(Goldschmidt-Clermont et aL, 1991) or in the presence of Ptdlns( 3 ,4 ) ? 2 or Ptdlnsf] (Lu et aL, 1996). Hydrolysis of these Ptdlnsf 2 molecules removes an inhibition of the interaction between profilin and actin (Lassing and Lindberg, 1985; Lassing and Lindberg, 1988).

PLCy can also be phosphorylated on one or more serine residues, although the effect on enzymatic activity is not apparent in vitro. The phosphorylation of serine residues in PLCy isolated from PDGF-treated NIH-3T3 cells and EGF-treated A431 cells rose several fold relative to that in PLCy isolated from untreated cells (Meisenhelder et aL, 1989). Specifically, Ser-1248 can be phosphorylated by PKC and cAMP-dependent protein kinase (Granja et aL, 1991; Kim et aL, 1989; Olashaw et aL, 1990; Park et aL, 1992).

32 The SH domains also appear to participate in mitogenic signalling because catalytically- inactive PLCy mutants still elicited a mitogenic response in NIH 3T3 cells and mitogenic activity was found to localize to the SH domains (Smith et aL, 1996; Smithet aL, 1994). Recently, the SH3 domain of PLCyl was reported to show GEF activity towards the nuclear GTPase, PI3K enhancer, that activates nuclear PI3K activity and which was necessary for mitogenesis in response to growth factors (Ye et aL, 2002). In addition, PLCy isozymes may be activated in the absence of tyrosine phosphorylation by several lipid-based second messengers such as PtdOH, arachidonic acid and Ptdlnsf] (Bae et aL, 1998; Falasca et aL, 1998; Hwang et aL, 1996; Jones and Carpenter, 1993; Rameh et aL, 1998).

In a whole organism, PLCyl has been found to be essential for normal growth and development. Homozygous disruption of the PLCyl gene caused early embryonic lethality in mice (Ji et aL, 1997); one possible cause could be that development of the haematopoietic and vascular systems were impaired (Liao et aL, 2002). However, fibroblasts derived from these PLCyl-null mutant mouse embryos were still responsive to EGF in assays of proliferation, receptor internalisation and MAPK activation (Ji et aL, 1998). On the other hand, PLCy2 is required to promote B-cell development in mice, apparently as a component of B-cell antigen receptor signalling pathways (Hashimoto et aL, 2000).

1.5.1.4 PLCs

The newly identified PLCs differs significantly from the other PLC classes in several aspects. All PLCs homologues cloned to date have molecular masses of at least 210 kDa, larger than the other classes (Kelley et aL, 2001; Lopez et aL, 2001; Shibatohge et aL, 1998; Song et aL, 2001a). In addition to the C2, X and Y domains, PLCs possesses a N- terminal CDC25-like domain and one or two Ras-associating (RA) domains at the C- terminus. The former is homologous to a family of RasGEF proteins such as CDC25 from yeast and Sosl from human, while the latter resembles those found in Ras-binding proteins such as Ral guanine nucleotide dissociation stimulator (RalGDS) (Ponting and Benjamin, 1996).

Several studies on the functional role of PLCs were reported recently. Upon EGF stimulation of COS-7 cells, recombinant PLCs was observed to translocate to the plasma membrane and the perinuclear region (Song et aL, 2001a). These translocations were 33 thought to be mediated by the small G proteins Ras and Rapl A, respectively. Preliminary studies also suggested that PLCe can activate MAPK via Ras and independently of

Ptdlns ? 2 hydrolysis (Lopez et aL, 2001). Furthermore, PLCs can act as a RasGEF by promoting the exchange of GTP for GDP. Thus PLCs may be participating in growth factor receptor-induced signal transduction. As the increase in PLC activity observed in growth factor-treated cells has been largely attributed to the activation of PLCy, it will be interesting to determine whether PLCs also play a substantial role.

In addition to interacting with small G proteins, PLCs can be stimulated by heterotrimeric G proteins. One study showed stimulation by constitutively active (GTPase-deficient) mutants of the heterotrimeric G protein a subunit G a^ but not Gas (Lopez et aL, 2001).

On the other hand, Schmidt and co-workers demonstrated that PLCs and Ca^"^ signalling can be activated by Gg-coupled receptors such as the p2-adrenoceptor and M 3 muscarinic acetylcholine receptor (mAChR), and proposed that the signal was relayed by the cAMP- regulated GEF Epac and the Ras-related GTPase Rap2B (Evellin et aL, 2002; Schmidt et aL, 2001). In the case of the M3 mAChR, PLCpl was also stimulated via Gaq. However, it remains to be established whether the M3 mAChR stimulates PLCpl and PLCs via two divergent pathways or whether PLCs was activated downstream of PLCpl (Evellin et aL, 2002). It will also be interesting to see whether PLCs can act as a GAP for heterotrimeric

G proteins, as in the case of PLC p. In summary, PLCs appears to be able to interact with both large and small G proteins and is further speculated to be relaying signals from selective GPCRs through two divergent pathways, namely, PI hydrolysis and activation of the Ras/MAPK pathway.

1.5.2 Functional Roles of Phosphatidylinositol 4,5-bisphosphate

In addition to serving as the substrate for PLC and PI3K isozymes, there is increasing evidence that Ptdlnsf] can act as a signalling molecule and modulate cellular processes in its own right, participating in membrane trafficking, secretion, cytoskeleton reorganization and apoptosis (Azuma et aL, 2000; de Camilli et aL, 1996; Eberhard et aL, 1990; Hsuan et aL, 1998). Binding sites for PPIs, particularly Ptdlnsf], have been found in many proteins and these non-covalent protein-PtdIns ? 2 interactions may cause the localization of proteins to subcellular membranes or allosteric stimulation of enzymatic activities.

34 Many cytoskeletal proteins have been shown to bind Ptdlnsf] in ^itro and include gelsolin (Janmey and Stossel, 1987), vinculin (Gilmore and Burridge, 1996), a-actinin (Fukami et aL, 1992) and profilin (Goldschmidt-Clermont et aL, 1990). When bound to these proteins, Ptdlnsfz can promote or inhibit their interactions with other cytoskeletal components and bring about consequent changes in cytoskeletal architecture. For example, profilins are small (12-15 kDa), ubiquitous proteins that bind globular (G) actin monomers, poly-L-proline and PPIs. Profilin-actin interactions are inhibited by

Ptdlns( 4 ,5 )? 2 , PtdIns( 3 ,4 )P2 or Ptdlnsf] (Lassing and Lindberg, 1988; Lu et aL, 1996), and Ptdlns( 4 ,5 ) ? 2 is able to rapidly dissociate profilin-actin complexes in vitro (Sohn and

Goldschmidt-Clermont, 1994). Profilin-bound Ptdlns( 4 ,5 ) ? 2 is not accessible to PLCy unless either PLCy is phosphorylated (Goldschmidt-Clermont et aL, 1991), or

Ptdlns( 3 ,4 ) ? 2 or Ptdlnsfg are produced following growth factor stimulation (Lu et aL, 1996).

In the case of vinculin, Ptdlns ? 2 inhibits intramolecular “head-tail” interaction in vitro and promotes increased binding of vinculin to structural and regulatory contact site proteins, such as talin and F-actin (filamentous actin forms) (Gilmore and Burridge, 1996), which in turn triggers assembly of focal adhesions. This is consistent with studies which have shown that formation of focal adhesions and stress fibres in serum-stimulated Swiss 3T3 cells is dependent on Rho activity (Ridley and Hall, 1992) and that Rho activates type I Ptdlns phosphate kinase (PtdlnsPK) in the same cell type (Chong et aL,

1994). The localized binding and release of Ptdlns? 2 -binding proteins such as gelsolin are also crucial during growth factor-induced cell motility to drive cytoskeletal reorganization and membrane protrusion (see, for example, Chou et aL, 2002).

However, the aforementioned interactions between Ptdlns ? 2 and various proteins are often observed using Ptdlns ? 2 in a non-physiological state. Pis in membrane preparations, liposomes and micelles of different compositions in the presence or absence of detergents and various divalent cations, can be expected to exhibit artefactual interactions with other molecules, arguably in a similar way to the anionic detergent, sodium dodecyl sulphate, which binds many proteins very strongly and with little specificity. Thus, demonstrations of proteins binding to PI micelles in vitro is a poor indicator of in vivo binding and many of these interactions may not be relevant physiologically. For example, the S. cerevisiae mss4-2‘^ mutant, in which the major PtdIns4P 5-kinase (PtdIns4P5K), Mss4p, carries a temperature-sensitive mutation, contains about 35% of the amount of PtdInsP 2 found in wildtype yeast. However, no effect was observed on the organization of the actin 35 cytoskeleton in the mutant (Desrivieres et aL, 1998). This suggests that, at least in yeast, functionally discrete pools of PtdInsP 2 rnay exist or that intracellular Ptdlnsfz levels may have to fluctuate dramatically to exert an effect on the actin cytoskeleton. In addition, although cytoskeletal proteins like profilin, gelsolin, a-actinin and talin bind Ptdlns? 2 , it is still unclear whether these protein-bound lipids are accessible to PLC in vivo.

In the nucleus, Ptdlns ? 2 can apparently modulate the association of SWI/SNF-like Brahma-XQXdXQÔL gene associated factors (BAFs) with actin filament, which leads to chromatin remodelling (Rando et aL, 2002; Zhao et aL, 1998). The lipid appears to localize at multiple nuclear sites, such as on the heterochromatin and interchromatin granule clusters (Mazzotti et aL, 1995) and can modulate the inhibition of RNA polymerase Il-mediated transcription by histone HI (Yu et aL, 1998).

Ptdlns ? 2 is also involved in membrane trafficking, endocytosis and secretion (de Camilli et aL, 1996; Hsuan et aL, 1998). For example, Ptdlns 4-kinase (PtdIns4K) and PtdIns4P5K have been found to be required for ATP-dependent priming of noradrenaline secretion (Hay et aL, 1995; Wiedemann et aL, 1996; see also below). Ptdlns ? 2 apparently participates in secretion in its own right, as opposed to being a substrate for a vesicle- associated PLC, because neither InsPg nor DAG triggers secretion in chromaffin or PC 12 cells (Eberhard et aL, 1990; Hay et aL, 1995). What role does Ptdlns ? 2 play in secretion and vesicle formation? Ptdlns ? 2 activates PLD and the ADP-ribosylation factor (ARE) G protein, both of which are involved in secretion (see Section 1 .5.4.2). Ptdlns ? 2 can also bind many cytosolic and cytoskeletal proteins which are involved in membrane budding (e.g., adaptins), vesicle fission (e.g., dynamin) and a post-docking step in exocytosis (e.g.,

CAPS). In addition, the increased amount of PtdInsP 2 in a membrane may affect its curvature caused by electrostatic repulsion (de Camilli et aL, 1996) and bring about invaginations of the lipid bilayer.

As changes in Ptdlnsf 2 levels are proposed to regulate several cellular processes, enzymes that participate in the biosynthesis and catabolism of Ptdlns ? 2 and its precursor PtdlnsP are likely to be subjected to both spatial and temporal regulation. Several studies have suggested that synthesis of Ptdlns? and Ptdlns ? 2 in response to receptor-derived signals is stringently regulated to maintain a sustained flux of intermediates through the pathway, so as to prevent biologically active PPIs from being synthesized inappropriately (Hsuan et aL, 1998). With the recent discovery of PtdIns5P (Rameh et aL, 1997b) and

Ptdlns( 3 ,5 ) ? 2 (Dove et aL, 1997) in cells, it is now clear that the D-3, D-4 and D-5

36 positions of the inositol ring can be phosphorylated separately and in all possible combinations. It is conceivable that most, if not all, Pis have distinct functional roles. For example, PtdIns( 3 ,5 ) /* 2 is required for normal vacuolar function in yeast (Cooke et aL,

1998). In Chlamydomonas cells, changes in levels of Ptdlns( 3 ,5 ) ? 2 and Ptdlns( 4 ,5 ) ? 2 differed with treatment of different osmolytes and stresses, suggesting that the two

Ptdlns ? 2 isomers have different functions (Meijer et aL, 2001). Thus, it is important to understand the biochemical and regulatory properties of the two principal enzyme families involved in PI metabolism - PI kinases and PI phosphatases (Majerus et aL, 1999).

There are two main classes of inositol phosphatase which, like protein phosphatases, tend to inhibit or terminate signalling (Leslie et aL, 2001; Woscholski and Parker, 1997). Type I inositol phosphatases can target only water-soluble inositol phosphates, while type II enzymes can target inositol phosphates and PPIs. Deletion of inositol polyphosphate 1- phosphatase in Drosophila results in dramatic defects in synaptic transmission and a “shaker” phenotype (Acharya et aL, 1998), while mutation of a type II inositol polyphosphate 5-phosphatase, OCRL-1, causes Lowe Syndrome or oculocerebrorenal dystrophy (Attree et aL, 1992).

1.5.3 Phosphatidylinositol 4-Kinase

1.5.3.1 Structures and Characteristics

PtdIns4Ks are conventionally known as type II and type III PI kinases (Fig. 1.3). (Type I enzymes are PI 3-kinases; see review by Vanhaesebroeck et a l, (1997).) They are characterized according to their sensitivity to adenosine and activation by non-ionic detergents such as Triton X-100 (Endemann et aL, 1987; Pike, 1992). The type II family (classically characterized as membrane-bound, 5 5-kDa proteins and thus termed p55) has a lower Km for ATP and is sensitive to inhibition by adenosine and Ca^"^ (Pike, 1992; Wetzker et aL, 1991). Type III isozymes have larger molecular masses, varying from 92 to over 200 kDa, a higher Km for ATP, but a lower affinity for Ptdlns (Endemann et aL, 1987). They are inhibited by relatively high concentration of wortmannin (WT) and LY- 294002 (a PI3K inhibitor) but are resistant to inhibition by adenosine and Ca^^ (Downing et aL, 1996).

Two isoforms (a and P) of type II PtdIns4K p55 were recently cloned and characterized. Their primary sequences share significant similarity (except in the N-terminal region) but

37 Type II M r (kPa)

PtdIns4K-IIa 54

PtdIns4K-Iip 55

PtdIns4Ka 97

Type III M r (kPal

PtdIns4K-IIIa 230 - S —H

Stt4p 200

PtdIns4K-IIip 92

PIKl 125

Fig. 1.3 Phosphatidylinositol 4-kinase Family A schematic representation of type II and III PtdIns4Ks. The lipid kinase homology, lipid kinase unique and PH domains are indicated in red, green and blue, respectiyely. PtdIns4Ka is a splice yariant of PtdIns4K-IIIa, which also has a SH3 domain (striped) and two proline- rich regions (yertical lines) at the N-terminus. PtdIns4K-IIip contains a domain (grey) that is homologous with PIKl, while a potential w palmitoylation site (white) is found in both PtdIns4K-IIa and PtdIns4K-IIp. Mr, predicted molecular mass. 00 differ from the type III PtdIns4K, PI3K and PtdlnsPK families, and represent a new family of PI kinases which are highly conserved across many species, including S. cerevisiae (Balia et aL, 2002; Barylko et aL, 2001; Minogue et aL, 2001). Both isoforms are able to phosphorylate Ptdlns but not PtdlnsSP, Ptdlns4P and PtdlnsSP, and Ptdlns4K- 11a is also inhibited by the anti-pSS monoclonal antibody 4C5G (Balia et aL, 2002; Minogue et aL, 2001).

Both Ptdlns4K-lla and Ptdlns4K-lip were found to be expressed in many human tissues and, when over-expressed in COS-7 cells, they exhibited a high degree of similarity in their intracellular localization to early and recycling endosomes (Balia et aL, 2002). Although p55 binds strongly to membranes, neither a transmembrane sequence nor a consensus sequence for acylation was detected. Nonetheless, labelling experiments revealed the occurrence of palmitoylation, presumably at a cysteine-rich region at the centre of the protein (Barylko et aL, 2001). It was thus suggested that both isoforms may be participating in clathrin-mediated endocytosis and playing a role in regulating endosomal membrane traffic. Nevertheless, recombinant Ptdlns4K-lla and Ptdlns4K-llp differed significantly in their level of expression and the ^^P-labelling of Ptdlns4P in COS-7 cells.

Ptdlns4Ka (97 kDa), the first mammalian Ptdlns4K to be cloned, exhibited the type 11 enzyme characteristic of being inhibited by the monoclonal antibody 4C5G, which had previously been thought to be specific for p55 (Wong and Cantley, 1994). However, amino acid sequence comparison and Northern blot analysis indicated that Ptdlns4Ka may be an alternatively spliced form of the type 111 Ptdlns4K p230 (see below), as the C- terminal half of rat p230 is 98.2% identical to the sequence of human Ptdlns4Ka (Nakagawa et aL, 1996b; Wong and Cantley, 1994). Corresponding regions of the human enzymes were then found to be virtually identical (Balia et aL, 1997) and human genome sequence data recently confirmed that the two cDNAs are derived from a common gene (NCBl LocuslD 5297). To date, no reports have since confirmed the existence of this isoform: an antiserum specific for recombinant Ptdlns4Ka detected a 180-kDa, but no 97- kDa protein in CHO cells (Wong et aL, 1997). Furthermore, Gehrmann et al. (1996) could find neither a cDNA clone nor a 97-kDa enzyme that corresponded to Ptdlns4Ka in bovine tissues.

Two type 111 Ptdlns4K isoforms (a and p) have been cloned to date and they share a few common homologous regions (Fig. 1.3). The last 254 residues at the C-terminus show 39 varying degrees of similarity to the catalytic domain of PI3Ks as well as a group of proteins which do not display PI kinase activity. This group includes the ataxia telangiectasia protein homologues, human DNA-dependent kinase and the yeast Tor and human FRAP and RAFT proteins. This putative lipid kinase homology (LKH) domain also shares distant homology with the catalytic domains of protein kinases. N-terminal to the LKH domain in type III PtdIns4Ks is a 120-residue “lipid kinase unique” (LKU) domain that is similar to those in PI3Ks, including Vps34p and PI3Kp.

In p230 (230 kDa; also denoted as PtdIns4K-IIIa), the LKU domain and a PH domain are found immediately N-terminal to the LKH domain (Nakagawa et aL, 1996b). Unlike other PtdIns4Ks, it possesses a SH3 domain and two proline-rich regions near its N- terminus. The enzyme also contains five motifs that indicate nuclear localization and DNA binding capabilities, but p230 was not detected in nuclei (Gehrmann et aL, 1996; Nakagawa et aL, 1996b; see also below). p230 is encoded by a 7.8-kb mRNA transcript and corresponds to the full-length version of the 200-kDa type III PtdIns4K purified by Gehrmann et aL (1996). The activity of p230 is enhanced by Triton X-100, but not significantly affected by adenosine and Ca^"^ ions, which is consistent with type III behaviour. However, unlike human PtdIns4Ka, bovine p230 is not inhibited by the anti­ type II PtdIns4K antibody 4C5G (Endemann et aL, 1991). Consequently, it will be interesting to test the specificity of 4C5G against p230 from different species.

In S. cerevisiae, the PtdIns4K-IIIa homologue Stt4p is a cytosolic enzyme of 200 kDa, whose mutation increases sensitivity to the PKC inhibitor staurosporine. Consequently, Stt4p may be involved in PKC signalling (Yoshida et aL, 1994a; Yoshida et aL, 1994b).

At a non-permissive temperature, STT4 mutant cells were arrested mostly at G 2/M phase, suggesting a role for PtdIns4K in cell cycle regulation. Disruption of the STT4 gene resulted in loss of spore viability and cell cycle defects similar to PKC mutations, supporting the notion that it is involved in PKC signalling (Yoshida et aL, 1994a). Furthermore, Stt4p appeared to be involved in maintaining normal vacuole morphology and actin polarization (Audhya et aL, 2000) and regulating some aspect of lipid transport in phosphatidylethanolamine (PtdEtn) biosynthesis (Trotter et aL, 1998).

PtdIns4K-IIip was first cloned as a soluble, 92-kDa enzyme (p92) which is sensitive to WT but not adenosine and whose pattern of expression is similar to that of p230 (Nakagawa et aL, 1996a). A 110-kDa splice variant (pi 10) was later cloned (Meyers and Cantley, 1997). In terms of molecular size and sensitivity to WT and adenosine, p92 and

40 pi 10 have type III PtdIns4K characteristics and correspond to a soluble, 125-kDa PtdIns4K, which has been suggested to allow sustained, agonist-induced PI turnover and

Ins ? 3 generation (Nakanishi et aL, 1995). Unlike the a isoform, the LKU domain is located near the N-terminus of PtdIns4K-IIIp. It also lacks a PH domain between its LKH and LKU domains; instead, it shares a region that is homologous with another yeast PtdIns4K, Piklp.

Piklp (125 kDa) exhibits catalytic properties (such as Km values for ATP and Ptdlns) that are similar to those reported for type II PtdIns4Ks. However, its high molecular weight and resistance to inhibition by adenosine resemble type III enzymes (Flanagan and Thomer, 1992). Interestingly, Piklp was originally reported to be a cytosolic enzyme (Flanagan et aL, 1993; Flanagan and Thomer, 1992), as well as part of the nuclear pore protein complex (Garcia-Bustos et aL, 1994). Indirect immunofluorescence analysis subsequently found Piklp to be localized at the Golgi (Walch-Solimena and Novick, 1999), which may depend on its interaction with Frqlp, a myristoylated, Ca^^-binding protein that also stimulates the lipid kinase activity of Piklp (Hendricks et aL, 1999). Recent studies indicated that Piklp may play a role in secretion (see also below), endocytosis and maintenance of Golgi stmcture (Audhya et aL, 2000). Defective protein transport from the Golgi to the vacuole was observed in p ikl mutant cells (Audhya et aL, 2000; Walch-Solimena and Novick, 1999), and overexpression of Piklp, but not Stt4p, suppressed secretory and growth defects of yeast sec 14-3^^ mutants (Hama et aL, 1999). In contrast to STT4, deletion of PIKl is lethal, indicating that PtdIns4P synthesized by this isozymes is indispensable (Flanagan et aL, 1993; Garcia-Bustos et aL, 1994). Thus, it is likely that Piklp and Stt4p synthesize distinct pools of Ptdlns4f and perform distinct, non-redundant functions in yeast (Audhya et aL, 2000).

Interestingly, p85/p 110-type PI3Ks had recently been shown to phosphorylate the D-4 position of Pis more efficiently than the D-3 position in vivo (Funaki et aL, 1999). In particular, p85/pl 10-type PI3Ks, but not the conventional PtdIns4Ks described above, appeared to be essential for the synthesis of D-4-phosphorylated Pis in insulin-treated 3T3-L1 adipocytes. The ability of PI3Ks to generate D-4-phosphorylated Pis was previously overlooked, possibly due to in vitro assay conditions as well as the fact that PLC is also co-activated in experiments using EGF, NGF or PDGF as agonists (Auger et aL, 1989; Carpenter et aL, 1992; Carter and Downes, 1992; Jackson et aL, 1992). Insulin does not induce activation of PLCy (Wahl et aL, 1992). Thus, at least in the case of

41 insulin and possibly for other agonists as well, the PI4K activity of p85/p 110-type PIK could be important for the production of D-4-phosphorylated Pis in vivo.

1.5.3.2 Regulations and Functions

Most studies of mammalian PtdIns4K activity have indicated that it is bound to membranes, including the plasma membrane, Golgi apparatus, ER, the nuclear envelope and various vesicles, although PtdIns4K activity has also been detected in the cytosol. In addition, type II PtdIns4K activity has also been found to be associated with the cytoskeleton in an EGF-dependent manner (Payrastre et aL, 1991). Thus, it is likely that distinct PtdIns4K isoforms are targeted to different subcellular locations and independently regulated to perform different physiological functions.

The p92 and p230 type III PtdIns4Ks have been detected in many rat tissues, including the brain, kidney, lung, heart, skeletal muscle and testis (Nakagawa et aL, 1996a; Nakagawa et aL, 1996b). In rats, they have both been found in most neurons in the fetal brain while expression is lower in adult brain, suggesting a role for both enzymes in neuronal differentiation and brain development. When expressed in COS-7 cells, both were found localized in the cytoplasmic space adjacent to Golgi vesicles and juxtaposed to the nucleus (Nakagawa et aL, 1996a; Nakagawa et aL, 1996b). p230 was also found on the membrane of Golgi vesicles. The intracellular localization of pi 10 appears to be similar to p92 and p230; anti-pl 10 antibody stained one side of the nucleus at the trans- Golgi network (TGN) and surrounding cytosol (Wong et aL, 1997). Thus, these three PtdIns4Ks may have roles in vesicular trafficking rather than agonist-induced PI turnover at the plasma membrane, since they are not detected at the plasma membrane or within the nucleus.

Indeed, PtdIns4K-IIIp has been shown to be required for the structural integrity of the Golgi complex (Godi et aL, 1999). Furthermore, like its yeast homologue Piklp, PtdIns4K-IIIp associates with and is stimulated by the Ca^^-binding neuronal calcium sensor-1 (NCS-1) or frequenin (Weisz et aL, 2000). NCS-1 has been shown to regulate neuroendocrine secretion, transmitter release and synaptic development (Pongs et aL, 1993) and can co-localize with PtdIns4K-IIIp to the Golgi in COS-7 cells (Zhao et aL, 2001). Ca^^ was found to further stimulate the activity of the NCS-1-associated PtdIns4K- IIip (Zhao et aL, 2001). This modulation of PtdIns4K activity also disrupts the delivery of an apical protein from the TGN to the plasma membrane.

42 Another role may be targeting vesicles to the cleavage furrow during cytokinesis. Mutations in the four wheel drive (Jwd) gene, which encodes a PtdIns4K-IIIp homologue in Drosophila, disrupted the formation of intercellular bridge during male germline cytokinesis (Brill et aL, 2000). The jw d PtdIns4K appears to be necessary for normal actin organization and for tyrosine phosphorylation in the cleavage furrow.

The effects of post-translational modifications on the activity of cloned PtdIns4Ks remain unknown. Early studies concentrated on the membrane-bound, p55 type II enzyme in A431 cells, in which EGF treatment led to an increase in PtdIns4K activity (Walker et ah, 1988). Subsequent studies also found increased p55 activity in anti-EGFR immunoprecipitates from EGF-stimulated A431 cells (Cochet et al., 1991; Kauffmann- Zeh et al., 1994). Similarly, EGFR was detected in anti-type 11 Ptdlns4K immunoprecipitates only after EGF treatment (Kauffmann-Zeh et al., 1994). Cochet et al. (1991) showed that type 11 Ptdlns4K and an uncharacterized Ptdlns4P5K bind to the juxtamembrane region of EGF-activated EGFR and that this interaction is independent of EGF and the receptor’s tyrosine kinase activity. The amino acid sequence of the juxtamembrane region is conserved among members of the EGFR family, suggesting that they share a common mechanism in PI signalling (Hsuan et al., 1998). For example, increased Ptdlns4K activity co-immunoprecipitated with HER2/neu when SK-Br-3 human breast cancer cells were treated with 4D5, a HER2 agonist (Scott et al., 1991). The C-terminus of CD4 receptors also has a sequence similar to the EGFR juxtamembrane region, and a dramatic increase in type 11 Ptdlns4K activity was detected in immunoprecipitated, cross-linked CD4 receptors from T cells (Prasad et al., 1993). However, the exact site of interaction with CD4 receptor was not identified.

Subsequent studies demonstrated that association appeared to be dependent on the phosphorylation state of the EGFR (Kauffmann-Zeh et al., 1994). However, the juxtamembrane region does not possess a tyrosine phosphorylation site, and Ptdlns4K and Ptdlns4P5K activities can associate with EGFR devoid of auto-phosphorylation sites in an EGF-dependent manner (Cochet et al., 1991). Hence, the mechanism of interaction with PI kinases remains unclear. Moreover, the specificity of binding between the acidic type 11 Ptdlns4K and the highly basic juxtamembrane peptides used in these experiments was questioned, because type 11 Ptdlns4K can be activated by polycations and binds anion-exchange columns at physiological pH (Vogel and Hoppe, 1986; Wetzker et al., 1991).

43 It is possible that the interactions of a PtdIns4K and a PtdlnsPK with the juxtamembrane region are mediated by Gq proteins. A synthetic peptide corresponding to the EGFR juxtamembrane region can regulate G proteins in vitro (Stryjek-Kaminska et aL, 1996; Sun et aL, 1995). This is consistent with earlier observations that cholera toxin inhibited increases in PtdIns4K activity in EGF-treated A431 cells (Pike and Eakes, 1987). This is also reminiscent of the key role Gj proteins play in the regulation of PLCy activity following EGF activation as described above. However, recombinant Gpy has no significant effect on PtdIns4Ka activity, even though PtdIns4Ka has a PH domain (Wong and Cantley, 1994). Although the mechanism by which RTKs regulate G proteins remains unclear, there is clear evidence that GPCRs can stimulate EGFR signalling via activated G proteins (Section 1.2.1).

The effect of phosphorylation on PtdIns4K activity is currently unclear. Early studies found that treating A431 membranes with a protein tyrosine phosphatase decreased PtdIns4K activity (Payrastre et aL, 1990). Furthermore, increased PtdIns4K activity was found in anti-pTyr immunoprecipitates from EGF-treated A431 cells and mouse B82L cells compared with untreated cells, and omission of orthovanadate (which inhibits protein phosphotyrosine phosphatases) during immunoprécipitation reduced these increases (Cochet et aL, 1991; Payrastre et aL, 1990). However, the distinction between dephosphorylation of PtdIns4K, EGFR or other co-immunoprecipitating proteins was not made in these experiments. In erythrocyte membranes, stimulation or inhibition of tyrosine phosphorylation led to increased and decreased PtdIns4K activity, respectively (de Neef et aL, 1996). Recent work by Balia and co-workers showed that PtdIns4K-IIip exhibited in vitro autophosphorylation at a site in the C-terminal catalytic core, which in turn reversibly inhibited its activity (Zhao et aL, 2000). Like PI3Ks, PtdIns4K-IIip autophosphorylation is enhanced by Mn^^ ions but inhibited by WT and LY294002 (both inhibitors of PI3K activity). However, the majority of in vivo phosphorylation of PtdIns4K-IIip appeared to be due to additional kinase(s) rather than autophosphorylation.

In addition to tyrosine phosphorylation, the activity of type II PtdIns4K was found to be differentially affected by serine phosphorylation. Surprisingly, dephosphorylation of phosphoserine and pTyr residues decreased and increased PtdIns4K activity respectively in anti-pTyr and anti-EGFR immunoprecipitates from EGF-treated A431 cells (Kauffmann-Zeh et aL, 1994). These observations suggest temporal regulation of PtdIns4K activity by phosphorylation on serine and tyrosine residues.

44 Type II and III PtdIns4Ks have both been proposed to be involved in hormone-stimulated PI turnover. Early biochemical studies suggested that type II PtdIns4K mediates PtdIns4P synthesis in agonist-treated cells (Cochet et aL, 1991; Pike, 1992). Specifically, the p55 type 11 Ptdlns4K co-immunoprecipitated with activated EGFR, and increased type 11 Ptdlns4K, PtdlnsPSK, PLC and DAGK activities associated with the cytoskeleton in A431 cells in an EGF-dependent manner (Payrastre et aL, 1991). Balia and colleagues subsequently demonstrated that WT abolished agonist-stimulated Ptdlns4P production in several cell types and proposed that type 111 Ptdlns4K, which is WT-sensitive, maintains the Ptdlns pools used in agonist-stimulated inositol phosphate production and Ca^^ signalling (Downing et aL, 1996; Nakanishi et aL, 1995). It remains possible that different transmembrane receptors or different cell types employ different Ptdlns4Ks to maintain this pool.

1.5.4 Phosphatidylinositol Phosphate Kinase

PtdlnsPi was once thought to be synthesized solely from Ptdlns4P by Ptdlns4P5K, which phosphorylates the D-5 hydroxyl of the inositol ring. Rapid progress in recent years have led to a reclassification of PtdlnsPKs into three major families, based primarily on their substrate specificities and sequence similarities (Kanaho and Suzuki, 2002): Ptdlns4P5K (also denoted as type 1 PtdlnsPK), Ptdlns-5-phosphate 4-kinase (Ptdlns5P4K; also type 11) and Ptdlns-3-phosphate 5-kinase (Ptdlns3P5K), which primarily phosphorylate Ptdlns4P at the D-5 position, Ptdlns5P at the D-4 position, and Ptdlns3P at the D-5 position, respectively.

1.5.4.1 Structures and Characteristics

Three Ptdlns4P5K isozymes have been found in bovine brain, with different names given by various research groups (Bazenet et aL, 1990; Divecha et aL, 1992; Jenkins et aL, 1991). Two type la isozymes have been cloned and designated PtdlnsPK-la (also denoted as Ptdlns4P5Ka) and PtdlnsPK-lp (Ptdlns4P5Kp) (Ishihara et aL, 1996). Both are 6 8 kDa in size and differ in length by only seven residues. Type lb was subsequently cloned and denoted PtdlnsPK-ly (Ptdlns4P5Ky) (Ishihara et aL, 1998). It has two alternatively spliced forms of apparent molecular mass 87 kDa and 90 kDa. All Ptdlns4P5Ks exhibit sequence homology (75%) in their central kinase domain, while their N- and C-terminal regions are different, suggesting that they probably have distinct functions (Ishihara et aL, 1996; Ishiharaet aL, 1998).

45 In addition to PtdIns4P, PtdIns4P5K isoforms can phosphorylate PtdlnsSf on the D-4 and D-5 positions to form Ptdlns( 3 ,4 ) ? 2 and PtdIns( 3 ,5 )P2 respectively (Rameh et aL, 1997b; Tolias et aL, 1998b; Zhang et aL, 1997). They can also generate Ptdlnsf] directly from Ptdlns3f in a concerted reaction (Zhang et aL, 1997). This dual specificity is consistent with the observation that some PtdIns4f5Ks can weakly phosphorylate Ptdlns to form PtdIns5P (Tolias et aL, 1998b), the probable substrate for PtdIns5P4Ks.

However, PtdIns4P5Ks are unable to phosphorylate PtdlnsSP to produce PtdIns( 4 ,5 )P2 (Rameh et aL, 1997b), suggesting that their 4-kinase activity may be restricted to 3- phosphate-containing Pis. More recently, PtdIns4P5K isoforms were reported to increase cellular PtdlnsPg levels when overexpressed in COS-7 cells (Halstead et aL, 2001). In particular, it appears that PtdIns4P5Ka is able to act as a Ptdlns( 3 ,4 ) ? 2 5-kinase in vivo and may be responsible for the increased PtdlnsP] levels in cells under oxidative stress. The PtdIns4P5K homologue in yeast, Mss4p, also has dual substrate specificity: besides converting PtdIns4P to PtdInsP 2 , it can phosphorylate PtdIns3P to form PtdIns( 3 ,4 )P2 (Desrivieres et aL, 1998).

Three mammalian PtdIns5P4K isoforms have been cloned, namely, a (53 kDa;

Boronenkov and Anderson, 1995; Divecha et aL, 1995), p (47 kDa; Castellino et aL,

1997), and y (47 kDa; Itoh et aL, 1998). Sequence comparison shows significant homology among PtdIns5P4Ks (>60% identity) and between PtdIns5P4Ks and PtdIns4P5Ks (-35% identity), which is predominantly localized to their kinase domain. In addition to PtdIns5P, PtdIns5P4Ks isoforms phosphorylate PtdIns3P, but not

PtdIns4P, to generate PtdIns( 3 ,4 )P2 (Rameh et aL, 1997b; Tolias et aL, 1998b).

Furthermore, PtdIns5P4Ka is unable to generate PtdlnsP] using PtdIns( 3 ,4 )P2 as substrate (Zhang et aL, 1997); the ability of PtdIns5P4Ka to produce PtdlnsP] from PtdIns3P in a concerted reaction appears to vary with reaction conditions (Rameh et aL, 1997b; Zhang et aL, 1997), Thus, at least in vitro, PtdIns5P4Ks are predominantly

PtdlnsP 4-kinases. Therefore, in addition to the canonical pathway of PtdInsP 2 synthesis via the PtdIns4P intermediate, there is the alternative pathway via the PtdIns5P intermediate. Since PtdIns5P is found in much lesser quantity than PtdIns4P in cells

(Rameh et aL, 1997b), most Ptdlns ? 2 in cells is likely to be produced via a PtdIns4P intermediate (Whiteford et aL, 1997). However, the wide expression pattern of PtdIns5P4Ks suggests that this alternative pathway may be important in several cell types (Boronenkov and Anderson, 1995; Divecha et aL, 1995; Hinchliffe et aL, 1996; Ling et aA, 1989).

46 Two PtdIns3P5K homologues have been cloned from yeast and mouse, and are termed Fab Ip and PlKfyve, respectively (Shisheva et aL, 1999; Yamamoto et aL, 1995). In yeast.

Fab Ip may be able to initiate a Ptdlns(3,5 )f 2-dependent signalling pathway (Odorizzi et aL, 2000). They are unique among PtdlnsfKs as each protein contains a FYVE domain (which specifically binds PtdIns3P in vitro) at the N-terminus. Thus, PtdIns3f5K may play a regulatory role as both an effector protein of Ptdlns3f and a kinase that terminates the recruitment and/or activation of other FYVE domain-containing proteins that bind PtdIns3P. PtdIns3P5K also contains a DEP (Dishevelled, Egl-10 and pleckstrin) domain, which was suggested to target proteins to specific membrane compartments (Koelle and Horvitz, 1996; Ponting and Bork, 1996; Rothbacher et aL, 2000), and a TCP-1 chaperonin homology domain, which predicts a role of PtdIns3f5K in non-covalent protein assembly (Ellis, 1998).

PlKfyve was also shown to have an intrinsic protein kinase activity that is inseparable from its lipid kinase activity (Sbrissa et aL, 2000). In addition to catalysing the formation of Ptdlns( 3 ,5 ) ? 2 and Ptdlns5f in vitro, PlKfyve can phosphorylate itself as well as other exogenous substrates on serine residue(s) in vitro. Autophosphorylation of PlKfyve down-regulates its in vitro lipid kinase activity and can be modulated by its lipid substrates Ptdlns and PtdIns3P. Thus, PlKfyve shows some similarity to the PI3K pi 10a and pi 10Ô subunits (Hunter, 1995).

Their primary sequences revealed that PtdlnsPKs are conserved from yeast to mammals - this led to proposals that they form a family distinct from other lipid kinases (Boronenkov and Anderson, 1995; Ishihara et aL, 1996; Loijens and Anderson, 1996). Despite their overall low sequence identity with other lipid and protein kinases, PtdlnsfKs possess specific conserved motifs and residues that are also found in other kinases as well as other ATP- and GTP-binding proteins. This is consistent with the ability of PtdlnsPKs to use both ATP and GTP as phosphate donors (Bazenet et aL, 1990). In the crystal structure of PtdlnsfK-Iip, the globular PtdlnsRK-Iip monomer has two domains with a deep cleft between them (Rao et aL, 1998). About 80 residues on both sides of the cleft can be superimposed on the ATP-binding and catalytic residues of protein kinase structures. The putative catalytic site is a shallow, open, binding pocket in which the phosphoinositol headgroup may be able to rotate freely such that Ptdlns3f or PtdIns5P can occupy the same binding site. Three catalytic site residues (Lys-150, Asp-278 and Asp-369) are fully conserved in all PtdlnsPKs and probably serve the same functions as their counterparts in protein kinases (Rao et aL, 1998). A flexible loop at the C-terminal region, which is the 47 topological counterpart to the activation loop of protein kinases, spans the catalytic site and was recently shown to determine both enzymatic specificity and subcellular targeting of PtdlnsfKs (Kunz et aL, 2000). In addition, a conserved glycine-rich loop (GA^GS), similar to the phosphate-binding loop {GXGXXG) found in protein kinases, is predicted to interact with the phosphotriester backbone of ATP, while the adenine moiety is thought to fit into a pocket containing conserved hydrophobic residues. Thus, there may be a conservation of catalytic mechanism between protein and lipid kinases (Rao et aL, 1998).

It is therefore not entirely surprising that, like PlKfyve, PtdIns4P5Ks also exhibit protein kinase activity in vitro and possibly in vivo (Itoh et aL, 2000). Autophosphorylation of PtdIns4R5Ks, on serine and threonine residues, is enhanced in vitro in the presence of Ptdlns (but not PtdOH or phosphorylated Ptdlns) and Mg^'^/Mn^'*', and strongly suppresses the lipid kinase activity. PtdIns4P5Ks were also found to be phosphorylated and down- regulated in resting cells (Itoh et aL, 2000). Thus, PtdIns4P5Ks are similar to class I PDKs and PtdIns4K-IIip, which also exhibit protein kinase activity (Hunter, 1995; Zhao et aL, 2000).

Two globular PtdlnsPK-IIp monomers interact at their N-termini to form a homodimer, which has an elongated, flat, disc-like structure (Rao et aL, 1998). One side of the disc is highly basic (net charge of + 1 0 ) and exceptionally flat, and probably functions as the interface for membrane association. In this orientation, the catalytic site on each monomer can access membrane lipids. The membrane-associating face is postulated to interact with phospholipid headgroups in a bilayer without penetration into hydrophobic membrane regions, suggesting that the enzyme docks at the membrane using only electrostatic interactions. Together, the flatness of the membrane-associating interface and the proximity of its shallow, substrate-binding pocket suggest that PtdlnsPKs phosphorylate their substrates in situ in membranes.

Despite the similarity in their kinase domains, the three PtdlnsPK families are biochemically and immunochemically distinct. For example, PtdIns4P5K, but not PtdIns5P4K, is able to phosphorylate PtdlnsP on isolated erythrocyte membranes in vitro (Bazenet et aL, 1990). PtdIns4P5K is stimulated by PtdOH, heparin and spermine, while PtdIns5P4K is inhibited or not affected by them, depending upon the assay conditions (Bazenet et aL, 1990; Jenkins et aL, 1994). Some PtdlnsPKs however cannot be readily assigned into any of the three families. For example, the recently-identified PtdlnsPK from Dictyostelium, denoted PIPkinA, is equally homologous to both PtdIn4P5Ks and

48 PtdIns5P4Ks, and the kinase domain from PtdIns4P5Kp and PtdIns5P4Kp are both able to substitute for the equivalent sequences in PIPkinA to effect phenotypic rescue (Guo et aL, 2001). PIPkinA also contains a DAG-binding domain, ankyrin repeats and an a- actinin homology region. Nevertheless, the different sensitivity of the various PtdlnsPKs to different cellular compounds may indicate that multiple regulatory pathways or mechanisms exist in vivo for the production of PtdlnsP].

1.5.4.2 Regulations and Functions

PtdlnsPKs have been found at the plasma membrane, ER, Golgi apparatus, cytoskeleton and in nuclei and the cytosol. For example, PtdIns4P5Ka has been found to localize at the plasma membrane or in the nucleus, while Mss4p was mostly found at the plasma membrane (Boronenkov et a l, 1998; Homma et aL, 1998; Jones et aL, 2000b; Shibasaki et aL, 1997). PtdIns5P4Ks were detected in the ER, actin cytoskeleton, plasma membrane, cytosol and nucleus (Boronenkov et aL, 1998; Hinchliffe et aL, 1996; Itoh et aL, 1998). In contrast, PtdIns3P5Ks are concentrated at intracellular membranes: Fab Ip is localized on the vacuolar membrane (Yamamoto et aL, 1995) and PlKfyve is mostly localized to multivesicular bodies (Shisheva et aL, 2001). Interestingly, a recent study using chimeric PtdlnsPKs suggested that PtdIns4P5Ks and PtdIns5P4Ks are primarily localized to the plasma membrane, and the nucleus and cytosol, respectively (Kunz et aL,

2000).

PtdlnsPKs are also re-distributed within cells following agonist treatment. For example, PtdIns4P5Ka was recently shown to translocate from the early TGN to the plasma membrane following stimulation of the G-protein-coupled thrombin receptor PARI in COS-7 cells (Chatah and Abrams, 2001). PtdIns5P4Kp appears to be a cytoplasmic enzyme that can bind directly to the juxtamembrane region of the p55 tumor necrosis factor (INF) receptor and may be responsible for increased PtdlnsP] levels in TNFa- treated HeLa cells (Castellino et aL, 1997). However, other studies showed that PtdIns5P4Kp is mostly localized to the nucleus (Boronenkov et aL, 1998; Ciruela et aL, 2000). Subcellular fractionation of 3T3-L1 adipocytes indicated that a portion of PlKfyve translocates from the cytosol to the low density microsomal fraction following stimulation with insulin (Shisheva et aL, 2001).

The localization to cellular membranes is in line with evidence indicating the participation of PtdlnsPKs and PtdInsP 2 in vesicle trafficking and secretion (de Camilli et

49 a l, 1996). In PCI2 cells, type I (but not type II) PtdInsPK was identified as one of three cytosolic proteins (termed PEP proteins) required for ATP-dependent priming of Ca^^- activated secretion of noradrenaline (Hay et a l, 1995). Type I PtdlnsfK is thought to act with the other PEP proteins, which include a Ptdlns transfer protein (Hay and Martin, 1993), to generate PtdlnsP] (Cunningham et a l, 1995), which is thought to bind effector proteins such as CAPS (calcium-dependent activator protein for secretion) (Loyet et al., 1998).

PtdIns4P5K also participates in the formation of coated transport vesicles in the Golgi that is regulated by the ARE family of small G proteins (Donaldson and Klausner, 1994). ARFs are directly and indirectly regulated by Ptdlnsf] (Brown et al., 1998; Paris et al., 1997; Terui et al., 1994), which can be produced by PtdIns4K-IIIp and a PtdIns4P5K that can be recruited by ARE to Golgi membranes independently of PLD activity (Godi et al., 1999; Godi et al., 1998). In separate studies, PtdIns4E5Ka was found to be a downstream effector of ARE I (Jones et al., 2000a) and ARE 6 (Honda et al., 1999). Ptdlnsf 2 thus plays a role as a co-factor in coat assembly because GTP-bound ARE triggers attachment of coat proteins. GTP-bound ARE can also activate PLD, for which Ptdlnsf] is a co­ factor (Brown et al., 1993; Cockcroft et al., 1994a; Divecha et al., 2000; Liscovitch et al., 1994; Pertile et al., 1995) and which interacts with PtdIns4P5Ka (Divecha et al., 2000). PLD cleaves PtdCho to generate PtdOH, which in turn stimulates PtdIns4E5K and consequently more PtdlnsE] production. The identification of these enzymatic properties led to the proposal that PLD and PtdIns4E5K can form a positive feedback loop via their respective products which leads to locally elevated amounts of Ptdlnsf] and PtdOH in vesicular membranes (Liscovitch et al., 1994). Cockcroft and co-workers further proposed that, at least in permeabilised HL60 cells, ARE activates PtdIns4E5K both directly and via stimulation of PtdOH-producing PLD (Skippen et al., 2002). At the same time, Ptdlns ? 2 and PtdOH also activate an ARE GTPase-activating protein (Brown et ah, 1998; Randazzo and Kahn, 1994), which may terminate this positive feedback loop by the hydrolysis of GTP to GDP. PLD would then be deactivated and disassembly of coat proteins initiated, thereby facilitating the subsequent fusion of the vesicle with its target membrane (Liscovitch and Cantley, 1995; Liscovitch et a l, 1994).

On the other hand, in streptolysin O (SLO)-permeabilised platelets, Ca^^-induced a- granule secretion was shown to be mediated, at least in part, by PtdInsP 2 that was synthesized via the PtdIns5P4K-dependent pathway (Rozenvayn and Elaumenhaft, 2001). This is in line with observations that PtdIns5P levels in platelets increased upon 50 stimulation with thrombin (Morris et al., 2000). Hence, current evidence suggests strongly that PtdlnsPKs play important regulatory roles in membrane trafficking and secretion.

In addition, PtdIns4P5Kp (but not PtdIns4P5Ka) was recently shown to be essential for endocytosis of the EGFR (Barbieri et al., 2001). Truncated or kinase-dead mutants of PtdIns4P5Kp disrupted the recruitment of the clathrin light chain and dynamin to the plasma membrane and inhibited coated vesicle formation in NR 6 cells. Furthermore,

PtdIns4P5Ky was identified as the major PtdInsP 2-synthesizing enzyme that plays a role in clathrin coat recruitment and actin polymerization at the endocytic zone of nerve terminals (Wenk et al., 2001). Recent reports also indicated that PIKIyve participates in the later steps of the endocytic and/or biosynthetic pathways (Shisheva et ah, 2001). In particular, it is the lipid kinase, but not the protein kinase, activity of PIKfyve that is crucial for endocytic membrane homeostasis (Ikonomov et al., 2002). The targeting of PIKfyve to the membranes of the late endocytic pathway is dependent on the FYVE domain, presumably via interacting with PtdIns3P (Sbrissa et al., 2002).

In yeast, the PtdIns3P5K homologue Fab Ip maintains normal vacuole function and morphology (Yamamoto et a l, 1995). Particularly, the PtdIns3f5K activity of Fab Ip is proposed to be required for selective cargo sorting within multivesicular bodies for transport to the vacuole lumen (Cooke et al., 1998; Gary et al., 1998; Odorizzi et al.,

1998) and is responsible for the in vivo production of Ptdlns( 3 ,5 ) ? 2 in yeast during osmotic stress (Cooke et al., 1998; Dove et al., 1997), which is dependent on the function of Vps34p, a yeast Ptdlns 3-kinase (PtdIns3K) that produces Ptdlns3f (Schu et a l, 1993). PIKfyve, but not PtdIns4P5Kp, is able to suppress the vacuolar defect and restore the basal Ptdlns( 3 ,5 ) ? 2 pool in Isfabl yeast mutant, supporting the notion that PIKfyve plays a role in membrane trafficking in mammalian cells (McEwen et al., 1999).

In addition, PtdIns4f5Ks play important roles in regulating actin filament formation. In this case, PtdlnsPK activity is apparently influenced by the Rho family of small G proteins. PtdIns4R5K activity in C3H 10T!4 cell lysate was stimulated by recombinant Rho A but not Racl; this effect appeared to be more potent in the presence of GTPyS (Chong et a l, 1994) and may be dependent on Rho-kinase (Oude Weemink et a l, 2000).

The same research group later found a 6 8 -kDa PtdIns4P5K from Swiss 3T3 cells associated with RhoA, apparently independent of whether RhoA was in a GTP- or GDP- bound state (Ren et a l, 1996). In another study, PtdIns4P5K activity in rat liver cytosol 51 associated with Racl but not RhoA, and the interaction was not dependent on Racl binding GTP (Tolias et a l, 1995). Similarly, PtdIns4R5K co-immunoprecipitated with Racl from serum-starved Swiss 3T3 cells and no increase in PtdlnsRK activity was detected when cells are treated with PDGF (Tolias et al., 1995). Furthermore, PtdIns4R5K and a DAGK in rat brain homogenate co-purified with recombinant Racl as a complex (Tolias et a l, 1998a). The Rac-associated PtdInsP5Ks were identified as PtdIns4P5Ka and PtdIns4P5Kp, but only the former was able to stimulate thrombin- initiated, Rac-dependent actin assembly (Tolias et a l, 2000). Specifically, Racl is able to induce EGF-initiated, ARF 6 -dependent translocation of PtdIns4R5Ka to, and production of Ptdlnsf] at, membrane ruffles (Honda et al., 1999). The association of these lipid kinases with Rac is enhanced by some phospholipids, including PtdIns4P and

Ptdlns( 3 ,4 )? 2 .

While the above conflicting results require reinvestigation using isozyme-specific reagents, a recent paper suggested that Rac, Rho and PtdIns4P5Ka are sequentially activated downstream of the G-protein-coupled thrombin receptor PARI (Chatah and Abrams, 2001). It is also noted that Rho can stimulate PLD (Exton, 1997; Mackay and Hall, 1998), which in turn can activate PtdIns4P5K in a positive feedback mechanism (see above). Thus, Rho and/or Rac may regulate actin filament formation via PtdlnsRKs. This is consistent with reports that overexpression of PtdIns4/*5Ks induces dramatic actin reorganization in COS cells and neuroblastoma NlE-115 cells, resulting in neurite retraction in the latter (Ishiharaet al., 1998; Shibasaki et al., 1997; van Horck et al., 2002; Yamazaki et al., 2002).

In yeast, the PtdIns4P5K homologue Mss4p controls actin-cytoskeletal organization (Desrivieres et ah, 1998; Homma et al., 1998). As alterations in Mss4p enzymatic activity resulted in major fluctuations of Ptdlns ? 2 levels, it was proposed that Mss4p may be the major PtdIns4P5K in S. cerevisiae under normal growth conditions, acting in the Stt4p/PLC/PKC pathway (Yoshida et a l, 1994b; Desrivieres et al., 1998). Unlike Fab Ip, Mss4p is essential for cell viability, but can be functionally replaced in vivo by murine type ip PtdlnsPK (Homma et al., 1998).

On the other hand, other studies point to a more prominent role for PtdlnsPg in regulating the actin cytoskeleton. For example, Cantley et al. (1991) reported that PI3K activation induces actin filament rearrangement, and cells with mutant PDGF P-receptors that lack PI3K-binding sites are unable to perform actin rearrangement and membrane ruffling in

52 response to PDGF (Kundra et a l, 1994; Wennstrom et al., 1994). Nevertheless, since

Ptdlnsf 2 is the primary substrate for PI3Ks, the roles suggested for PtdlnsPSKs and PBKs are not mutually exclusive and may even be complementary. It was thus speculated instead that PtdInsf5K may be regulating (rather than be regulated by) Rho and Rac (Hsuan et al., 1998), since PtdInsR5K can associate with these G proteins in their GDP- bound state (Ren et al., 1996; Tolias et a l, 1995). Indeed, Ptdlns ? 2 strongly stimulates GDP dissociation from RhoA and Racl (Zheng et al., 1996), possibly in a similar way to the Ptdlnsf 2 -stimulated exchange of GDP for GTP by ART (Terui et al., 1994).

The increased flux through PLC-mediated PI signalling pathways and the stimulation of PI kinase activities in agonist-treated cells prompted investigations into the possible association of PtdlnsRKs with receptors (Pike and Bakes, 1987; Walker and Pike, 1987). PtdlnsPK activity co-immunoprecipitated with the EGFR from B82L cells (Cochet et a l, 1991), while in NR6 cells, stimulation with EGF induced increased association between the EGFR and PtdIns4P5Kp (Barbieri et a l, 2001). In addition, PtdIns5R4Kp (but not the highly homologous PtdIns5P4Ka) associated with the p55 TNF receptor (Castellino et a l, 1997). In these cases, PtdlnsPK appeared to interact with the juxtamembrane region of the receptor. On the other hand, biochemical data suggests that PtdIns5P4Ka is likely to be the platelet PtdIns3P4K activity, which is stimulated by thrombin receptors and PKC (Divecha et a l, 1995; Hinchliffe et a l, 1996; Yamamoto et a l, 1990; Yamamoto and Lapetina, 1990).

The role played by receptors in the regulation of PtdlnsPKs however remains unclear. One possible mechanism is through phosphorylation of PtdlnsPK enzymes. For example, increased PtdlnsPK activity was detected in anti-pTyr immunoprecipitates from EGF- treated B82L cells (Cochet et a l, 1991) and increased PtdInsP 2 production by EGF- treated A431 plasma membranes could be delayed by exogenous phosphotyrosine phosphatase activity (Payrastre et a l, 1990). Specifically, PtdIns5P4Ky is phosphorylated following mitogenic stimulation (Itoh et a l, 1998). The enzymatic activity of PtdIns4P5Ka in NIH-3T3 cells is also regulated by phosphorylation: suppression via phosphorylation in resting cells, and activation via dephosphorylation in lysoPtdOH- treated cells (Park et a l, 2001). However, cyclic AMP-dependent protein kinase and protein phosphatase 1 were identified as the primary protein kinase and phosphatase, respectively.

53 In an interesting report, PtdIns5P4Ka has been found to be phosphorylated on serine and threonine (but not tyrosine) residues in unstimulated platelets (Hinchliffe et al., 1998). Thrombin treatment caused dephosphorylation of PtdIns5P4Ka and a corresponding two- to three-fold increase in PtdlnsPK activity, suggesting the existence of inhibitory phosphorylation site(s) on PtdIns5P4Ka. At the same time, PtdIns5P4Ka was also phosphorylated on at least two new residues, which contributed to an increase in activity. In Arabidopsis thaliana, the PtdlnsPK AtPIP5Kl, which is more similar to Mss4p than other PtdlnsPKs, can be phosphorylated by protein kinase A, leading to a 40% decrease in its catalytic activity (Westergren et al., 2001).

Another way by which receptors can regulate PtdlnsPK activity is through G proteins, including members of the ART and Rho/Rac families (see above). PtdIns4T5K activity is stimulated by GTPyS in rat brain membranes and HL60 promyelocytic cells (Cunningham et al., 1995; Smith and Chang, 1989). In neutrophils and partially-purified rat liver membrane preparations, PtdIns4P5K activity is also sensitive to pertussis and cholera toxins, respectively (Stephens et a l, 1993; Urumow and Wieland, 1988; Urumow and Wieland, 1990). More recently, it was suggested that the thrombin receptor PARI stimulates PtdIns4P5Ka via sequential activation of Gaq, Rac and Rho following treatment with agonist (Chatah and Abrams, 2001).

The stimulation of PtdIns4P5K by PtdOH may also allow a regulatory feedback mechanism to operate in PLC signalling. PLCyl has been found to be allosterically stimulated by PtdOH after being phosphorylated by the EGFR (Jones and Carpenter, 1993). PLC generates DAG, which is subsequently phosphorylated to PtdOH.

Consequently, PtdOH may then enhance Ptdlns ? 2 production and then hydrolysis, thereby forming a feedback loop by coupling Ptdlns ? 2 hydrolysis with Ptdlns ? 2 generation.

It remains unclear whether the substrate preferences of PtdlnsPKs can be altered by association with receptors or G proteins, or by phosphorylation. Given the lack of sequence homology outside each kinase domain, it is likely that the various PtdlnsPKs have different patterns of spatial and temporal regulation.

1.5.5 Phosphatidylinositol Transfer Protein (PtdlnsTP)

The conventional model for PLC-mediated PI signalling pathways assumes that the Ptdlns substrate for sequential phosphorylation by PtdIns4K and PtdIns4T5K resides 54 freely in the cytoplasmic leaflet of the plasma membrane lipid bilayer. The resultant Ptdlnsf] is subsequently cleaved by receptor-activated PLC to generate InsPs and DAG.

This model was challenged when it was observed that PI signalling in permeabilised, cytosol-depleted HL-60 cells was not restored by the addition of exogenous PLCp,

GTPyS and fMLP (Thomas et a l, 1991). This suggested that the PI pool in the plasma membrane may not be accessible to PLCp. GTPyS-dependent PLCp activity was restored when rat brain cytosol was also added to the cells, indicating that one or more additional cytosolic factors were required. PtdlnsTP was then identified as the major cytosolic factor that could reconstitute GTPyS-dependent PI signalling (Thomas et al., 1993).

PtdlnsTP is a member of a superfamily of lipid transfer proteins that transport lipids between membranes in vitro (Wirtz, 1991; Wirtz, 1997). The superfamily members differ considerably in size and amino acid sequence and can be divided into three main classes, based on their in vitro lipid-transporting specificity: (i) non-specific lipid transfer protein (nsL-TP), (ii) PtdCho transfer protein (PtdChoTP), and (iii) PtdlnsTP. The nsL-TP class transfers a wide range of lipids, including Ptdlns, PtdCho, phosphatidylserine, PtdEtn, PtdOH, sphingomyelin (SM) and cholesterol, while the PtdChoTP class transfers only PtdCho. As its name implies, the PtdlnsTP class prefers Ptdlns, but can also transfer PtdCho and, to a lesser extent, SM and phosphatidylglycerol.

PtdlnsTPs have been found in most, if not all, eukaryotic organisms studied. Anti-bovine PtdlnsTP antibodies recognise a 35-36-kDa cytosolic protein in mammals, birds, reptiles, amphibians and insects (Dickeson et al., 1989). Four isoforms of cytosolic PtdlnsTPs have been identified: three (PITPa, PITPp and rdgBp) from higher eukaryotes and one (Sec14) originally from yeast Saccharomyces cerevisiae. The larger, membrane-bound RdgBs were first discovered in Drosophila melanogaster, but homologues have since been cloned in mammals (Fig. 1.4).

1.5.5.1 PITP

PITPa has been cloned from several mammals and is a protein of 270-271 residues, with an apparent molecular mass of 35 kDa in SDS-PAGE (Dickeson et al., 1994; Dickeson et al., 1989). PITPp was discovered via its ability to complement seel4 mutations in yeast

(Tanaka and Hosaka, 1994). Its molecular size is similar to that of PITPa, with which it

55 PITPa

PITPp

DrdgB T-K 4-1— k

Nir3 4— H 4 - k

MrdgBp

Seel 4

Sshl

Ssh2

Figure 1.4 Structure of PtdlnsTPs A schematic diagram of domain topology of PITP, RdgB and Sec 14 homologues. The PITP domain (red) is conserved in various isoforms of PITP and RdgB. DrdgB and the Nir proteins also have two and one acidic domain respectively (white), six putative transmembrane regions (vertical lines) and the PYK2-binding domain (striped) at the C-terminus. The Sec 14 protein shares sequence similarity with the Ssh proteins but not the PITPs or RdgBs. shares 70% sequence identity. In addition to Ptdlns and PtdCho, both PITP^ isoforms can transfer SM between membranes in vitro, albeit less actively (de Vries et al., 1995; Li et al., 2002; Segui et al., 2002), and PITPp may be playing an important role in SM metabolism (van Tiel et al., 2000a). Due to the single negative charge difference between Ptdlns and PtdCho, PITPa purified from mammalian sources occurs in two forms with slightly different isoelectric points (Wirtz, 1991; see also Chapter 5).

The crystal structures of phospholipid-free PITPa (‘apo’ state) and PITPa in complex with PtdCho (‘holo’ state) were elucidated recently (Schouten et al., 2002; Yoder et al., 2001). Holo-PlTPa has a globular, single-domain structure that comprises a large concave p-sheet and seven a-helices (Fig. 1.5), while apo-PlTPa has a more ‘open’ conformation. They differ from the structure of Secl4p (Sha et al., 1998), even though PITPs and Secl4p have similar biochemical properties and demonstrate partial genetic complementation. Interestingly, PITPa shares significant 3-D structural similarity with the START (StAR-related lipid transfer) domain, which contains a hydrophobic chamber formed from a 9-stranded p-sheet and two a-helices (Ponting and Aravind, 1999). START domain-containing proteins are proposed to bind lipids and interact with membranes, but do not share protein sequence homology with PITPs.

The holo-PlTPa crystal structure showed the PtdCho molecule to be bound in a cavity formed by the p-sheet and two helices (A and F). This lipid-binding core is highly conserved among PITPs, and many of the conserved amino acid residues in the core interact with the bound PtdCho molecule. Wirtz and co-workers further proposed the presence of (i) a lipid exchange loop that acts as a lid to the lipid-binding core and (ii) two separate, adjacent sites for binding the phosphorylinositol and phosphorylcholine headgroups in the more open ‘apo’ conformation. The crystal structure also confirmed earlier observations that PITPa has specific binding sites for both acyl chains of the PtdCho or Ptdlns and that each binding site has a distinct preference for acyl chain length (van Paridon et a l, 1988a; van Paridon et a l, 1987b). The cavity within the lipid-binding core around the sn-2 acyl chain is more confining than that around the sn-\ acyl chain (Yoder et al., 2001). This could perhaps explain the broader range of acyl moieties that can be accommodated in the sn-\ acyl region of the cavity.

^ PITPa, PITPP and rdgBP will hereinafter be referred together as the PITP sub-class within the PtdlnsTP family of lipid transfer proteins. 57 Figure 1.5 Ribbon Diagram of the PITPa Structure in Complex with PtdCho The lipid-binding core, the regulatory loop and the C-terminal region are coloured blue, green and red, respectively. The (3-sheets are numbered 1-8, and the a-helices labelled A-G. The PtdCho molecule is depicted as space-filled spheres (carbon, grey; oxygen, orange; phosphorus, yellow). Schouten et al. (2002) proposed that helix B constitutes a distinct functional region, the lipid exchange loop, which acts as a lid to the lipid binding core. This figure is reproduced from Yoder et al. (2001). 58 It was also proposed that the extensive loop region between p-strands 7 and 8 has regulatory functions in the association and interaction of PITPa with lipid- and protein- modifying enzymes, such as PI kinases (Yoder et al., 2001). This regulatory loop is found in all RdgBs and PITPs but not Secl4ps, and contains Ser-166, which was recently identified as a PKC-dependent phosphorylation site (van Tiel et a l, 2000b). The residue equivalent to Ser-166 is however absent in PITPs from C elegans.

Previous studies of PITPa which was truncated at various C-terminal residues indicated that the C-terminal region (CTR) helps PITPa stabilize a tight, compact native structure (Voziyan et al., 1996) and is critical for lipid transfer by modulating the affinity of full- length PITPa for membranes, particularly those that contain acidic phospholipids (Tremblay et al., 1996; Tremblay et al., 1998). Yoder et al. (2001) postulated that the CTR participates in membrane association and dissociation, and proposed a model by which PITPa interacts with membranes and exchanges its bound phospholipid: the CTR is displaced from the lipid-binding core on contact with membrane surface, such that the bound phospholipid is exposed, absorbed into the membrane and replaced by a membrane lipid. At the same time, the displaced CTR becomes more flexible such that helix G (which constitutes part of the CTR) is susceptible to proteolysis (Tremblay et al., 1996). Dissociation of the protein from the membrane is thought to be driven by the refolding and reassociation of the CTR with the rest of the protein. This may explain why PITPa that is truncated at the C-terminus dissociates more slowly from membranes and is inefficient in lipid transfer (Hara et al., 1997; Prosser et al., 1997). The above model is in agreement with the displacement and partial unfolding of the CTR in the apo-PITPa crystal structure, which is proposed to resemble a membrane-bound state of PITP (Schouten et a l, 2002).

1.5.5.2 RdgB

The Drosophila PtdlnsTP, RdgB, is encoded by the retinal degeneration-B gene which, when mutated, causes an abnormal photoreceptor response and light-enhanced retinal degeneration. Genetic and biochemical data suggest that rdgB mutations disrupt an aspect of the visual transduction cascade, which is mediated by PLC hydrolysis of PtdInsP 2 - Mutations in the ninaE (rhodopsin), norpA (PLC) and trp (light-activated Ca^^ channel) suppress rd^R-dependent photoreceptor degeneration (Zuker, 1996). In photoreceptor cells, RdgB localizes to the subrhabdomeric cistemae (SRC), which are extensions of the ER that lie adjacent to the rhabdomeres (Suzuki and Hirosawa, 1994; Vihtelic et al., 59 1993). In addition to photoreceptors, RdgB is expressed in the antennae and regions of the brain, indicating that it functions in other sensory and neuronal cells. RdgB is unique among PtdlnsTPs as it is a probable transmembrane protein. It has six putative transmembrane domains, and the N- and C-termini have been proposed to lie in the cytosol between the SRC and rhabdomere (Vihtelic et a l, 1993; Vihtelic et al., 1991). The PITP-homologous domain at the N-terminus is about 280 residues long and 42% identical to rat PITPa (Vihtelic et al., 1993).

Mammalian RdgB homologues were cloned recently and found to be highly similar to the fruit fly protein, with highest homology occurring at the N-terminal PITP domain and a region near the C-terminus (Aikawa et al., 1997; Chang et al., 1997; Guo and Yu, 1997; Lu et al., 1999a). Like their counterpart in the fruit fly, both mammalian isoforms, MrdgBl and MrdgB2, are primarily expressed in neuronal tissues in the mouse, such as retina and brain, but their different spatial expression patterns suggest that they perform different, non-redundant roles (Chang et al., 1997; Guo and Yu, 1997; Lu et al., 1999a). In particular, MrdgBl may play a crucial role in mouse brain development (Aikawa et al., 1997). Interestingly, the mouse mrdgBl gene maps to the same location on chromosome 19 as Mvb-1, which encodes a major suppressor of the mouse vibrator (PITPa) mutation (Aikawa et al., 1997; Hamilton et al., 1997). The mrdgB2 gene is on the distal end of mouse chromosome 5 that is syntenic to human chromosome 7pl2-q22, in which one form of retinitis pigmentosa and autism were mapped (Lu et al., 1999a). The humanrdgB gene maps to 1 lql3.1, a locus where several retinal diseases have also been mapped (Chang etal., 1997).

Human orthologues of MrdgBl and MrdgB2 and a third, novel human RdgB homologue were shown to bind the N-terminal domain of PYK2, a protein tyrosine kinase, and thus designated Nir (PYK2 N-terminal domain-interacting receptor) proteins (Lev et al., 1999). They interacted and co-immunoprecipitated with PYK2 via their C-terminal domains. Protein sequence comparisons indicate that MrdgBl and MrdgB2 are orthologues of the Nir2 and Nir3 proteins, respectively. Interestingly, Nirl does not contain a PITP domain while a Nir3 splice variant has a truncated PITP domain. Immunohistochemical analysis detected expression of the Nir proteins in different retinal cell types, suggesting different functions for Nir proteins (Lev et al., 1999). Both Nir and PYK2 proteins are tyrosine phosphorylated in cells that are treated with agonists, including phorbol myristate acetate.

60 Another novel human RdgB protein was recently cloned (Fullwood et a l, 1999). MrdgBp (also denoted as rdgBp) is comprised of only a N-terminal PITP domain and a short C- terminal domain, but unlike typical RdgB proteins, it lacks the transmembrane motifs and conserved C-terminal domain. MrdgBp is ubiquitously expressed and exhibits Ptdlns transfer activity that is comparable to PITPs.

Although both MrdgBl and MrdgB2 isoforms suppress retinal degeneration in rdgB^ null mutant flies, only MrdgBl was able to fully restore the electrophysiological light response (Chang et al., 1997; Lu et al., 1999a). This points to some underlying functional similarity between invertebrate and mammalian photoreceptors and suggests that the Ptdlns cycle also participates in the physiology of the latter. While MrdgBl was shown to be an integral membrane protein as predicted (Aikawa et ah, 1999), MrdgB2 surprisingly does not behave as an integral membrane protein in protein extraction studies, although it also possesses the putative transmembrane domains that are common features of RdgB proteins. Instead, MrdgB2 may be associated with cytoskeletal complexes in vivo (Lu et a l, 1999a).

When expressed as a soluble protein, the RdgB-PITP domain displayed Ptdlns/PtdCho transfer activity in vitro and its ectopic expression was apparently sufficient to suppress both retinal degeneration and the abnormal light response electroretinogram in rdgB^ null mutants, suggesting that one or more of the functions of RdgB required during phototransduction reside in its PITP domain (Milligan et a l, 1997; Vihtelic et a l, 1993). However, substitution of the RdgB-PITP domain with rat PITPa did not restore a wild- type phenotype to rdgB^ flies, showing that other PtdlnsTPs exhibiting Ptdlns/PtdCho transfer activity in vitro cannot replace RdgB functions in vivo, and that the phospholipid transfer activity of RdgB is not functionally sufficient in vivo (Milligan et a l, 1997).

RdgB also possesses a Ca^^-binding domain near its N-terminus, suggesting that Ca^^ is involved in RdgB function (Vihtelic et a l, 1993). This is consistent with findings that (i) voltage-gated Ca^^-channel blockers inhibit rJgR-mediated retinal degeneration (Sahly et a l, 1992), and (ii) loss of one of the light-activated Ca^^-channels (trp^^) significantly slowed retinal degeneration in rdgB flies (Paetkau et a l, 1999).

1.5.5.3 Secl4

The Sec 14 proteins are in a distinct subclass of their own and share no apparent sequence homology with PITPa, PITPp and RdgB (Wirtz, 1991). Instead, they are homologous to 61 mammalian retinaldehyde- and a-tocopherol-binding proteins and the non-catalytic domain of the human MEG2 protein tyrosine phosphatase (Gu et a l, 1992; Salama et a l, 1990; Sato et a l, 1993). Nevertheless, the molecular mass (35 kDa) and p / (5.3) of Secl4p from S. cerevisiae are similar to PITPa (Aitken et aL, 1990; Bankaitis et al.,

1989). S. cerevisiae Secl4p is able to substitute for PITPa and PITPp in many functional assays (see below) and PITPa can partially rescue sec 14 mutants (Skinner et a l, 1993). Recently, five yeast proteins with modest sequence homology to Secl4p were found (Phillips et a l, 1999). Some of these SFH (Sec Fourteen Homologue) proteins were able to transfer Ptdlns and/or complement seel 4 defects in yeast, but none was able to transfer PtdCho (Li et al., 2000a; Phillips et ah, 1999). Although SFH proteins are not individually or collectively essential for cell viability, they are required for regulation of PLD activity in Sec 14p-independent cell growth (Li et al., 2000a; see also below). Interestingly, the SFH4 gene is identical to the PSTB2/PDR17 gene, which appeared to play a role in lipid transport processes in phosphatidylserine metabolism (Wu et al.,

2000).

Sec 14 proteins of significant sequence similarity are also found in other yeasts, including Yarrowia lipolytica, Schizosaccharomyces pombe, Kluyveromyces lactis, Candida albicans and Candida glabarata. However, these proteins may have distinct physiological functions. For example, the sec 14 gene in S. cerevisiae is required for cell viability and the export of secretory proteins from the Golgi, while the Y. lipolytica sec 14 gene is necessary for differentiation from a yeast to a mycelial mode of growth (Lopez et al., 1994).

Recently, SecMp homologues have been cloned in Arabidopsis thaliana and soybean by functional complementation of temperature-sensitive sec 14 mutants (Jouannic et al., 1998; Kearns et al., 1998b). They share 20-25% sequence identity with Secl4p in yeasts. In A. thaliana, the AtSEC14 gene product is expressed in all tissues examined and has strong Ptdlns but weak PtdCho transfer activity (Jouannic et al., 1998). In soybean, the two SecMp homologues, denoted Sshlp and Ssh2p, exhibit novel biochemical properties. Unlike previously characterized PtdlnsTPs, Sshlp is unable to transfer Ptdlns or PtdCho in vitro, while Ssh2p is capable of transferring Ptdlns but not PtdCho (Kearns et a l, 1998b). In vitro binding assays further suggested that Sshlp and Ssh2p may bind Ptdlns?]. In addition, they appear to perform different physiological functions: Sshlp is expressed mostly in roots and leaves, while Ssh2p expression is highest in developing

62 seeds. In particular, Sshlp probably participates in a stress response pathway in plants that are exposed to an osmotic insult.

The crystal structure of SecMp appears to be structurally unique (Sha et aL, 1998). It consists of two domains held together by hydrophobic stacking interactions and has twelve a-helices, six p-strands and eight 3]o-helices. The phospholipid-binding site on the C-terminal domain is a hydrophobic pocket, which is stabilised by a string-like motif behind the p-sheet pocket floor. Three conserved residues (Lys- 6 6 , Glu-207 and Lys-239) that facilitate specific inositol headgroup binding were also identified in the pocket

(Phillips et aL, 1999). For example, the simultaneous mutation of Lys - 6 6 and Lys-239 to alanine abolished the binding and transfer of Ptdlns, but not PtdCho. Furthermore, the surface helix A10/T4 (a-helix AlO; 3io-helix T4), which protrudes away from the mouth of the binding site, has a hydrophobic face and presumably makes extensive van der Waals interactions with phospholipid acyl chains. Thus, it was postulated that when SecMp binds to a membrane, the A10/T4 helix swings away from the protein and inserts into the lipid bilayer. At the same time, the lipid on SecMp is deposited into the bilayer. As the A10/T4 helix retracts from the membrane, it acts as a “bulldozer” and abstracts another lipid molecule into the unoccupied hydrophobic pocket (Sha et aL, 1998).

1.5.5.4 Functions

Until almost ten years ago, the accepted role of PtdlnsTPs was primarily to transport Ptdlns from its site of synthesis (mostly in the ER) to other membrane compartments in the cell, since Ptdlns is not freely diffusible in the cytosol (Wirtz, 1991). Its characteristic difference in affinity for Ptdlns and PtdCho would help to maintain a distinct Ptdlns/PtdCho ratio in the membranes with which it interacted. If the ratio changed due to, for example, membrane biogenesis or repair or the consumption of Ptdlns during cell stimulation, it could be restored by PtdlnsTP-mediated mass action, that is, by predominantly inserting Ptdlns into a membrane in exchange for PtdCho, or vice versa (Wirtz, 1991).

In 1990, the SEC 14 gene was discovered to be identical to the PIT! gene, which encodes a PtdlnsTP in S. cerevisiae (Bankaitis et aL, 1990). Later studies postulated that SecMp uses its Ptdlns/PtdCho-binding ability to maintain an essential pool of DAG in the Golgi to support production of secretory vesicles (Kearns et aL, 1997) (Fig. 1.6). PtdCho-bound SecMp (but not Ptdlns-bound SecMp) inhibits choline-phosphate cytidylyltransferase

63 Cho ATP

CKIase ADP Secl4p-PI Secl4p-PC Cho-P

CTP CDP-Cho

( + )

CDP CCTase CPTase Golgi Ptdlns DAG PtdCho membranes metabolism

SL metabolism PtdOH metabolism Golgi secretory function

Fig. 1.6 Proposed Model of Secl4 Function at the Golgi SecMp is proposed to promote the secretory function of the Golgi apparatus in yeast through maintaining the integrity of a crucial DAG pool by regulating its production and consumption. The PtdCho-bound form of SecMp (SecMp-PC) feedback inhibits the activity of the CDP-choline (CDP-Cho) pathway for PtdCho biosynthesis, while the Ptdlns-bound form (SecMp-PI) might potentiate Ptdlns metabolism that resupplies the Golgi DAG pool. Other metabolic pathways coupled to inositol sphingolipid (SL) and PtdOH might also help resupply the DAG pool. CCTase, choline-phosphate cytidylyltransferase; CPTase, choline phosphotransferase; CKIase, choline kinase; Cho, choline; Cho-P, phosphocholine. 2 (CCTase), the rate-determining enzyme in the cytosine 5’-diphosphate (CDP)-choline pathway for PtdCho biosynthesis (Skinner et a l, 1995). This inhibition represses depletion of DAG, a substrate in the CDP-choline pathway. On the other hand, Ptdlns- bound SecMp is thought to increase turnover of Ptdlns and generate DAG as a by­ product (Kearns et aL, 1997). Thus, it has been hypothesised that SecMp regulates, via a dual mechanism of independent yet complementary functions, the production and consumption of DAG by different phospholipid metabolic pathways so that the level of DAG at the Golgi is preserved in order to maintain its secretory competence (Martin, 1997). At the same time, the Ptdlns/PtdCho ratio at the Golgi is also modulated, albeit indirectly, by regulating metabolic pathways involving the two phospholipids (McGee et aL, 1994). At this stage, it is still unclear what role DAG plays in vesicle budding at the yeast Golgi, but in mammalian cells, DAG is essential for recruiting protein kinase D to the TGN membranes during the early stages of vesicle biogenesis (Baron and Malhotra,

2002).

However, data from several recent papers indicate that accumulation of PtdIns4P, but not DAG, is required for bypass of growth and secretory phenotypes in seel4 cells by mutations in the SACl gene, which encodes an integral membrane protein with intrinsic PI phosphatase activity (Guo et aL, 1999; Hama et aL, 1999; Stock et aL, 1999). Similar levels of DAG were also found in other bypass mutants of secI4-l^\ in which the CDP- choline pathway is blocked (Sreenivas et aL, 1998). The PtdIns4P level decreased significantly in S. cerevisiae when secl4-3‘^ mutants were subjected to a non-permissive temperature, suggesting that SecMp-mediated secretion defects may be partly caused by limited amounts of Ptdlns4f (Hama et aL, 1999). It was thus proposed that Ptdlns4f, instead of DAG, might have a key role in constitutive secretion in yeast. Indeed, overexpression of Piklp (but not Stt4p or Vps34p) or loss of Sac Ip suppressed growth and secretion defects of seel4 mutants, apparently by increasing cellular level of PtdIns4P (Hama et aL, 1999; Stock et aL, 1999). The above data appear contradictory, but it may be explained if the increased PtdIns4P level in see 14 mutants expands the pool of DAG directly through PLC-mediated hydrolysis of Ptdlns4f or indirectly through influencing other PI metabolic pathways that lead to increased DAG production. For example, PLD activity (which hydrolyses PtdCho to produce PtdOH) is required for the bypass of seel 4 phenotype by Sac Ip or via the CDP-choline pathway (Sreenivas et aL, 1998; Xie et aL, 1998), and PtdOH can be dephosphorylated to generate DAG. Nevertheless, the above data also indicate that SecMp may be coupled to the action of

65 Piklp, as both seel4 and p ikl mutants are defective in the Golgi-to-plasma membrane phase of secretion (Hama et aL, 1999; Salama et aL, 1990).

PITPa and PITPp are able to rescue only temperature-sensitive (but not null) secl4 mutations (Skinner et aL, 1993), suggesting incomplete functional complementation. Indeed, not only did PITPa not inhibit CCTase activity, but instead modestly stimulates it. It is also unable to associate stably with yeast Golgi membranes (Skinner et aL, 1993). This suggests that phospholipid transfer activity is required but not sufficient to fulfil the functions of SecMp (Skinner et aL, 1993). However, recent data indicated that the Ptdlns binding and transfer ability of SecMp is dispensable for its function in vivo (Phillips et aL, 1999). The SecMp^^^’^^^^ double mutant exhibits about 60% of the wildtype PtdCho transfer ability but no detectable Ptdlns transfer activity, and is unable to stimulate Ptdlns4f production in vivo. Surprisingly, SecMp^^^'^^^^ was nonetheless able to efficiently complement both seel 4 temperature-sensitive and null yeast mutants (Phillips et aL, 1999). This led to the conclusion that PtdCho binding and/or transfer presumably constitutes an important facet of SecMp function. Hence, it is unclear how PITPa and

PITPp rescue see 14 mutants since inhibition of CCTase (and thus PtdCho synthesis) and association with the Golgi are key features of SecMp in the model proposed by Bankaitis (Li et aL, 2000b).

Mammalian PITPs have also been found to be involved in cellular secretion and vesicular transport. PITPa was identified as one of three cytosolic factors that are required for ATP-dependent priming of Ca^^-activated secretion of noradrenaline (Hay and Martin, 1993). Type I PtdlnsPK, also identified as a factor (Hay et aL, 1995), and a PtdIns4K on chromaffin granules (Wiedemann et aL, 1996) were also required for priming of exocytosis. In addition, both PITPa and PITPp isoforms are individually able to stimulate ATP-dependent formation of constitutive secretory vesicles and immature secretory granules from the TGN in vitro (Ohashi et aL, 1995). Moreover, PITPa acts synergistically with a novel PI3K in a hepatocyte-derived, cell-free system to stimulate formation of exocytic vesicles containing the poIy(IgA) receptor from the TGN (Jones et aL, 1998). Interestingly, Simon et al. reported that PITPa, probably in its Ptdlns-bound form, can trigger ATP-independent scission of COPI-coated vesicles from the TGN and that production of Ptdlns? and Ptdlns?] is not obligatory for such vesicle release. Instead, it was proposed that PITP inserts Ptdlns near the neck of coated buds to induce local changes in the organization of the lipid bilayer, which in turn triggers fusion of

66 ectoplasmic faces of the Golgi membrane and brings about scission of the COPI-coated vesicles (Simon et aL, 1998). In these experiments, SecMp was able to substitute for PITPs in forming secretory vesicles and priming secretion. In addition, PITPa can reconstitute cis-to-medial intra-Golgi vesicular transport (Paul et aL, 1998). These reports support the idea that PtdlnsTPs play a highly conserved role in cellular secretion from yeast to mammals but probably via different mechanisms (see above).

RdgB proteins may also be involved in secretion. In rat retinal photoreceptor cells, MrdgBl is primarily localized to the inner segment, suggesting that it may play a role in membrane sorting and trafficking and in neurotransmitter release (Guo and Yu, 1997). In addition, the N-terminal PITP domain formed a complex with a type III PtdIns4K in the Golgi, hinting that it may be involved in PI synthesis (Aikawa et aL, 1999).

As mentioned above, PITPa was shown to be required for reconstituting agonist- stimulated PLC signalling in permeabilised cells. In permeabilised HL60 cells, PITPa is essential for restoring fMLP- or GTPyS-dependent PLCp signalling (Thomas et aL,

1993), Ca^^-activated Insfg production by PLCÔ1 (Allen et aL, 1997), GTPyS-dependent

Ptdlns ? 2 synthesis (Cunningham et aL, 1995), and fMLP-dependent PI 3-kinase signalling (Kular et aL, 1997). In EGF-treated A431 cells, PITPa is also required for

Ptdlns phosphorylation and Insfg production by PLCyl (Kauffmann-Zeh et aL, 1995).

The requirement for PITPa in PI signalling via PLCô, PLCp, PLCy, and PI 3-kinase suggests that PITPa may be functioning at a common point in these differentially regulated signalling pathways. PITPa could be transporting Ptdlns down a concentration gradient from one or more intracellular sites (where Ptdlns is synthesized) to replenish the agonist-responsive Ptdlns pool, which is depleted by increased production and consumption of Ptdlns? and Ptdlns? 2 - This is perhaps achieved via specific recognition of PI kinase-associated receptors by PITP - the ‘targeted replenishment model’ of PITP function (Batty et aL, 1998; Currie et aL, 1997). This would help explain the observation that the amount of Ins ? 3 generated can exceed resting levels of Ptdlns ? 2 and that Ptdlns ? 2 concentration decreases only transiently during agonist stimulation (Cunningham et aL, 1995). Thus, sustained PLC activity is apparently dependent on the ability of PITPa to stimulate Ptdlns ? 2 synthesis in pools which serve as substrate for receptor-stimulated PLC. However, experiments demonstrating an acute requirement for PITPa led to the proposal of the ‘co-factor model’ of PITP function, whereby PITP may be regulating PI signalling by presenting the Ptdlns substrate to PtdIns4K, and probably sequentially to

67 PtdInsP5K and PLC as well (Cunningham et aL, 1995; Hsuan, 1993; Liscovitch and Cantley, 1995).

PITPs may be acting as a co-factor in Ptdlns ? 2 synthesis and hydrolysis via two mechanisms (Fig. 1.7). Ptdlns could be presented directly to PtdIns4K (or in the vicinity of the complex) for phosphorylation and then channelled through the enzyme complex containing PtdIns4K, PtdlnsfK and PLC. Alternatively, Ptdlns could remain bound on the PITP as it is sequentially phosphorylated by PI kinases, and the resultant PITP-bound

Ptdlns ? 2 is then cleaved by PLC. A common feature of both mechanisms is that PITP- bound Ptdlns is the preferred substrate for the enzymes in the PI signalling complex. Available data appear to support the latter mechanism. For example, PITPa is able to bind strongly but unable to transport or release PPIs (Schermoly and Helmkamp, 1983; van Paridon et aL, 1988b). More significantly, PITPa was found to associate, in an EGF- dependent manner, with EGFR, type II PtdIns4K and PLCyl in A431 cells (Kauffmann-

Zeh et aL, 1995), suggesting that PITPa may be a component of a multi-enzyme PI signalling complex assembled on the EGFR.

The RdgB proteins have also been suggested to participate in PLC-mediated phototransduction cascade in the Drosophila eye. As mentioned above, RdgB is localised to the SRC, which is in close proximity (10 nm apart) to the rhabdomere (Suzuki and Hirosawa, 1994) that contains rhodopsin and the phototransduction machinery (Zuker, 1996). This gap may be sufficiently narrow for the PITP domain to transfer Ptdlns between the two membranes (Hsuan and Cockcroft, 2001; Hsuan and Tan, 1997; Suzuki and Hirosawa, 1994; Vihtelic et aL, 1993). It is also possible that RdgB presents Ptdlns to PI metabolising enzymes on the rhabdomere. However, as recent reports have shown that RdgB proteins may not be integral membrane proteins (Litvak et aL, 2002; Lu et aL, 1999a), the biochemical role of RdgBs in PI signalling (or other metabolic pathways) remains to be determined.

1.6 Compartmentation of Signalling

In recent years, there is an increasing consensus that signal transduction reactions do not occur freely in the cytosol but are organized into architecturally and spatially discrete complexes or domains of signalling (Monaco and Gershengom, 1992; Pawson and Scott, 1997). Specific signalling molecules are recruited into different transduction complexes.

68 A PIP PIP, DAG PIP5K PI4K PLC or

PITP PITP

PC

B

PC DAG PIP5K PI4K PLC

IPs PITP PITP

PC PITP PITP

PIP PIP2

Fig. 1.7 Two Alternative Models for the Co-Factor Function of PITP during PtdlnsjPi Synthesis (A) In this model, Ptdlns (PI) is either delivered directly to PtdIns4K (PI4K) or to the vicinity of the signalling complex and subsequently channelled through the complex. (B) In the second model, Ptdlns remains bound on PITP through the sequential phosphorylation and Ptdlns(4,5)?2 (PIP2 ) bound to PITP is the preferred substrate for PLC. PIP5K, phosphatidylinositol phosphate 5-kinase; PC, phosphatidylcholine; IPs, inositol-1,4,5-trisphosphate. 69 so that cellular responses are fine-tuned and optimised, thereby enhancing specificity, selectivity and speed while preventing unintended cross-talk between different pathways.

A good example is the protein InaD, which has five PDZ domains and regulates phototransduction in D. melanogaster photoreceptor cells (Shieh and Zhu, 1996), Using distinct PDZ domains, InaD interacts with three components of the light-triggered PI cascade for phototransduction in the fruit fly visual system: the transient receptor potential (TRP) Ca^^ channel, PLCp and PKC. InaD apparently acts as a scaffolding protein to assemble these proteins into a signalling complex so that (i) efficient activation of the TRP channel by PLCp in response to rhodopsin stimulation and (ii) deactivation of TRP through phosphorylation by PKC can occur (Tsunoda et aL, 1997). InaD mutants defective in a particular PDZ domain produced incomplete complexes lacking the appropriate target protein and displayed corresponding defects in their physiology (Tsunoda et aL, 1997). In null mutants, signalling complexes were not formed, the three signalling molecules were mislocalized, and their protein levels markedly reduced. Thus, InaD is a modular, multivalent adaptor protein that not only links multiple signalling components within the same cascade but is also required for their stability and subcellular localization. These studies also demonstrated that it is the location of a transduction molecule in the cell and not its presence per se that promotes effective signalling (Scott and Zuker, 1998; Tsunoda et aL, 1997).

An important implication of the above-mentioned ‘co-factor model’ is that all intermediates in the PI signalling pathway are isolated from their counterparts in the plasma membrane, resulting in metabolic compartmentation. This could help explain the empirical phenomenon of agonist-responsive and -unresponsive pools of cellular Ptdlns (Monaco and Gershengom, 1992). A pool is defined as a population of PPIs which can be distinguished biochemically following cell-labelling experiments or fractionation of cell lysates. Different lipid pools do not necessarily represent synthesis or localization at different intracellular membranes but pools may be on the same membrane and possibly sequestered by membrane proteins.

The agonist-responsive Ptdlns pool is the fraction of total cellular Ptdlns that can be hydrolysed in response to receptor agonist, while the agonist-unresponsive pool is apparently resistant to hydrolysis, even after prolonged stimulation. It is presumed that distinct Ptdlns pools serve different cellular functions and may employ different PI kinase isoforms to mediate and regulate their metabolism. For example, overexpressing the yeast

70 PtdlnsPK Mss4p did not completely suppress the G 2/M boundary blockage caused by mutation of stt4, a yeast PtdIns4K gene (Yoshida et aL, 1994b). This suggests that there may be a branch point in the signalling pathway between Stt4p and Mss4p and/or that

Stt4p-dependent G 2/M cell progression uses a Ptdlns4f pool different from that used by Mss4p.

The existence of distinct Ptdlns pools will also require segregation of the remaining intermediates within the Ptdlns cycle. The DAGK a-isozyme has been found to associate with the activated EGFR (Schaap et aL, 1993) and may be a component of a PI signalling complex associated with the EGFR (Payrastre et aL, 1991). DAG molecules which were generated randomly in the plasma membrane by an exogenous bacterial Ptdlns-specific PLC were not accessible to receptor-activated DAGK (van der Bend et aL, 1994). This topological restriction of DAGK activity suggests that DAGK phosphorylates only DAG generated as a result of receptor stimulation and is in line with the concept of an agonist- responsive Ptdlns pool in which receptor-induced DAG is channelled from endogenous PLC to the DAGK aetive site. The data also imply the compartmentation of DAG and possibly PtdOH. Such segregation of intermediate metabolites is also reflected in experimental data which suggest that hydrolysis and resynthesis of Pis occurred in a ‘closed’ PI cycle, whereby Ptdlns synthesized in response to agonist-induced depletion is preferentially sensitive to subsequent agonist-induced hydrolysis (Monaco and Gershengom, 1992, and references therein).

Compartmentation of Pis and their derivatives have also been observed in the nucleus (D'Santos et aL, 1999; Vann et aL, 1997). For example, there may be at least two distinct pools of DAG in the nucleus: one generated from Pis and the other from PtdCho, and the former was suggested to be the pool that mediates PKC translocation into the nucleus (Neri et aL, 1998). In addition, PI kinases and lipases are localized to different nuclear sites (Payrastre et aL, 1992). PtdIns4K-IIip was concentrated at lamina-pore complexes in the nucleus of NIH 3T3 cells (de Graaf et aL, 2002), while PtdlnsfKs and PLCs were found in the inner nuclear compartment (Boronenkov et aL, 1998; Cocco et aL, 2000; Zini et aL, 1993).

Despite years of studies, the physical manifestation of these PI pools remains enigmatic, and different data have both supported and questioned their existence (Monaco and Gershengom, 1992). Nevertheless, at least three mechanisms have been proposed to help explain the metabolic compartmentation of Pis. Firstly, EGF-induced PI tumover has

71 been suggested to take place in caveolae, which are discrete detergent-insoluble plasma membrane structures that are enriched in caveolin (Parton, 1996). Caveolae appear to contain most, if not all, of the molecules involved in PI signalling, including receptors, PI kinases and Pis. Secondly, Pis in agonist-responsive pools have predominantly stearyl and arachidonyl acyl chains, while those in agonist-unresponsive pools contain other types of acyl chain (Lee et aL, 1991). This is in line with studies showing that PITPs and DAGKs discriminate between different PI acyl groups (Tang et aL, 1996; Wirtz, 1991). Finally, the in vitro enzymatic activities of PtdIns4K, PtdlnsPK and PLC isozymes differ with different modes of substrate presentation, for example, in the presence of carrier lipids, detergents and PtdlnsTP. These three mechanisms are not mutually exclusive and pave the way to a testable hypothesis for PI compartmentation.

Most of the above data support the concept that PITP plays an important role in the compartmentalised synthesis of Ptdlnsf] in the cell, although the mechanism remains unresolved. The main objective of this thesis is therefore to investigate and test the central tenets of the co-factor model of PITP function in PI signalling, that is, (i) the ability of PITP to interact with PI signalling complex(es), and (ii) the viability of PITP-bound Ptdlns as a direct substrate to PI kinases.

72 Chapter 2

Materials and Methods

Materials

Most biochemicals and reagents were of analytical grade and purchased from BDH and Sigma. When available, radioactive and radio-labelled reagents and restriction enzymes were purchased from Amersham-Pharmacia; other suppliers included New England Biolabs. Antibiotics and protease inhibitors were from Sigma and Life Technologies. HPLC equipment and gel filtration columns (pre-packed) were from Amersham- Pharmacia. Unless otherwise stated, bacterial and cell culture media were from Life Technologies.

Buffers supplied by manufacturers were used, unless indicated otherwise. Calcium buffers were obtained from Prof. Cockcroft’s laboratory and prepared as described by Gomperts and Tatham (1992). All other buffers were prepared as described in Sambrook et al. (1989).

Amersham-Pharmacia: alkaline phosphatase (calf intestinal), ECL reagents, high range prestained molecular weight markers (for SDS-PAGE), Novablot electrode paper. Protein A Sepharose, T4 DNA ligase Beckman: Ready Safe^*^ liquid scintillation cocktail Bio-Rad: Bio-Lyte 5-8 or 3-10, Bradford reagent, Dowex 1-X8 anion-exchange resin, lEF Standards, protein molecular weight standards (high and low range; for SDS- PAGE) Boehringer Mannheim: isopropylthio-P-D-galactoside (IPTG) Eastman Kodak Co.: anti-FLAG M2 affinity gel, anti-FLAG M2 monoclonal antibody Genetic Research Instrumentation: X-ray film (blue-sensitive) In vitro gen: pcDNA3 expression vector Jackson ImmunoResearch: FITC-conjugated anti-mouse IgG Life Technologies: 4% agarose solution, DNA ladder (1 kb), foetal calf serum, lipid concentrate, Lipofectin, T4 DNA ligase, trypsin Millipore: Immobilon-P (polyvinylidene fluoride, PVDF) transfer membrane Murex: Streptolysin O New England BioLabs: Vent^ DNA Polymerase Novagen: pET protein expression system, BL21(DE3) cells 73 Novex: Native Running Buffer, Native Sample Buffer, pre-cast Tris-Glycine gels Perkin-Elmer: PRISM Ready Reaction DyeDeoxy Terminator Cycle Sequencing Kit Packard: Ultima-Flo AF scintillant Pharmingen: BaculoGold DNA Promega: deoxyribonucleotide triphosphate, Taq DNA polymerase Oiagen: EndoFree Plasmid Maxi Kit, nickel-nitrilotriacetic acid (Ni^^-NTA) Metal Affinity Chromatography resin, QIAEX II Extraction Kit, QIAprep Spin Miniprep, QIAquick Gel Extraction Kit, SuperFect Sigma: anti-rabbit or anti-mouse immunoglobulin G (IgG) conjugated to horseradish peroxidase (HRP), epidermal growth factor, goat serum, lysozyme. Medium 199, P-mercaptoethanol, phalloidin-TRITC, phosphatidylcholine, polyethylene glycol

8000, X-gal (5-bromo-4-chloro-3-indolyl p-D-galacto-pyranoside), yeast extract ultrafiltrate Stratagene: Bluescript SK plasmid, XL 1-Blue cells Whitman: thin-layer chromatography (TEC) plate

2.1 Molecular Biology

2.1.1 DNA Biochemistry

2.1.1.1 Polymerase Chain Reaction (PCR)

The following PCR protocol was adapted from that described in Sambrook et al. (1989). Either of two DNA polymerases was used: Taq DNA polymerase for amplification of low-copy-number cDNA, or Vent^ DNA Polymerase for PCR with adequate template DNA. Typically, a PCR was performed in a 50 pi volume containing not more than 0.2 pg template DNA, 1 pg of each primer, 200 pM deoxyribonucleotide triphosphate (dNTPs), 1 unit of the indicated DNA polymerase and, unless indicated, the recommended concentration of the manufacturer’s reaction buffer. Mg^"*" was added (as

MgCl 2 or MgS 0 4 ) where indicated, and the solution overlaid with 40 pi of mineral oil. PCRs were run on a DNA Thermal Cycler (Perkin Elmer 480) for 30 cycles of the following three sequential steps: dénaturation (94°C, 40 s), annealing (temperature varied with primers, 1 min), and extension (73°C, 1 min). The solution was then kept at 4°C until required.

Overlapping PCR is a two-stage PCR strategy for introducing desired mutations in an open reading frame (ORF). The template is initially amplified in two separate reactions, 74 with each amplifying half of the ORP. The products of these two reactions overlap in the ORP where the mutated codon is located and the mutation is incorporated by the primers. These PCR products are gel-purified and re-amplified together in a single reaction to generate a full-length ORP with the desired mutation.

2.1.1.1.1 Polymerase Chain Reaction Primers

The forward primer for human PITPa, a5, contained sequences complementary to 21 nucleotides at the 5' end of human PITPa (Dickeson et aL, 1994). The two restriction sites, BamUl and Nde\, allowed cloning into vectors pBSII and pET21a, respectively. Use of the Ndel site would also remove the nucleotide sequence coding for the T7 tag from the pET21a vector.

BamHI start a5: 5' AAA AGG ATC CCA TAT GOT GCT GCT CAA GGA GTA T 3‘ Ndel

The reverse primer for human PITPa, a3, contained sequences complementary to 22 nucleotides at the 3' end of human PITPa, but excluded the stop codon. The SaK restriction site enabled cloning into both vectors pBSII and pET21a. The stop codon was removed to allow translation to include a C-terminal 6 -histidine (Hise) tag encoded by pET21a.

Sail a3: 5' AAA AGT CGA CGT CAT CTG CTG TCA TTC CTT TC 3'

The forward primer for rat PITPp, P5, had a sequence matching the 21 nucleotides at the

5' end of rat PITPp (Tanaka and Hosaka, 1994), as well as BamYil and Nco\ restriction sites to enable cloning into vectors pBSII and pET21d, respectively. The Nco\ site also removed the T7 tag-coding sequence from the pET21d vector.

BamHI start p5: 5' AAA AGG ATC CCC ATG GTG CTG ATT AAG GAATTC CG 3' Ncol

The reverse primer for rat PITPp, p3, contained a sequence complementary to the 19 nucleotides at the 3' end of rat PITPp. Its SaR restriction site behaved as described for primer a3.

75 Sali P3: 5' AAA AGT CGA CGG CAT GAG GAG GGG AGG TG 3'

A primer (aP) common to both PITPa and PITPp was synthesised for screening bacterial colonies. Its sequence was identical to human PITPa nucleotides 166-189, and the corresponding sequence in rat PITPp save nucleotide 180 (A to G substitution). ap: 5' GAG TAG AGA GAG AAG ATG TAG GAG 3'

The forward and reverse primers for human PITPa, JHAl and JHA4 respectively, were used for the baculovirus expression system (BES). JHA4 was also used for expression in mammalian cells. They contained sequences complementary to 5' or 3' ends of human PITPa and each has a restriction site EcoRl or Bam\{\, respectively.

EcoRI JHA1 : 5 ‘ ATA TGA ATT GGA TAT GGT GGT GGT GAA GGA G 3'

BamHI JHA4: 5' AAA AGG ATG GTT AGG AGG GGG 3'

The forward and reverse primers for mutant human PITPa (T59D), JH59A+ and JH59A-, were substituted the Thr (ACA) codon for Asp (GAG) or the Thr (TOT) anti-codon for Asp (GTG) respectively, and were used for BES and expression in mammalian cells.

Asp JH59A+: 5' GGA GTA GGÂCGA GAA GAT GTA GGA G 3'

JH59A-: 5 ‘ TGT TGT GGT GGT AGT GGG GTT TOT G 3'

The forward and reverse primers for rat PITPp, JHBl and JHB4, were used for BES. JHBl was also used for mammalian cell expression. They contained sequences complementary to 5' or 3' ends of rat PITPp and had an Nco\ or BamYH restriction site respectively.

Ncol JHB1: 5' ATA TGG GGG GTA GGA TGG TGC TGA TTA AGG A AT T 3'

76 BamHI JHB4: 5‘ AAA AGG ATC CTC AGT GGT GGT G 3‘

The forward and reverse primers for mutant rat PITPp (T59D), JH59B+ and JH59B-, were substituted the Thr (ACA) codon for Asp (GAG) or the Thr (TGT) anti-codon for Asp (GTG), and were used for BES.

JH59B+: 5' ACA GTA CGA CCA CAA AAT CTA CCA C 3‘

JH59B-: 5‘ TIT TGT GGT CGT ACT GTC CCT TCT C 3‘

The forward and reverse primers for human PITPp, JHB3 and JHB2, were used for mammalian cell expression. They contained sequences complementary to the 5' or 3' ends of human PITPp (Tanaka et al., 1995) and had an Ncol or Banûll restriction site respectively.

Ncol JHB3: 5' AAA ACC ATG GTG CTG ATC AAG GAATTC CGT 3‘

BamHI JHB2: 5' AAA AGG ATC CGA CAT CAG CAG CCG ACG TGC CTC G 3'

The forward and reverse primers for mutant human PITPp (T59D), JHBM+ and JHBM-, were used to substitute the Thr (AGG) codon for Asp (GAG) or the Thr (GGT) anti-codon for Asp (GTG) respectively for expression in mammalian cells.

Asp JHBM+: 5 ’ ACA GTA TGA CCA CAA AAT TTA TCA C 3'

JHBM-: 5" TTT TGT GGT CAT ACT GTC CCTTTT C 3'

2.1.1.2 Bacterial Transformation

The following methods were adapted from those described in Sambrook et al. (1989).

To prepare competent XL 1-Blue Escherichia coli cells, a 2 ml culture was grown overnight in Luria-Bertani (LB) medium under selection in 10 pg/ml tetracycline at 37°G in a shaking incubator. 500 pi of this overnight culture was added to 50 ml of LB medium 77 and shaken under selection until cell density reached an absorbance at 600 nm (Aôoo) of about 0.6. The cells were then placed on ice for 10 min, centrifuged at 2000 rpm for 15 min at 4°C and resuspended gently in 17 ml of sterile storage buffer (100 mM KCl, 50 mM CaCl2, 10% glycerol, 10 mM potassium acetate). After a further 10 min on ice, cells were pelleted at 1000 rpm for 15 min at 4°C. Cells were subsequently resuspended in 2ml of storage buffer and dispensed into pre-cbilled, sterile polypropylene tubes. They were flasb-frozen in a dry ice/etbanol bath for 10 min and stored at -70°C.

To transform competent XL 1-Blue cells, one plate of LB agar containing 50 pg/ml carbenicillin was prepared for each transformation. 50 pi of 2% X-gal (dissolved in dimetbylformamide) was spread on the agar and left to absorb. When necessary, 10 pi of 1 M IPTG was then added and the plate was covered for 1 b at room temperature. For each transformation, an 80-pl aliquot of competent cells was thawed on ice, mixed with

10 pi of DNA and left on ice for 30 min. The mixture was then beat-shocked for 2 min at 37°C, returned to ice for 2 min and incubated for 1 b at 37°C after adding 200 pi of LB medium. The cells were subsequently spread onto an agar plate, left at room temperature for 10 min and then inverted and incubated for 12-16 b at 37°C.

Competent BL21(DE3) cells were transformed as described by the manufacturer (Novagen). BL21(DE3) cells were chosen as an expression host because they lack the Ion protease and ompT outer membrane protease that can degrade some proteins after cell lysis and during protein purification.

2.1.1.3 Plasmid Preparation

Plasmids for all molecular biology work were prepared using the QIAprep Spin Miniprep Kit or as follows: 5 ml of Terrific Broth (TB) or LB medium containing the appropriate antibiotic was inoculated with a single bacterial colony picked from a plate and incubated overnight at 37°C with vigorous shaking. 3 ml of the culture was pelleted, resuspended in

200 pi of sterile Solution I (50 mM glucose, 10 mM EDTA, 25 mM Tris-HCl pH 8 ) and lysed gently with 400 pi of freshly-prepared Solution II (0.2 N NaOH, 1% SDS). After 5 min on ice, 200 pi of ice-cold Solution III (3 M potassium acetate) was added and mixed rapidly into the lysate. After a further 5 min on ice, the mixture was centrifuged at 12,000 rpm for 10 min at 4°C. The supernatant was added to an equal volume of phenol:chloroform (1:1), vortexed and centrifuged for 5 min. The supernatant was further mixed with 0 . 6 volume of isopropanol and centrifuged for 1 0 min at room temperature.

78 The pellet was rinsed with cold 70% ethanol, spun at 4°C and vacuum-dried for 3 min, after which it was incubated in 100 pi of TE buffer (1 mM EDTA, 10 mM Tris-HCl pH 8 ) containing 10 pg/ml ribonuclease A for at least 1 h at 37°C. To precipitate the DNA, 60 pi of 20% polyethylene glycol (PEG) 8000, 2.5 M NaCl was added for incubation for 1 h on ice before centrifugation for 10 min at 4°C. The purified plasmid pellet was rinsed with 70% ethanol, spun and dried as above, and resuspended in 50 pi of TE Buffer pH 8 for storage at -20°C.

Plasmids for use in tissue culture were prepared using the EndoFree Plasmid Maxi Kit (Qiagen).

2.1.1.4 DNA Digestion with Restriction Enzymes

When available, restriction enzymes from Amersham-Pharmacia were used; other sources included Life Technologies and Roche. Typically, 0.5-1.0 pg of DNA was incubated for at least 1 h at 37°C in 20 pi (final volume) of OPA+ Buffer (Amersham-Pharmacia) of the recommended concentration and 1 unit of each restriction enzyme. When a vector plasmid was digested, an additional step was performed: restriction enzymes were heat- inactivated first before 0.5 unit of calf intestinal alkaline phosphatase was added for an additional 30 min to prevent recircularisation of the vector.

2.1.1.5 Agarose Gel Electrophoresis of DNA

The following method was adapted from that described in Sambrook et al. (1989). A 1% agarose gel was prepared by dissolving 1 g of agarose in 100 ml of TAE buffer (40 mM Tris-acetate, 1 mM EDTA, 0.5 pg/ml ethidium bromide) and setting the solution in a gel mould. DNA was mixed with 0.2 volume of Gel-Loading Buffer (0.25% bromophenol blue, 0.25% xylene cyanol FF, 30% glycerol) and loaded onto the gel. A voltage of 5 V/cm was then applied. After sufficient separation had been achieved, DNA was visualised under UV light and purified using the QIAEX II Extraction Kit or QIAquick Gel Extraction Kit. The sizes of DNA fragments were estimated using a 1 kb DNA ladder (Life Technologies).

2.1.1.6 DNA Ligation

T4 DNA ligase was used according to instructions provided by the respective manufacturers. Typically, 3 pg of insert cDNA was mixed with every 1 pg of vector

79 DNA plasmid and incubated with 1 unit of ligase for at least 1 h at 16°C in the ligation buffer provided.

2.1.1.7 Automated DNA Sequencing

The PRISM Ready Reaction DyeDeoxy Terminator Cycle Sequencing Kit was used according to the manufacturer’s instructions. Purified DNA sequencing products were stored at 4°C prior to loading onto a sequencing gel. Gels were run and the data analysed by a service facility within the Ludwig Institute.

2.1.2 Protein Biochemistry

2.1.2.1 Polyacrylamide Gel Electrophoresis (PAGE)

2.1.2.1.1 Sodium Dodecyl Sulphate (SDS)-PAGE

Solutions for both the resolving and stacking portions of the gel were prepared as described (Sambrook et al., 1989) and set to a thickness of 0.75mm in a Hoefer SE250 Mighty Small unit or SE400 Sturdier Vertical unit (Amersham-Pharmacia). Depending on the molecular mass of the proteins to be analysed, gels of 7.5%, 10% or 15% acrylamide were used. Either high or low range mixtures of molecular standards were used if gels were to be subsequently stained, or high range prestained molecular weight markers (Rainbow) for subsequent electroblotting of the separated proteins. Before loading onto a gel, protein samples were mixed with an equal volume of 2 x Loading Buffer (0.125 M

Tris-HCl pH 6 .8 , 4% SDS, 20% glycerol, 0.2 M dithiothreitol (DTT), 0.02% bromophenol blue) and heated for 2-3 min at 100°C. Electrophoresis was carried out in Tris-glycine buffer (25 mM Tris, 190 mM glycine, 0.1% SDS) at 450 V for 3 h (or until the dye-front had reached the edge of the gel) or at 50 V overnight.

2.1.2.1.2 Native PAGE

Pre-cast Tris-Glycine gels (8-16%) and Native Running and Sample buffers were used according to the manufacturer’s instructions. The gel cassette was assembled on an XCell Mini-Cell apparatus (Novex) and the buffer chambers filled with Native Running Buffer. Samples were mixed with equal volume of Native Sample Buffer at 4°C and loaded onto the gel. A constant voltage of 125 V was applied until the dye front reached the bottom of the gel.

80 2.1.2.1.3 Iso-Electric Focusing

Iso-electric Focusing (lEF) was performed on the Mini lEF Cell (Model 111; Bio-Rad) as described in the instruction manual. The 0.4 mm, 5% acrylamide lEF gels were cast on the casting cassette provided, using Bio-Lyte 5-8 or 3-10 as indicated to achieve optimal resolution at pi 5-8 or 3-10. PITP samples (0.3-0.5 pg, unless otherwise indicated) and lEF Standards were applied at various points between the anode and cathode. The gel was run at constant voltage until electric current was negligible.

To visualise the positions of migrated proteins, lEF gels were stained with Coomassie Blue as described in Section 2.1.2.2.1. Alternatively, proteins were blotted onto Immobilon-P (polyvinylidene fluoride, PVDF) transfer membrane. The membrane was first soaked briefly in methanol and then immersed in lEF Blotting Buffer (25 mM Tris pH 8.3, 192 mM glycine, 20% methanol) for at least 5 min, before being placed on top of the lEF gel for 15 min. lEF Blotting Buffer was periodically applied to the lEF gel-PVDF membrane sandwich to prevent the membrane from drying out. The membrane was subsequently analysed with antibodies as described in Section 2.1.2.4.

2.1.2.2 Staining of Protein Gels

The following methods and reagents were modified from those described in Sambrook et a l (1989).

2.1.2.2.1 Coomassie Staining

The gel was submerged for at least 10 min in Staining Solution (0.025% Coomassie Brilliant blue R250 dissolved in 40% methanol, 7% acetic acid) on a shaking platform before de-staining in several changes of Destaining Solution (40% methanol, 7% acetic acid).

2.1.2.2.2 Silver Staining

After electrophoresis, the separated proteins were fixed in the gel overnight with Destaining Solution (Section 2.1.2.2.1) and then washed with distilled water (dHjO) for 5 min. The gel was subsequently sensitised with 12.5% (v/v) glutaraldehyde for 7.5 min before two 5-min washes in dH 2 0 and one 7.5-min wash in 20% ethanol. Staining solution (20% ethanol, 0.2% NaOH, 0.25% ammonia, 0.2% silver nitrate) was then added for 15 min, followed by two 5-min washes in 20% ethanol. The gel was developed in

81 Developing Solution (20% ethanol, 0.037% formaldehyde, 575 pM citric acid) until protein bands were visible. The gel was preserved in 10% acetic acid, 5% glycerol.

2.1.2.3 Semi-dry Electrophoretic Transfer of Protein

This method was adapted from that described elsewhere (Burnette, 1981; Sambrook et al., 1989) One piece of Immobilon-P transfer membrane and 12 sheets of Novablot electrode paper were cut to the size of the gel. The membrane was first soaked briefly in methanol and then at least 5 min in Transfer Buffer (48 mM Tris, 39 mM glycine, 20% methanol). The electrode paper was also soaked thoroughly in Transfer Buffer before use. The following assembly was then set up on the Multiphor Novablot unit (Amersham- Pharmacia) anode plate: six sheets of pre-soaked electrode paper were stacked on the anode and the transfer membrane was then placed on top. The protein gel was quickly placed over the membrane, followed by the other six sheets of electrode paper. Any bubbles trapped in this sandwich were forced out with a clean pipette. The electrophoretic transfer was run for 1-1.5 h at 0.8 mA/cm^ of gel.

2.1.2.4 Western Blotting

The Immobilon-P membrane was blocked for 1 h with Blocking Solution (5% low-fat milk, 2.5% goat serum, 0.02% Tween 20 in PBS) on a rotary wheel and washed in generous volumes of PBST (0.02% Tween 20 in PBS) in three steps: one 15-min wash, followed by two 5-min washes. The membrane was then incubated for 1 h in diluted primary antibody in PBST, washed as above and further incubated for 1 h in the secondary antibody in PBST. The secondary antibody was either an anti-rabbit or anti­ mouse IgG conjugated to HRP. Excess secondary antibody was removed using the above three washes in PBST. Immuno-chemiluminescence was initiated by covering the membrane with ECL reagents for 1 min, and a sheet of Blue-sensitive X-ray film was then placed over the membrane for the indicated time.

2.1.2.5 Bradford Assay

Protein concentration in a given solution was estimated using the Bradford reagent (Bradford, 1976). Typically, 20 pi of protein solution was mixed in a disposable cuvette with 1 ml of the Bradford reagent (diluted 5-fold in distilled water) for at least 5 min but not longer than 30 min at room temperature. Absorbance at 595 nm (A 595) was then

82 measured in a spectrophotometer (versus A 595 of solvent) and the protein concentration estimated by comparison with a series of bovine serum albumin standards.

2.1.3 Lipid Biochemistry

2.1.3.1 Separation of Inositol Phosphates

Inositol monophosphate (InsP), inositol bisphosphate (Insfz) and inositol trisphosphate (InsPs) can be separated using Dowex 1-X8 anion-exchange resin, as described by Cockcroft et al. (1994b). 100 g of Dowex resin was stirred in 400 ml of 1 M NaOH and then allowed to settle for 1-2 h. This step was repeated using 400 ml of IM formic acid.

The resin was then washed thoroughly with dH2 0 and stored as a 50% slurry at 4°C.

0.5 ml (bed volume) aliquots of the resin containing formate counterions, were transferred to glass wool-plugged Pasteur pipettes and washed sequentially with 6 ml of 2 M ammonium formate/0.1 M formic acid and 20 ml of dH 2 0 . These resin columns were stored at 4°C and washed extensively with dH 2 0 before use. After the sample was loaded, inositol was washed off with 6 ml of dH 2 0 , and glycerophosphoinositol was removed with 6 ml of 5 mM sodium tetraborate/60 mM sodium formate. InsP, InsP 2 and InsP^ were then separately and sequentially eluted with 3 ml each of 0.2 M ammonium formate/0.1 M formic acid, 0.4 M ammonium formate/0.1 M formic acid, and 1 M ammonium formate/0.1 M formic acid, respectively. Each eluate was mixed with 10 ml of Ultima-Flo AF scintillant.

2.1.3.2 Identification of Inositol Phospholipids

2.1.3.2.1 Thin-layer Chromatography

The following method was adapted from Wetzker et al. (1991). TEC plates were soaked briefly in 1% potassium oxalate, 50% methanol, 2 mM EDTA and then dried for 30 min at 110°C. Oxalate treatment obviated the need to remove divalent cations when the resolved lipids were subsequently extracted for déacylation (Section 2.1.3.2.2). Lipid samples were dried in vacuo^ resuspended in 5 pi of TLC Running Buffer (propan-1- ol:2M acetic acid (65:35) containing 1% H3PO4) and applied in 1 pi aliquots onto the TLC plate at the origin. The applied samples were left to dry before the plate was placed in a buffer chamber saturated with Running Buffer. The plate was removed when the solvent front was near the top of the plate and left to dry in a fume hood.

83 Two methods were used to detect separated lipids on TLC plates. In the first method, iodine crystals were vaporised in a closed box at 60°C for 10 min. The box was subsequently transferred to a fume hood. The dry TLC plate was then placed in the closed box for 3-10 min until lipids were stained brown. In the second method, the plate was sprayed generously with ImM 2-(p-toluidino)naphthalene-6-sulphonic acid (TNS) in 50mM Tris/HCl, pH7.4 and then visualised under ultra-violet light. Lipid spots gave a blue fluorescence using this stain.

2.1.3.2.2 Lipid Déacylation

The specific identities of inositol phospholipids were established through déacylation using methylamine (Clarke and Dawson, 1981) and subsequent liquid chromatography (Stephens et al., 1991). Each lipid spot on the TLC plate was scraped off together with the silica gel and then treated with 200-500 pi of room temperature methylamine reagent

(33% monomethylamine in ethanol : H 2O : butan-1 -ol ; 50:15:5) for 30-50 min at 53°C. The mixture was cooled on ice and excess methylamine removed in vacuo. The dry residue was extracted with 500 pi of H 2O and 600 pi of petroleum mixture (butan-1 -ohpetroleum ether (b.p. 40-60°C):ethyl formate; 20:4:1). The mixture was then vortexed and centrifuged for 2 min. The lower aqueous phase was treated with another 600 pi of petroleum mixture, vortexed and spun for 2 min. Butan-l-ol in the biphasic separation mixture facilitated the removal of lysoplasmalogens (alkenyl) and lysoglycerol ether phospholipids (alkyl) from the aqueous layer. The final aqueous phase was dried to 100-

2 0 0 pi before 1 ml of H 2O was added and the mixture subsequently filtered to remove silica debris.

The recovered glycerophosphoesters (in the aqueous phase) were separated using an HPLC Partisphere 5-SAX anion-exchange column developed at 1 ml/min with a gradient in Buffer B (1.25 M NaH 2P0 4 , pH adjusted to 3.8 using NaOH); fractions were collected every 0.5 min. This HPLC analysis was kindly performed by Dr. Frank Cooke (Imperial Cancer Research Fund Laboratories, London).

2.2 Cell Culture

2.2.1 Insect Cell Culture

2.2.1.1 Routine Maintenance

84 Spodoptera frugiperda Sf9 cells were maintained in monolayer cultures at 27°C in 175 cm2 culture flasks as described (King and Possee, 1992). The IPL41 medium, supplemented with 10% foetal calf serum (PCS), yeast extract ultrafiltrate (Ix), lipid concentrate ( 1 x), penicillin ( 1 0 0 units/ml) and streptomycin ( 1 0 0 pg/ml), was warmed to 27°C before use. When 90% confluent, the cells were detached from the flask by gentle agitation and passaged at 1 : 1 0 dilution into fresh medium.

To prepare Sf9 cells for protein expression, they were grown in suspension cultures using spinner flasks (Techne), with constant stirring at 65 rpm. Passage was required every 3-4 days when the cell density reached 3 -4 x 1 0 ^ cells/ml.

2.2.1.2 Generation of Baculovirus Stock

The following methods were adapted from those recommended by Pharmingen and Life Technologies and described by King and Possee (1992).

2.2.1.2.1 Transfection of Sf9 cells.

For each transfection, 3x10^ cells were seeded in three 25 cm^ culture flasks and allowed to attach firmly for 1 h at 27°C. For each flask, 2 pg of recombinant pVL1392 baculovirus transfer vector was mixed gently in a polystyrene tube with 0.3 pg of

BaculoGold DNA in a total volume of 16 pi of sterile water. 8 pi of Lipofectin was then added and allowed to form a complex with the plasmid DNA by incubating the tube for 15 min at room temperature. The attached Sf9 cells were gently washed once with IPL41 medium supplemented only with yeast extract and lipid concentrate (IPLyl), and then covered with 1.5 ml of IPLyl. The DNA-Lipofectin mix was added to the cells and distributed evenly. After overnight incubation at 27°C, the DNA-Lipofectin mix was removed and replaced with 5 ml of IPL41 medium containing the full supplement of additives (Section 2.2.1.1). The flasks were returned to incubation chambers for 4 days before the supernatant medium (containing virus) was collected. This medium was centrifuged at 3000 rpm for 5 min to remove cells and stored at 4°C. All solutions containing BaculoGold DNA were pipetted using wide-aperture tips to minimise DNA shearing.

2.2.1.2.2 Plaque Assay

For each transfection supernatant, six 3 5-mm culture dishes were each seeded with 1.5 X10^ cells and left for the cells to attach for 1-2 h at 27°C. Serial dilutions (10'^ to 10’*)

85 of each supernatant were prepared in volumes of 250 |l i 1. After removing the medium from the dishes, 2 0 0 pi of each diluted supernatant was added gently at the centre of the dish, and left to incubate at 27°C for 1 h. The dishes were rocked gently every 15 min to maximise even infection throughout the dish. 2 ml of 1 % agarose was prepared for each dish, by diluting 4% agarose solution with IPL41 medium containing the full supplement of additives (Section 2.2.1.1). At the end of the infection time, the viral supernatant was removed and 2 ml of 1 % agarose solution added gently to the cells from the side of the dish. The agarose was allowed to set before the dishes were transferred to a moist box at 27°C for 3-4 days or until plaques were visible.

2.2.1.2.3 Virus Amplification

Typically, four viral plaques for each recombinant protein were selected for amplification. Each plaque was picked up using a wide-aperture pipette tip and dispersed in 0.5 ml of IPL41 medium. 1 .5 x 1 0 ^ Sf9 cells were seeded in a 25 cm^ flask for each plaque and left to attach for 1 h at 27°C. The medium was removed and 250 pi of the viral suspension was added to the cells for 1 h at 28°C. The viral suspension was then replaced by 4 ml of IPL41 medium and the infected cells left to incubate at 28°C for 4-5 days. The medium was collected, briefly centrifuged and the supernatant stored at 4°C as a seed stock. A plaque assay was then performed for each viral seed stock to estimate its titre (in pfu/ml).

To further amplify each viral seed stock, 50 ml of Sf9 suspension culture at an initial cell density of 1x10^ cells/ml was prepared. When the cell density reached 5x10^ cells/ml, the culture was infected with viral seed stock at 0.2 pfti/cell at 28°C for 5 days. Cells were then removed by spinning at 500xg at 4°C for 10 min and the supernatant stored at 4°C in the dark. A plaque assay was performed to determine the titre of this stock and, if necessary, a further round of viral amplification was carried out.

2.2.2 Mammalian Cell Culture

2.2.2.1 Routine Maintenance

All cell cultures were maintained at 37°C in 10% CO 2 using a water-jacketed incubator. Adherent cell cultures were grown in Dulbecco's Modified Eagle Medium (DMEM) supplemented with 10% heat-inactivated PCS, penicillin (100 units/ml) and streptomycin (100 pg/ml). When confluent, different cell lines were passaged as described below.

86 A431 and COS-7 cell lines were washed once with PBS and then incubated with 1 ml of trypsin at 37°C until the cells detached. 10 ml of DMEM was added to inactivate the trypsin and cells were diluted 1 : 1 0 on a new dish.

Human embryonic kidney (HEK) cells, on the other hand, were washed with PBS and treated briefly with 3 ml of trypsin which had been diluted 5-fold in PBS. As cells started to detach from the dish, the trypsin was aspirated and 10 ml of DMEM was added to break up aggregated cells before dilution.

HL60 cells were grown in suspension in 175 cm^ flask with RPMI 1640 medium supplemented with 10% heat-inactivated PCS, penicillin (100 units/ml) and streptomycin (100 pg/ml). When confluent, cells were pelleted at 450xg for 5 min, resuspended in 5 ml of RPMI 1640 medium and passaged at 1:10 dilution.

2.2.2.2 In situ Immunofluorescence and Confocal Microscopy

Sterile glass coverslips were dipped in 70% ethanol and flamed before being placed in 12- well culture dishes. The required number of cells was then seeded in each well and incubated for 24 h at 37°C in 10% CO]. Cell transfection with DNA plasmid was performed as described in Section 2.3.3.1.

Where permeabilisation was required, cells were treated with streptolysin O two days after passage as described in Section 2.4.1.1.3. All cells were subsequently fixed with 1 ml of 3.7% paraformaldehyde in PBS for 20 min at room temperature, washed six times with PBS and treated with 0.2% Triton X-100 in PBS for 5 min. This was followed by two washes with PBS and staining with anti-FLAG M2 monoclonal antibody (10 pg/ml in PBS, 0.1% BSA) on a rocking platform for 1 h. Excess antibody was then removed using six washes with PBS before treatment with a second staining cocktail containing FITC-conjugated anti-mouse IgG (diluted 1:400), phalloidin-TRITC (diluted 1:1000), and, where indicated, 1 pg/ml 4',6-diamidino-2-phenylindole (DAPI). Staining was carried out in the dark for at least 30 min on a rocking platform. After six more washes with PBS, each coverslip was inverted on 5 pi of Mowiol mountant and incubated overnight in the dark. Cells were viewed using a scanning confocal microscope (Zeiss 510) operated by Drs. J. Hsuan and M. dos Santos.

2.3 Protein Expression and Purification

87 2.3.1 Expression System in Bacteria

2.3.1.1 Expression of Recombinant Proteins in E. coli.

The protocol was as described in the pET System Manual. 5 ml of LB medium containing 50 pg/ml carbenicillin was inoculated with a single bacterial colony or a stab from a glycerol stock and shaken at 37°C overnight or until the Aeoo was about 0.6. An aliquot of this culture corresponding to 4% of the required final volume was pelleted at 4,000 rpm for 5 min, resuspended in the required final volume of selective LB medium and shaken as above. When the culture reached an A ôoo of 0.6, protein expression was induced using indicated concentrations of IPTG for 3 h at the indicated temperature.

2.3.1.2 Purification of Recombinant PITP

IPTG-treated cells (Section 2.3.1.1) were pelleted and resuspended in 10% of the culture volume of cold Sonication Buffer (50 mM sodium phosphate pH8, 100 mM NaCl) containing 1 mg/ml lysozyme, 5 pg/ml deoxyribonuclease, 1 pg/ml aprotinin, 10 pM leupeptin, 1 pM pepstatin, 1 mM phenylmethylsulfonyl fluoride (PMSF). After incubation on ice for 30 min, cells were sonicated in six 10 s bursts (with at least 10 s of cooling on ice between each burst) and the lysate centrifuged at 39,000 for 20 min at 4°C. The supernatant was then mixed with Ni^^-NTA Metal Affinity Chromatography resin (pre-equilibrated with Sonication Buffer) on a rotary wheel for 1 h at 4°C. The resin was pelleted at 800 and washed with cold Wash Buffer (50 mM sodium phosphate pH7, 300 mM NaCl, 10% glycerol, 1 mM P-mercaptoethanol, 10 mM imidazole) until the

A280 of the supernatant was less than 0.01. PITP was eluted using Elution Buffer (50 mM sodium phosphate pH7, 150 mM NaCl, 100 mM imidazole) and then solvent-exchanged into Ptdlns 4-kinase buffer (25 mM Tris-HCl pH7.4, 150 mM NaCl, 2 mM MgCl 2) using either Microsep Microconcentrators (Filtron Technology Corporation) or HiTrap Desalting columns (Amersham-Pharmacia).

2.3.2 Expression in Insect Cells

The following methods were adapted from those recommended by Pharmingen, Life Technologies and Qiagen, and described by King and Possee (1992).

2.3.2.1 Expression of Recombinant Proteins in Sf9 Cells One day before infection, suspension cultures of Sf9 cells were prepared at a cell density of 0.6 X10^ cells/ml. When the cell density reached 1.2x106 cells/ml, high-titre baculovirus stock was added at a multiplicity of infection (MCI) of 7.5. The infected cells were incubated for 2 days at 28°C and then harvested.

2.3.2.2 Purification of PITP

Infected Sf9 cells were lysed in ice-cold Lysis Buffer (10 mM Tris-HCl pH7.4, 130 mM NaCl, 10 mM NaF, 20 mM sodium phosphate, 10 mM sodium pyrophosphate, 0.3% Triton X-100, 10 pg/ml aprotinin, 10 pg/ml leupeptin, 16 pg/ml benzamidine, 10 pg/ml phenanthroline, 10 pg/ml pepstatin, 1 mM PMSF) for 45 min, and then centrifuged at 39,000Xg for 20 min at 4°C. Ni^^-NTA resin, which had been pre-equilibrated with Lysis Buffer, was mixed with the supernatant for 1 h at 4°C and then washed with Wash Buffer (50 mM sodium phosphate pH7.4, 300 mM NaCl, 10% glycerol, 30 mM imidazole) until the A280 of the supernatant was less than 0.01. PITP was eluted in Elution Buffer (50 mM sodium phosphate pH6, 300 mM NaCl, 10% glycerol, 400 mM imidazole) and solvent- exchanged as described in Section 2.3.1.2.

2.3.3 Expression in Mammalian Cells

2.3.3.1 Transfection of Cells

The pcDNA3 expression vector, the transfection reagents Lipofectin and SuperFect were used according to the respective manufacturer’s instructions. COS-7 and A431 cells were transfected at 40-50% confluency, while HEK cells were transfected at 70% confluency.

2.3.3.2 Immunoaffmity Purification of PITP

All steps were performed on ice or at 4°C, and ice-cold reagents were used, unless otherwise stated. Transfected cells were washed with PBS and lysed in non-denaturing Lysis Buffer (50 mM Tris-HCl pH7.4, 150 mM NaCl, 50 mM NaF, 1% Triton X-100, 2 mM EDTA, 2 mM EOT A, 250 pM sodium orthovanadate, 10 pg/ml aprotinin, 10 pg/ml leupeptin, 16 pg/ml benzamidine, 10 pg/ml pepstatin, 1 mM PMSF). Cell lysate was cleared by centrifuging at 14,000 rpm for 10 min and the supernatant was incubated with pre-swollen Protein A Sepharose (50% v/v in Lysis Buffer) on a rotary wheel for 30 min. The Protein A Sepharose was pelleted by spinning at 2000 x g for 5 min and the supernatant incubated with anti-FLAG M2 affinity gel on a rotary wheel for 1 h. (Before use, the affinity gel had been washed with O.IM glycine-HCl pH3.5 for not longer than 20 89 min and then equilibrated with Lysis Buffer.) The gel was then washed twice with Lysis Buffer and then twice with Ptdlns 4-kinase Buffer.

The amount of affinity gel used varied according to the expected amount of protein expressed. By comparing the intensity of Western blots, it was calculated that a 15 cm dish of HEK cells expressed about 3 pg (approximately 88 pmol) of recombinant wildtype PITPa. Since the binding capacity of anti-FLAG M2 affinity gel is about 10 nmol per ml of suspended affinity gel, approximately 8.8 pi of gel should have been sufficient to immunoprecipitate the expressed protein.

Typically, the level of expression of mutant T58D PITPa was only a quarter that of wildtype PITPa (see Fig. 4.12A). Hence, the amount of affinity gel used was proportionally reduced.

2.4 Assays for PITP, PLC and Lipid Kinases

2.4.1 PITP Assays

2.4.1.1 In vitro Ptdlns Transfer

2.4.1.1.1 Preparation of Reagents

The reagents, donor microsomes and acceptor micelles were prepared as described (Thomas et al., 1993; Thomas et a l, 1994). Three rat livers were homogenised in 150 ml of SET Buffer (0.25 M sucrose, 1 mM EDTA, 10 mM Tris-HCl pH 7.4) and then centrifuged, first at 2000 xg for 5 min and again at 10,000xg for 15 min, to pellet unbroken cells and mitochondria, respectively. After further centrifugation at 100,000 x g for 1 h, the microsomal pellet was resuspended and rehomogenised in 40 ml of Labelling

Buffer (50 mM Tris pH 7.4, 2 mM MnCl 2), after which 300 pCi of pH]-inositol was added for 1.5 h at 37°C to bring about exchange of the inositol headgroup on Ptdlns. To remove excess radiolabel, microsomes were pelleted at 100,000xg for 1.5 h and resuspended in 100 ml of 10 mM Tris-HCl pH 8.6 containing 2 mM inositol. They were re-centrifuged, resuspended in 100 ml of 1 mM Tris-HCl pH 8.6, 2 mM inositol, and re­ pelleted. The microsomal pellet was resuspended in 20 ml of SET Buffer and the protein concentration adjusted to 25 mg/ml, before storage in aliquots at -80°C. All steps were carried out at 4°C unless otherwise indicated.

90 To prepare the micelles, 784 pg of PtdCho and 16 pg of Ptdlns were dried in a glass tube and then resuspended using sonication in 1 ml of SET Buffer.

2.4.1.1.2 Ptdlns Transfer Assay

Microsomes were first diluted 20-fold with SET Buffer to a protein concentration of 1.25 mg/ml. Each assay was initiated by mixing 100 pi of microsomes, 100 pi of micelles and

50 pi of sample (adjusted to the required concentration). After incubation at 25°C for 30 min, assays were terminated using 50 pi of 0.2 M sodium acetate, 0.25 M sucrose, pH 5 to aggregate the microsomes. The solution was then vigorously mixed and centrifuged at 15,000 Xg for 15 min at 4°C, and 100 pi of the supernatant was mixed with 5 ml of Ready Safe^”^ liquid scintillation cocktail before counting.

2.4.1.2 Reconstitution of G-protein-regulated PLC Signalling

This assay was adapted from Cockcroft et al. (1994b) and Cunningham et al. (1995) and all steps were carried out at 4°C unless otherwise stated. HL60 cells were incubated for two days at 37°C with 1 pCi/ml [^H]-inositol in Medium 199 supplemented with 5 pg/ml insulin, 5 pg/ml transferrin, 4 mM glutamine, 100 units/ml penicillin and 100 pg/ml streptomycin. 50 ml of labelled cells (1-2x10^ cells/ml) were pelleted at 450 xg for 5 min and resuspended in 40 ml of PIPES Buffer (20 mM PIPES pH6.8, 137 mM NaCl, 3 mM KCl). Cells were suspended and pelleted again, and then resuspended in 4.5 ml of PIPES Buffer. Permeabilisation Solution (0.5 ml) was then added such that the final volume had 0.6 lU/ml streptolysin O, 2 mM MgATP at pCa7 (with 3 mM EOT A). The cell suspension was incubated for not more than 10 min at 37°C with occasional mixing; the incubation period was sufficient to deplete most of the endogenous PITP but retain most PLCp (Cunningham et a l, 1995). The permeabilised cells were then diluted with 40 ml of cold PIPES Buffer, sedimented at 2000 xg for 5 min and resuspended in 2 ml of PIPES Buffer supplemented with 4 mM MgATP, 4 mM 2,3-diphosphoglycerate, 20 mM LiCI, 4 mM MgCIi. The free calcium ion concentration was maintained at 1 pM.

20 pi of each protein sample was mixed with 5 ml of 90 pM GTPyS and 20 pi of cells was then added. The mixture was incubated at 37°C for 20 min before quenching with 1 ml of cold 0.9% NaCl. Cells were pelleted at 2000 xg for 5 min and 0.9 ml of supernatant was removed and assayed for labelled inositol phosphates (Section 2.1.3.1).

91 2.4.1.3 Reconstitution of EGF-dependent Phosphoinositide Kinases

This assay was adapted from Kauffmann-Zeh et al. (1995). A431 cells at 40% confluence in 6-well dishes were incubated for 24 h with DMEM supplemented with antibiotics, 5 pg/ml insulin and 5 pg/ml transferrin. Cells were washed with PIPES-Li buffer (PIPES Buffer containing 10 mM LiCl) and permeabilised for 20 min with PIPES-Li Buffer supplemented with Permeabilisation Solution at 37°C. The incubation period was sufficient to deplete A431 cells of endogenous PITP and PLCy, and to render them refractory to EGF-dependent PPI and inositol phosphate production (Kauffmann-Zeh et al., 1995). Cells were rinsed with cold PIPES-Li Buffer and treated for 2 min at 37°C with 2 mM MgATP, 10 mM LiCl and 2 mM MgCb at pCa6 (with 3 mM EGTA), and with or without 100 nM EGF, to stimulate EGFRs. Cells were then washed with cold PtdIns4K assay buffer (25 mM Tris-HCl pH7.5, 10 mM MgCb) and further incubated for 5 min at 37°C with PtdIns4K assay buffer containing 70 pM [y-^^P]ATP (10 pCi/ml). Reactions were terminated with an equal volume of PtdIns4K assay buffer containing 3% Triton X-100, followed by 300 pi of methanol, 200 pi of 1 M HCl and finally 300 pi of chloroform, and the entire mixture was vortexed for at least 5 s. The lower phase was washed with 400 pi of 1 M HCl and then removed for counting.

2.4.2 Ptdlns 4-Kinase Assay

This assay was adapted from Kauffmann-Zeh et al. (1994). Samples were incubated for 20 min at 37°C with PtdIns4K assay buffer containing 70 pM [y-^^P]ATP (5 pCi per reaction) or 200 pM unlabelled ATP, in the presence or absence of 1.5% Triton X-100 or

200 pM Ptdlns. Reactions were stopped by addition of 1 ml of hexane/isopropanol (13:7, v:v) and 0.2 ml of 2 M KC 1/concentrated HCl (8:0.25, v:v) and then vortexed. The organic phase was mixed with an equal volume of 0.1 M HCl and vortexed again. The final organic phase was extracted for further analysis.

PtdIns4K activity was assayed following SDS-PAGE of purified A431 plasma membranes at 4°C (Wetzker et a l, 1991; Kauffmann-Zeh et al., 1994). Each gel lane was sliced into 2-mm pieces and incubated overnight in Renaturing Buffer (25 mM Tris-HCl pH7.2, 30 mM MgCh, 1.5% TXlOO, 70 pM [y-” P]ATP (5 pCi per reaction)) with 200 pM Ptdlns or 10 pg of PITP at room temperature. Pis were extracted as described above.

2.4.3 PLC Assay

92 2.4.3.1 PLC Assay using Exogenous PtdlnsPa

This assay was developed based on conditions described by Paterson et al. (1997),

Homma and Emori (1997) and Wahl et al. (1992). ^^P-labelled Ptdlns ? 2 was synthesized by the type I PtdlnsPSK, STM7, using PtdIns4P and [y-^^P]ATP, and kindly provided by Dr. M. dos Santos. The lipid products were dried under nitrogen before resuspension and sonication in 4xPLC assay buffer (40 mM Tris-HCl pH7.4, 480 mM KCl, 20 mM

MgS 0 4 , 80 mM LiCl, 20 mM P-mercaptoethanol, 1% Triton X-100, 60 pM unlabelled

Ptdlns? 2, pCa5.5 (with 12 mM EGTA)). Each sample was mixed with the lipid suspension at a ratio of 3:1 (v:v) and incubated for 10 min at 37°C. Reactions were stopped using 500 pi of cold chloroform/methanol/concentrated HCl (50:50:1) and 150 pi of 1 M HCl, and the mixtures were centrifuged at 2000xg, 4°C, 5 min. 200 pi of the upper, aqueous phase was counted

2.4.3.2 PLC Assay using Endogenous, Labelled PtdInsP 2

Sf9 cells were incubated for six days (HA growth cycle) in IPL-41 Medium (without inositol), supplemented with 10% dialysed PCS, lipid concentrate (1X), myo-[2- ^H]inositol (1.9 pCi/ml), penicillin (100 units/ml) and streptomycin (100 pg/ml). Cells were then infected and PITP purified as described in Section 2.3.2. PITP preparations were then washed and incubated in PLC Buffer (10 mM Tris-HCl pH7.4, 120 mM KCl, 5 mM MgS 0 4 , 20 mM LiCl, pCa7 (with 3 mM EGTA)) containing 2 mM ATP for 15 min at 37°C. Inositol phosphates were extracted as described in Section 2.4.3.1.

2.4.4 DAG Kinase Assay

DAGK activity was measured according to Sakane et a/. (1991) and Klauck et al. (1996) with some modifications. Each sample (typically 50 pi) was incubated for 20-30 min at

30°C with an equal volume of DAGK Buffer (25 mM Tris-HCl pH 7.4, 18 mM MgCli, 2 mM EGTA, 2 mM DTT, 4 mM DAG, 100 mM octyl glucoside, 2 mM deoxycholate) containing either 2 mM unlabelled ATP and l-stearoyl-2-[l-^"^C]arachidonyl-5«-glycerol (0.5 pCi per reaction) or 140 pM ATP (20 pCi per reaction). Reactions were stopped by addition of 1 ml of hexane/isopropanol (13:7, v:v) and 0.2 ml of 2 M KCl/concentrated HCl (8:0.25, v:v), and then vortexed. The organic phase was removed to fresh tubes, mixed with an equal volume of 0.1 M HCl and vortexed. The final organic phase was extracted for analysis by TLC (Section 2.1.3.2.1).

93 Chapter 3

Reconstitution of Polyphosphoinositide Signalling by Both Isoforms of

PITP

3.1 Introduction

As mentioned in Chapter 1, the concept of signalling molecules being organized into functionally distinct complexes has been gaining acceptance in recent years. Many studies relied on cell lysis and the isolation of stable, agonist-triggered signalling complexes to reveal information on the components. However, this approach does not reveal how the supply and consumption of substrate or precursor is controlled. At the same time, in vitro assay conditions which use excess exogenous substrate, such as phospholipids in micelles, often ignore the effects of the physical environment of the cell on enzymatic activity and may compromise our understanding of the signalling process. In addition, the existence of multiple isozymes further complicates attempts to designate a specific functional role to any individual isoform.

Streptolysin O (SLO) is a bacterial cytolysin which sequesters cholesterol and generates plasma membrane lesions sufficiently large to allow the leakage of cytosolic proteins of molecular mass up to 400 kDa (Buckingham and Duncan, 1983). When treated with SLO, permeabilised cells will be depleted of cytosolic components but still retain their native membrane architecture and subtle compartmentation. Purified and/or recombinant proteins can then be assayed, using these cells, to elucidate their roles in various cellular processes. This approach can be useful for understanding the supply and demand of lipid signalling molecules.

This approach has been used to identify, for the first time, the functional roles of PITP in mammalian cells. After treatment with SLO to deplete their cytosolic contents, human pro-myelocytic HL60 cells become refractory to GTPyS-stimulation of PLCp (Cockcroft et aL, 1994b). Rat brain cytosol was able to reconstitute GTPyS-mediated inositol lipid signalling, and the reconstituting factor was identified as PITPa (Thomas et aL, 1993). Subsequent work using permeabilised A431 human epidermoid carcinoma cells demonstrated that PITPa was also required for EGF-triggered Ptdlns phosphorylation and inositol phosphate production (Kauffmann-Zeh et aL, 1995). Thus, PITPa was identified

94 to be essential for PLC-mediated PI signalling by two major classes of cell surface receptor.

The above observations supported the “replenishment model”, in which the role of PITP is to transport Ptdlns from its site of synthesis in the ER and replenish Ptdlns in the plasma membrane, which has been depleted by phosphorylation and hydrolysis (Downes and Batty, 1993; Michell, 1975). Indeed, PLCpi alone is capable of restoring GTPyS- stimulated inositol phosphate production in SLO-permeabilised HL60 cells, presumably by using inositol lipids already present in the plasma membrane (Allan and Quinn, 1989; Cunningham et al., 1995). However, addition of exogenous PITPa greatly enhanced the initial rate of In sf 3 production (Cunningham et aL, 1995). Furthermore, PITPa was also found to co-immunoprecipitate with the EGFR, type II PtdIns4K and PLCy in an EGF- dependent manner (Kauffmann-Zeh et aL, 1995). These data are not consistent with the model in which PITPa randomly exchanges its bound Ptdlns for a molecule of PtdCho in the plasma membrane. Instead, it was proposed that PITPa is a component of a multi­ enzyme PI signalling complex, whose components include PtdIns4K, PtdInsR5K and PLC, and its function is to present Ptdlns substrate to the lipid kinases and lipase (Cunningham et aL, 1995; Hsuan, 1993; Liscovitch and Cantley, 1995). This “co-factor model” postulates that PITP-bound Ptdlns is the preferred substrate for the PI- metabolising enzymes and is metabolically distinct from PI lipids resident in the plasma membrane. This model may provide, at least in part, the mechanism for the segregation of agonist-responsive and -unresponsive inositol lipid pools (Monaco and Gershengom, 1992; see also Section 1.6).

3.2 Cloning and Expression of PITP

Just before this project started, the p isoform of rat PITP (PITPP) was cloned by complementation of temperature-sensitive sec 14 mutations in S. cerevisiae (Tanaka and Hosaka, 1994). Its amino acid sequence shows 94% similarity (77% identity) with that of rat PITPa (Dickeson et aL, 1989) and it is expressed highly in rat brain, kidney, liver and lung. At that time, PITPp was relatively uncharacterized biochemically and further work was required to understand its functional properties, especially in relation to those of PITPa.

95 In order to compare the biochemical and physiological properties of PITPa and PITPp, they were expressed in systems that could potentially produce large amounts of each recombinant protein. A bacterial host was chosen in the first instance for simplicity.

The pET System is a powerful system for expression of recombinant proteins in E. coli. In pET plasmids, target genes are under the control of the bacteriophage T7 promoter and their expression are induced by T7 RNA polymerase in an appropriate host containing a chromosomal copy of the T7 RNA polymerase gene under lacUV5 control. Addition of IPTG to the bacterial culture induces expression of the polymerase, which in turn transcribes the target gene. In addition, target genes are transcriptionally silent or poorly expressed in the uninduced state. This tight control of expression will minimise plasmid instability, especially if the recombinant protein is toxic to the host cell. The ORFs of both PlTPs were first amplified from cDNA libraries and eventually sub-cloned into pET vectors downstream of the T7 promoter. Protein expression and purification conditions were then optimised to obtain highly-purified recombinant PITP proteins for further biochemical characterization.

3.2.1 Cloning

Similar strategies were employed for cloning human PITPa (hPlTPa) and rat PITPp

(rPlTPp). The cDNAs for the two PlTPs were amplified by PCR, restriction-digested and spliced into the cloning vector pBluescript 11 SK+ (pBSll; Stratagene).

PCR primers (Section 2.1.1.1.1) were designed with reference to the cDNA sequences of hPlTPa and rPlTPp (Dickeson et aL, 1994; Tanaka and Hosaka, 1994). Each primer contained one or more unique restriction sites, a nucleotide sequence of appropriate annealing temperature (Tm), and complementarity to either end of each ORE. The choice of restriction sites was influenced by the possibility that a second expression system, the glutathione S-transferase expression system (Pharmacia), would be employed, should the pET plasmids fail to produce PITP proteins of sufficient quantity or quality.

3.2.1.1 Cloning Human PITPa

The template used for cloning hPlTPa was a human liver carcinoma HepG2 cDNA library. In the first PCR amplification, faint products between 500 and 1000 bp were observed. These were gel-purified and re-amplified using the proof-reading Venf^ DNA polymerase to generate more DNA for further manipulation. Products were obtained at

96 the expected size of approximately 800 bp (data not shown). These were purified, digested with BamRl and SaR, ligated with the corresponding restriction sites in pBSII, and transformed into XLl Blue cells.

Nine white colonies from each ligation were randomly selected and screened by PCR using primers aP and a3. Five positives were sequenced and only one of these was free of mutations. This clone, hereafter called 3H4, was used for subsequent expression studies.

The 3H4 ORF was sub-cloned into the bacterial expression vector pET21a using Ndel and SaR restrictions sites and transformed into BL21(DE3) cells. The Ndel site incorporated the ATG translation start codon, while the SaR site allowed in-frame translation beyond the 3' end of hPITPa coding region to include the Hise tag at the C- terminus of the protein.

3.2.1.2 Cloning RatPITPp

A rat brain cDNA library was used as the template. Using a PCR hybridisation temperature (Tm) of 59°C, very small amounts of products in the expected 800 bp range were produced. The comparatively small amount of PCR product indicated that PITPa may be more abundantly expressed than PITPp. Gel-purified products were subjected to a second PCR amplification, in which the Tm was increased to 62°C for more specific hybridisation. The products were purified, digested with BamlAl and SaR, ligated into corresponding sites in pBSII, and transformed into XLl Blue cells. Only two bacterial colonies were obtained.

Plasmids were prepared from both colonies and sequenced. Frame-shift mutations were detected in both clones: an extra adenosine (A) at nucleotide 356 in the first clone (hereafter called Bl) and the omission of a cytidine (C) at nucleotide 692 in the second clone (hereafter called B2).

Hence, a chimera of the 3' portion of clone Bl and the 5' portion of clone B2 was constructed to eliminate the mutations. The restriction site BstXl was chosen as the site of splicing because it is (i) unique in the rPITPp ORF at nucleotides 388-399, and (ii) within the multiple cloning site of pBSII. A 420bp BstXl fragment from B2 comprising the unmutated 5' region of the rPITPp ORF was ligated with a 3312bp BstXl fragment of Bl

(which contained the unmutated 3' region of the rPITPp ORF), and transformed into XLl

97 Blue cells. Restriction-digestion analysis with BamWl and SaR identified two purified plasmids containing unmutated rPITPp cDNA in the correct orientation.

Finally, the rPITPp cDNA was subsequently sub-cloned into the expression vector pET21d using Ncol and SaR restriction sites, and transformed into BL21(DE3) cells. The Ncol site incorporated the translation start codon ATG, and the SaR site again allowed the inclusion of the His6 tag at the C-terminus of the recombinant rPITPp protein.

3.2.2 Expression and Purification of PITP

Selected clones (Sections 3.2.1.1 and 3.2.1.2) were induced for 3 h with 1 mM IPTG at 37°C to determine their level of PITP expression (Fig. 3.1). Bacterial cells were then lysed and PITP extracted from the cleared lysate using Ni^"^-NTA resin (Sections 2.3.1.1 and 2.3.1.2), The purified proteins migrated at their expected molecular mass of 35 kDa (PITPa) and 34 kDa (PITPp) and protein sequencing (kindly performed by Nicholas Totty) confirmed their identities (data not shown).

By comparing the staining of purified PITP proteins with standards on SDS-PAGE, the estimated yield of the colonies showing highest PITP expression (a4 and bl in Fig. 3.1) was less than 700 pg per litre of culture. This was unusually low compared to the typical yields of 100 mg/1 culture for other soluble proteins (pET System Manual and references therein). This could be partly explained by the observation that most of the expressed PITP proteins precipitated as inclusion bodies (Fig. 3.1).

Bacterial cultures overexpressing rPITPp typically grew slower than those overexpressing hPITPa and took -30-40% longer to reach an A ôoo of 0.6. Lower amounts of soluble rPITPp were also produced per volume of culture for the same length of IPTG-induction time. This phenomenon of higher yields of PITPa relative to PITPp was also observed in insect and mammalian expression systems (see Chapter 4).

One option to obtain more soluble protein was to renature the insoluble PITP in the inclusion bodies. However, other laboratories had only managed to partially restore the functional activities of renatured PITPa: although the renatured protein could transfer phospholipid in vitro, it was unable to reconstitute PLC signalling in permeabilised HL60 cells (S. Cockcroft, personal communication).

98 hPITPa rPITPp cell pellet

al a2 a3 a4 bl b2 b3 b4 b5 A B kDa : c f t --v't — 97.4

— 66

— 45

PITP

— 31

Fig. 3.1 Comparison of Protein Yields of Selected Clones Colonies were cultured in 30 ml LB medium at 37°C to an A^qq of 0.6. Protein expression was then induced with 1 mM IPTG for 3 h and the cells lysed in non­ denaturing Sonication Buffer. Recombinant PITP was purified using NP^-NTA resin and then examined using SDS-PAGE on a 10% gel. Insoluble cell lysate debris from bacteria expressing (A) hPlTPa and (B) rPlTPp were also analysed for PITP in inclusion bodies.

99 The second option was to reduce the incubation temperature and IPTG concentration to enhance solubility of the protein. Incubation temperature often affects both expression level and protein solubility, and lower temperatures can lead to increased yields of soluble and active proteins (Schein, 1989; Schein and Notebom, 1988). However, there appeared to be only small differences in the final amount of soluble PITP when the bacterial cultures were induced at lower temperatures and IPTG concentrations (data not shown). A low level of PITP solubility in the bacterial protoplasm was inferred. Consequently, any further adjustment of the external environment would probably not have increased the amount of soluble and functional PITP. Thus, the IPTG concentration was kept to 0.1 mM and the temperature at room temperature (approximately 20°C).

Since the recombinant PITPs were purified under native conditions, there is a higher potential for binding bacterial contaminants. To minimise non-specific binding, the stringency of the Wash Buffer was increased (Section 2.3.1.2), by adjusting the concentrations of sodium chloride and glycerol to reduce non-specific ionic and hydrophobic interactions, respectively. Despite washing the resin with such stringent conditions, the PITPs were only about 90% pure (Fig. 3.2).

To remove the remaining contaminants the PITP proteins were gel-filtered through a Superdex 75 16/60 column following elution from the NP^-NTA resin. Surprisingly, the gel filtration profiles of recombinant hPITPa and rPITPp were different (Figs. 3.3 & 3.4).

For hPITPa, most of the protein eluted as a peak at around 75 ml post sample injection (Fig. 3.3A). Calibration of the gel filtration column indicated that the apparent molecular mass of this peak was approximately 19 kDa. This differs from the calculated molecular mass of recombinant PITPa (33 kDa) and its behaviour as a 35-kDa protein in SDS- PAGE (Fig. 3.3B). A larger-than-expected elution volume was also observed during gel filtration of native PITPa from bovine brain (Thomas et al., 1994). Furthermore, recombinant hPITPa purified from Sf9 cells also displayed this phenomenon (Section 5.3.1).

In addition, there was a broad shoulder ahead of the main peak (Fig. 3.3A). SDS-PAGE revealed the presence of hPITPa and contaminants with molecular masses of 28 kDa and 70 kDa in this shoulder (Fig. 3.3B). The latter may be the bacterial HSP70 heat shock protein homolog, DnaK.

100 room temperature 30°C

A FT W 5 20 El E2 M FT W 5 20 El E2

kDa 97 —

66 —

45 —

hPITPa

31 —

room temperature 30°C B FT W 5 20 El E2 M FT W 5 20 El E2

kDa 97 —

66 —

45 —

rPlTPp 31 — I

Fig. 3.2 Optimisation of Conditions for Recombinant PITP Purification 100 ml of clones (A) a4 or (B) bl were induced with 0.1 mM IPTG at room temperature or 30°C and then lysed and incubated with Ni^^-NTA resin for 1 h. The resin was washed with Wash Buffer containing 0 mM (W), 5 mM (5) or 20 mM (20) imidazole before PITP was eluted (El, E2). The loadings in each SDS-PAGE gel were proportional except for the unbound cell lysate (FT), which is 1:8 to others. M, molecular weight standards.

101 Chymotrypsinogen A (25 kDa) 0.75 -

280 0,50 - Ovalbumin Ribonuclease A (43 kDa) (13.7 kDa) 0.25 -

Fraction 1 510 15 20 25 30 35 40 i-,------1------L_------1------U ------L Volume (ml) 65 70 75 80

Fractions

B M FT W E 2 4 15 26 28 31 36 39 40 41 42 43 kDa 97 —

66 —

45 —

# # m -

21.5 — %

Fig. 3.3 Gel Filtration Chromatography of hPITPa Recombinant hPITPa on Ni^^-NTA resin was washed (W) and eluted (E) as described in Section 2.3.1.2, and further analysed (1 mg; 1.5 ml) using a Superdex 75 16/60 column (flow rate = 1 ml/min; in PIPES Buffer). Fractions (0.5 min) were (A) collected and (B) analysed in proportion using SDS-PAGE. The proportion between the unbound lysate (F), W and E was 1:1:5. The Superdex column was calibrated using a calibration kit from Pharmacia. The elution position of three molecular weight standards were indicated. The arrow in panel B indicates the major 70-kDa contaminant. M, molecular weight standards. 102 A

0.2 - 200kD 60kD 17.5RD

^280

0 - Fraction 1 510 15 20 25 30 35 40 Volume (ml) 40 45 50 55 60 65 70 75 80

B Gel-filtered Fractions

4 5 -

3 1 -

Fig. 3.4 Gel Filtration Chromatography of rPITPP (A) Purified recombinant rPlTPp (0.4 mg; 1.5 ml) was gel-filtered as described for hPITPa in Fig. 3.3. One fraction was collected per min. (B) rPITPp fractions were analysed using SDS-PAGE and then silver-stained to better detect minor components. Fractions 19 and 20 were duplicated, while fractions 32-37 were stained for a longer period to check for contaminants. 103 In contrast, the gel filtration profile of rPITPp showed that the protein eluted in three distinct peaks, with apparent molecular masses of 200 kDa, 60 kDa and 17.5 kDa (Fig. 3.4A). Thus, PITPp also did not elute at its calculated molecular mass of 32 kDa. Nevertheless, the peaks, especially the latter two, appeared to be similar to those observed for rPITPp purified from Sf9 cells and may indicate a constitutive organization of PITPp into discrete oligomers when overexpressed (see Section 5.3.1.2). SDS-PAGE of the middle peak revealed a weakly-stained protein band (-28 kDa), which corresponded in size with one of the contaminants in PITPa. No protein (such as the aforementioned 70- kDa contaminant) was detected at the higher molecular mass range.

The aforementioned contaminants in the PITP preparations were in relatively small amounts and were not pursued further.

3.3 Functional Comparison of PITPa and PITPp.

3.3.1 In vitro Ptdlns Transfer

The ability to shuttle Ptdlns (and PtdCho) between membranes in vitro is the hallmark of PtdlnsTPs (Wirtz, 1991). To test the integrity of the purified recombinant PITPs, their relative abilities to transfer Ptdlns were investigated. The experimental protocol and preparation of reagents (Section 2.4.1.1) were adapted from those published elsewhere (Thomas et ah, 1993; Thomas et al., 1994). Microsomes were purified from homogenised rat livers by ultra-centrifugation and then labelled with [^H]inositol by headgroup exchange with microsomal Ptdlns. Micelles were prepared by sonicating and resuspending dried PtdCho/Ptdlns mixtures (98:2 (w:w) PtdCho/Ptdlns) in SET Buffer (Section 2.4.1.1.1). In the assay, the transfer of [^H]PtdIns from donor microsomes to acceptor micelles was measured.

Gel filtration fractions corresponding to each peak containing hPITPa and rPITPp were pooled and concentrated using centrifugal concentrators. Recombinant hPITPa (fractions 23-30 in Fig. 3.3; “al” in Fig. 3.5A) was able to transfer Ptdlns at a similar rate to rat brain PITPa (S. Cockcroft, personal communication), with maximal activity at a hPITPa concentration of 50-100 pg/ml. Therefore, the presence of the 6-histidine peptide at the

C-terminus of the protein did not appear to affect its lipid transfer ability. hPITPa forming the broad shoulder preceding the main peak in Fig. 3.3A, seems to be able to

104 A 5000

^ 4000-

g J 3000- § 1 al a2 O 2000 - T3 i 1000 -

0 50 100 150 200 250 PITPa concentration (jug/ml)

B 5000

4000-

3000- bl II8 a b2 o 2000 -

§ 1000 -

0 25 50 75 100 125 PITPp concentration (pg/ml)

Fig. 3.5 In vitro Ptdlns Transfer by PITP Gel filtration fractions 23-30 and 5-18 of PITPa (A; Fig. 3.3) were separately pooled and concentrated using centrifugal concentrators, and denoted 'aT and 'a2% respectively. In the case for PITPp (B; Fig. 3.4), fractions 32-37 were pooled, concentrated and denoted 'bT, while fractions 4-8 and 17-22 constituted ’b2'. The concentrates were assayed in triplicates at different concentrations for their Ptdlns transfer activity (Section 2.4.1.1.2). The data is representative of five independent experiments. 105 transfer Ptdlns equally well (fractions 5-18 in Fig. 3.3; “a2” in Fig. 3.5A), although a complete dose curve could not be established due to the low amount of material.

The first two rPITPp elution peaks (fractions 4-8 and 17-22 in Fig. 3.4; “b2” in Fig. 3.5B) were pooled and compared with the third peak (fractions 32-37 in Fig. 3.4; “b l” in Fig. 3.5B), There appeared to be no significant difference in their Ptdlns transfer abilities. Furthermore, the PITPp Ptdlns-transfer dose curves were very similar to those of hPITPa. This is consistent with reports that demonstrated equal in vitro Ptdlns transfer activity of PITPa and PITPp purified from bovine brain (de Vries et al., 1995).

3.3.2 Reconstitution of PI Signalling

To further compare the functional properties of the two recombinant isoforms of PITP, their respective abilities to reconstitute PI signalling via PLC were tested in both A431 and HL60 cells.

3.3.2.1 Reconstitution in A431 Cells

This laboratory had demonstrated previously that PITPa was essential for restoring EGF- stimulated PPI production in permeabilised A431 cells (Kauffmann-Zeh et al., 1995). PITPa was then the only known cytosolic PITP in higher eukaryotes (Wirtz, 1991). With the subsequent cloning of PITPp, it became necessary to determine whether or not there is any specificity for PITP isoforms in agonist-induced PI signalling.

The reconstitution assay was performed as described in Section 2.4.1.3. While an intact C-terminus of PITPa has recently been reported to be important in inositol lipid signalling (Hara et al., 1997; Prosser et al., 1997; Tremblay et al., 1996; Tremblay et al., 1998), the presence of the His6 tag (plus a nine-residue linker) did not appear to affect the ability of the protein to stimulate PI kinases in cytosol-depleted, EGF-stimulated A431 cells (Fig. 3.6), and the concentration (approximately 100 pg/ml) required for maximum kinase activity is similar to those reported for native PITPa (Kauffmaim-Zeh et al., 1995).

A more detailed comparison of PtdlnsP and Ptdlns ? 2 synthesis in permeabilised A431 cells revealed no obvious difference between the isoforms of PITP (Fig. 3.7). Furthermore, the forms of PITPp present in the different elution peaks shown in Fig. 3.4

106 I 2500

c/5 M Ii fS

100 150 200 250

PITPa concentration (|ig/ml)

Fig. 3.6 Restoration of EGF-stimulated Ptdlns Phosphorylation by PIT Pa Serum-starved A431 cells were permeabilised and then treated with (squares) or without (circle) EGF as described in Section 2.4.1.3. PITPa (“a l” in Fig. 3.5) was then added at various concentrations, together with 70 pM [y-^^P]ATP (10 pCi/ml), to assay its ability to restore Ptdlns phosphorylation (Section 2.4.1.3). Phosphoinositides were extracted and counted (mean ± SD; « > 2). A PITP dose response in EGF-untreated cells could not be completed due to the limited amount of PITP available. The data is representative of three independent experiments.

107 A o 5 X

4 I ê 3 g

<+H 2 I 1 EGF+ EGF+ EGF+ control EGF hPITPa rPITPp rPITPp (b l) (b2)

% 30 - B X

25 -

1 20 - i 15 -

1 10 - .2 % ! 5 - EGF+ EGF+ EGF+ control EGF hPITPa rPITPp rPITPp (b l) (b2)

Fig. 3.7 Comparative Reconstitution of Phosphoinositide Kinase Activities in A431 by PITPa and PITPP Serum-starved A431 cells were permeabilised and then treated with or without EGF (Section 2.4.1.3). 100 jLtg/ml of PITPa ("al" in Fig. 3.5) or 50 pg/ml of PITPp ("bl" or "b2" in Fig. 3.5) were added where indicated with PtdIns4K assay buffer containing [y-^^P]ATP. Polyphospho­ inositides (A, PtdlnsP; B, Ptdlns? 2 ) were extracted, resolved by TLC and counted (mean ± SD; n = 3). The data is representative of three independent experiments. 108 were also able to reconstitute both Ptdlns 4-kinase and PtdlnsP 5-kinase activities with similar efficacy (Fig. 3.7).

Lastly, the dose response of cellular Ptdlns kinase activities to PITP was similar to that of the in vitro Ptdlns transfer activity (compare Figs. 3.5 and 3.6). This suggests that (i) the differences in protein sequence between PITPa and PITPp (77% identity) have no apparent bearing on their respective Ptdlns transfer or signalling potential, and (ii) a property of PITPs that was required to restore Ptdlns kinase activities was their Ptdlns transfer ability.

3.3.2.2 Reconstitution in HL60 Cells

The requirement for PITP in PI signalling was first demonstrated in HL60 cells which were labelled with [^H]inositol. After permeabilisation by SLO, PITPa was found to be essential for restoring PLCp-mediated inositol phosphate production in these cells

(Thomas et a l, 1993). As in the case of A431 cells, maximal GTPyS-stimulated PLCp activity was restored by similar concentrations (100 |ig/ml) of hPITPa and rPITPp (Fig. 3.!%.

3.4 Discussion

In this chapter, recombinant PITPa and PITPp were compared and some unexpected results were obtained. cDNAs for hPITPa and rPITPp were amplified from human and rat cDNA libraries, respectively. Comparatively, more PITPa PCR products were obtained than that for PITPp, even after two rounds of amplification. This reflects the generally lower expression of PITPp in many cells when compared to PITPa, and may partly explain the failure to detect and/or isolate PITPp in the many reports of PITPa purification from several different tissues, including blood platelets, bovine brain and heart (see for example, Thomas et a l, 1994; van Paridon et a l, 1987b).

Incidentally, the overexpression of PITPp in bacterial, insect and mammalian cells were also lower than those of PITPa, despite using the same host cells and expression vectors (and thus the same expression regulatory elements) (see Section 3.2.2 and Chapter 4). In addition, host cells overexpressing PITPp exhibited a slower growth rate than those overexpressing PITPa. Van Tiel et al. (2000) also reported an increase in the doubling time of mouse NIH3T3 fibroblasts that overexpressed mouse PITP p. It is unclear why

109 A 8000

6000- È

I 4000- I

I 2000 -

100 200 300 400 500 600 PITPa concentration (pg/ml)

B 8000

T ■o _L I 6000-

I 4000- I

I 2000 - -n

100 200 300 400 500 600 PITPp concentration (^g/ml)

Fig. 3.8 Reconstitution of Phosphoinositide Signalling in HL-60 cells Serum-starved HL-60 cells were labelled with [^HJinostiol for 48 h and then permeabilised (Section 2.4.1.2). (A) hPITPa or (B) rPITPp were added in the presence (circle) or absence (square) of 10 pM GTPyS and the production of inositol phosphates was measured (mean ± SD; n = 3). The data is representative of two independent experiments. 110 intracellular levels of PITPp are kept comparatively low and, when increased, slowed cellular growth rate. One may speculate that high PITPp concentrations affect (i) unidentified but common metabolic pathways and/or (ii) membrane properties in both prokaryotic and eukaryotic cells.

The yields of soluble PITPs from bacterial cultures (typically 500-700 pg/1) were up to two orders of magnitude lower than those reported for other proteins (pET System Manual and references therein). Wirtz and co-workers reported a yield of up to 3 mg/1 of recombinant mouse PITPa that was expressed in E. coli (Geijtenbeek et a l, 1994). This could perhaps be due to the use of different expression vectors, host strains, culture media, incubation and expression conditions, and purification protocols. In our experience, more than 90% of recombinant PITPs were found as inclusion bodies. Adjustment of IPTG-induction and other external conditions did not improve protein yield in the soluble form, suggesting that PITPs may have low levels of saturation in the bacterial protoplasm. As bacterially-expressed PITPs mostly bind phosphatidylglycerol (Geijtenbeek et al., 1994; Hara et al., 1997; and data not shown), it is tempting to speculate that phosphatidylglycerol-loaded PITPs may be structurally unstable or poorly soluble and prone to forming insoluble aggregates at high concentrations.

Although highly similar in molecular mass and amino acid sequence and immunologically cross-reactive (de Vries et al., 1995; de Vries et al., 1996), PITPa and

PITPp eluted from Superdex gel filtration columns in dissimilar patterns (compare Figs. 3.3 and 3.4; see also Chapter 5). PITPa, with a molecular mass of 32 kDa, eluted as a 19- kDa protein, but migrated as a 35-kDa protein in SDS-PAGE. This phenomenon of eluting at a lower molecular mass was also previously observed in the purification of endogenous PITPa from bovine brain and heart (DiCorleto et al., 1979; Thomas et a l,

1994) and hence is unlikely to be an anomaly due to PITPa overexpression, the C- terminal His6 tag, the type of phospholipid bound, proteolysis or post-translational modification(s) in a different expression host. In addition, Cockcroft and co-workers observed that, while wildtype PITPa eluted with an apparent molecular mass of 20 kDa, putative dimers of PITPa that was truncated at the C-terminus region eluted earlier (i.e., with a higher apparent molecular mass) than a hypothetical dimer of full-length PITPa (Prosser et al., 1997). It is thus possible that the C-terminal region of.PITP may have limited affinity for the gel filtration matrix and partially retards its passage through the gel column.

Ill In contrast, PITPp peaked at three different elution points, none of which correspond with its expected molecular mass (32 kDa). The last peak, with an apparent molecular mass of 17.5 kDa, contained virtually pure rPITPp, which may be in an equivalent physical form to the PITPa in the single peak in Fig. 3.3. As PITPp is highly similar in sequence to

PITPa, it is likely to share tertiary structural features with PITPa and show affinity for the Superdex gel matrix, causing it to elute at a lower apparent molecular mass. SDS- PAGE and silver staining analysis of the two other peaks indicated the presence of minor contaminants. The high apparent molecular mass of these proteins hinted at the dimérisation and/or oligomerisation of PITPp. Interestingly, Wirtz and co-workers recently reported the dimérisation of rat PITPa in the ‘apo’ (i.e., phospholipid-free) state (Schouten et a l, 2002). During chromatographic fractionation of bovine brain cytosol, de Vries et al. (1995) also detected multiple peaks (apparent molecular masses undetermined) of immunoreactivity to an antibody that recognises PITPp, some of which were unable to transfer Ptdlns in vitro. In contrast, PITPp from all three peaks (Fig. 3.4) have Ptdlns transfer capability (Fig. 3.5). It is unclear what causes this difference in lipid transfer capability, but it is possible that other bovine brain proteins in the same eluate fraction inhibited PITPp activity or the assay itself (de Vries et al., 1995).

The data reported above demonstrate clearly that PITPa and PITPp showed no apparent difference in the three assays of PITP function. PITPp was found to be able to transfer

Ptdlns in vitro as efficiently as PITPa and reached a similar maximum. This is consistent with findings reported elsewhere (Cunningham et al., 1996; de Vries et al., 1995; Segui et al., 2002; Westerman et al., 1995). The maximal Ptdlns transfer rate at 50-100 pg/ml indicates that, at this concentration, there was no longer any net transfer or exchange of Ptdlns and PtdCho between the microsomes and acceptor micelles by the end of the 30- min assay, and an equilibrium had been reached. This figure is dependent however on various experimental conditions, such as assay time.

Exogenous PITPp was able to restore agonist-stimulated PtdIns4K and PLC activities in permeabilised cells as efficiently as PITPa (Figs. 3.7 and 3.8), suggesting an overlap in functional capability between the two isoforms. This was also observed in other independent assays, such as the formation of secretory vesicles from the TON (Ohashi et al., 1995) and the stimulation of p i50 Ptdlns 3-kinase activity in vitro (Panaretou et al., 1997). Furthermore, the yeast PtdlnsTP, Secl4p, was also able to substitute for PITPa and PITPp in reconstituting G-protein-mediated PLCp and IgE receptor-stimulated PLCy 112 signalling in HL60 and RBL-2H3 cells, respectively (Cunningham et al., 1996). This occurred despite the lack of primary sequence homology between Secl4p and PITPs (Wirtz, 1991). The Ptdlns-binding/transfer capability of Secl4p appeared to be necessary for the substitution, at least in HL60 cells (Phillips et al., 1999).

The above data can be interpreted in several ways. Firstly, it could be that PITPs are merely transporting Ptdlns from the ER, where it is synthesized, to the plasma membrane, perhaps in the vicinity of activated receptors, to replenish the depleted Ptdlns pool during PLC-mediated signalling (Batty et al., 1998; Downes and Batty, 1993). Such a role would probably only require the Ptdlns transfer activity of the lipid transfer protein, which is the common biochemical property of PITPa, PITPp and Secl4p.

Another interpretation is that PITPs present Ptdlns to component enzymes of the PI signalling complex. When stimulated by agonists, the activities of these enzymes are low in the absence of PITPs, possibly due to the limited amount of Ptdlns in the surrounding membrane and/or such Ptdlns being a poor substrate compared to Ptdlns bound to PITP. The primary and tertiary structures of the transfer proteins may not be crucial in performing this function. Instead, the agonist-stimulated PI kinases and lipase within the signalling complex need only recognise the exposed inositol headgroup of Ptdlns on the surface of a juxtaposed PITP in order to phosphorylate or cleave it.

Neither interpretation requires the secondary and tertiary protein structures of PITP to play a significant role in its mechanism of signalling reconstitution. The crystal structures of SecI4p and PITPa were solved recently (Sha et al., 1998; Yoder et al., 2001). Although they have little structural similarity, the proposed mechanisms by which they mediate phospholipid exchange at the membrane surface share similar aspects. One part of the PtdlnsTP (helix A10/T4 of Secl4p and the C-terminal region of PITPa) is suggested to move away from the body of the protein to expose the bound lipid when it contacts the membrane lipid bilayer (Sha and Luo, 1999; Yoder et al., 2001). The protein- bound lipid is then exchanged for another phospholipid (Ptdlns or PtdCho) in the bilayer. It is thus tempting to speculate that, during PI signalling, the exposed Ptdlns on PITPa or Secl4p is recognised and phosphorylated by lipid kinase(s) within the signalling complex, instead of being deposited into the membrane bilayer. This is in agreement with the functional model proposed by Wirtz and co-workers, that the inositol moiety of Ptdlns would be exposed at one end of PITP’s lipid-binding site and is accessible to Pl-specific enzymes, including kinases and lipases (Schouten et al., 2002).

113 It is noted that the presence of the Hise tag, plus a two-residue linker, at the C-terminus of PITPa and PITPp did not appear to affect or inhibit in vitro Ptdlns transfer or the reconstitution of PI signalling. Studies have shown that the C-terminal region modulates the affinity of PITPa for vesicles and the efficiency of Ptdlns transfer (Hara et al., 1997; Tremblay et al., 1998) and helps maintain and stabilize the native structure of the protein (Voziyan et al., 1996). In particular, structural perturbations within the C-terminal region caused an increase in electrostatic interactions between PITPa and membranes and had variable effects on lipid transfer activity (Tremblay et a l, 1998). The apparent lack of effect of the His6 tag on both PITP isoforms in functional assays supports the notion that the C-terminal region has a highly flexible structure. This is consistent with its X-ray data (Schouten et al., 2002; Yoder et al., 2001).

Another case in point is that the deletion of 5 or 10 residues from the C-terminus of native PITPa rendered it impotent in restoring PLCp signalling but potent in inhibiting the signalling function of wildtype PITPa (Hara et al., 1997). As the truncated PITPa can still transfer Ptdlns, albeit less efficiently, and has no effect on the in vitro lipid transfer capability of wildtype PITPa, it appears to function as a dominant-negative mutant, blocking the association of wildtype PITPa with PI signalling enzymes. This suggests that the C-terminal region of PITP may play a crucial role in regulating interaction with the PI signalling complex, and also indicates that the reconstitution of PLC signalling involves more than just the passive transfer of Ptdlns to the plasma membrane. One may speculate that the deletion increased the association constant (or lowered the dissociation constant) of the PITPa with the complex (or the membrane bilayer) (Hara et al., 1997). This would be in agreement with the model proposed by Yoder et al. (2001) to describe the interaction between membrane and PITPa, in which the refolding and reassociation of the C-terminal region with PITPa provides a driving force for the dissociation of the protein from the membrane surface.

PITPa shares significant similarity in tertiary structure, but no obvious sequence similarity, with the START domain (Yoder et al., 2001). The START domain has been found in such proteins as pl22-RhoGAP, PtdCho transfer protein and the Gla2 homeodomain proteins, and it is proposed to bind lipids and interact with membranes (Ponting and Aravind, 1999). However, the lipid-binding cavity in the START domain of the MLN64 protein (Tsujishita and Hurley, 2000) is smaller than the cavity in PITPa (Yoder et al., 2001). This is consistent with the MLN64 protein binding the more

114 compact cholesterol molecule rather than diacyl phospholipids. Nevertheless, it will still be of great interest to investigate whether a Ptdlns-binding START domain is able to substitute PITP in reconstituting PLC signalling. Interestingly, pl22-RhoGAP has been shown to bind to and activate PLCô (Homma and Emori, 1995).

Non-specific lipid transfer protein (nsL-TP), also referred to as sterol carrier protein-2, was recently reported to possess a lipid-binding core similar to those of PITPa and MLN64 (Lopez Garcia et al., 2000). However, another nsL-TP, which can transfer Ptdlns in vitro, was unable to restore PLC signalling in HL60 cells (Cunningham, 1994). Furthermore, C-terminally truncated PITPa, which can still transfer Ptdlns, is unable to restore PLCp signalling (Hara et ah, 1997). Hence, it is unlikely that the reconstitution of PI signalling by PtdlnsTP depends on merely the net transfer of Ptdlns to the plasma membrane from the ER, or the non-selective enzymatic activity of lipid kinases and lipase on Ptdlns bound to any lipid binding protein.

A third interpretation presumes that Secl4p and PITPs share structural homology at one or more motifs that interact with one or more components of the signalling complex, and present the Ptdlns substrate as a constituent of the complex. For example, Hara et al. (1997) have proposed the existence of a ‘PITP receptor’ within the PI signalling complex. In both Secl4p and PITPa, at least one region of the protein has been postulated to be crucial for association with other proteins: the N-terminal domain of Secl4p and the regulatory loop of PITPa (Sha and Luo, 1999; Yoder et al., 2001, and references therein). However, these corresponding regions on Secl4p and PITPa do not share any apparent similarity in their primary or tertiary structures.

The complementation of function between Secl4p and PITP is not restricted to plasma membrane signalling. PITPa and PITPp were able to rescue temperature-sensitive sec 14 {seel4*^) mutants, but not null mutants (Skinner et a l, 1993; Tanaka and Hosaka, 1994). This complementation required, at least in part, the Ptdlns transfer capability of PITPs (Alb et al., 1995). Mutations of four residues which abolished PITPa Ptdlns transfer (but not PtdCho transfer) capability resulted in loss of ability to phenotypically rescue seel 4*^ growth defects in S. eerevisiae. The mechanism of complementation of Secl4p by PITPa and PITPp remains unclear, and is further complicated by the demonstration that Ptdlns- binding/transfer is dispensable for Secl4p function in vivo (Phillips et al., 1999). Conversely, Secl4p appeared to be able to substitute for PITPa and/or PITPp in (i) ATP-

115 dependent priming of Ca^^-triggered secretion (Hay and Martin, 1993), (ii) ATP- dependent formation of constitutive secretory vesicles and immature secretory granules from the TGN in vitro (Ohashiet al., 1995), (iii) formation from the TON of exocytic vesicles containing the poly(IgA) receptor (Jones et a l, 1998), (iv) ATP-independent scission of COPI-coated vesicles from the TGN (Simon et a l, 1998), and (v) stimulation of the pl50-PtdIns 3-kinase complex (Panaretou et ah, 1997). The ability of Secl4p to bind or transfer Ptdlns is apparently necessary to stimulate priming of secretion in substitution of PITPa (Phillips et al., 1999).

Most of these studies on complementation of function utilised permeabilised cells, to which exogenous PtdlnsTPs were added. As such, the transfer proteins may be delocalized from their normal sub-cellular targets. PITPa is found in both the cytosol and nucleus, while PITPp is primarily localized to the Golgi apparatus (de Vries et al., 1995; de Vries et al., 1996; see also Chapter 5). This differential localization of endogenous isoforms raised questions over the physiological significance of the demonstrated ability of various PITPs to reconstitute a PI metabolic process that occurs at a remote sub- cellular location. Indeed, PlTPp apparently was unable to compensate for the reduction in

PITPa level in the mouse vibrator mutation (Hamilton et al., 1997), while PITPp deficiency results in early embryonic death in mice (Alb et al., 2002).

Thus, identification of which endogenous PITP isoform participates in each distinct PI metabolic pathways requires careful further investigation. The recent cloning of multiple isoforms of the MrdgB subfamily will add further complexity to such studies.

116 Chapter 4

Presentation of Phosphatidylinositol to Lipid Kinases and Lipases by Phosphatidylinositol Transfer Proteins

4.1 Introduction

A primary difference between the ‘replenishment model’ and the ‘co-factor model’ of PITP function in PI signalling is the viability of the PITP-bound Ptdlns as a direct and preferred substrate for PI kinases. The latter model would therefore require (i) the localization of PITP at site(s) of PI signalling, (ii) the ability of PITP to associate in vivo with enzymes of the PLC signalling pathway, and (iii) the ability of these enzymes to utilise PITP-bound Ptdlns as a direct substrate.

This chapter will discuss experiments that were devised to test the above issues.

4.2 In vitro Interactions between PITP and Ptdlns 4-kinase

4.2.1 Interaction with Unpurified Ptdlns 4-kinase

To determine whether PtdIns4K can recognise and phosphorylate the exposed inositol headgroup of the PITP-bound Ptdlns, bacterially-expressed recombinant human PITPa

(hPITPa) still bound on Ni^^-NTA resin (see Section 2.3.1.2) was first mixed with Ptdlns micelles for 15 min to replace phosphatidylglycerol in the lipid-binding site of the transfer protein (Geijtenbeek et a l, 1994; Hara et a l, 1997; and data not shovm). After washing, purified A431 plasma membranes which are rich in type II PtdIns4K activity (Pike and Bakes, 1987) were added for a 30 min incubation, to allow any interaction to occur between the Ptdlns-loaded hPITPa and membrane proteins. Unbound membranes were then washed off and [y-^^P]ATP was added for 10 min. Lipids were then extracted and analysed by TLC.

A ^^P-Iabelled lipid was found to migrate to the same final position as the Ptdlns4f standard (Fig. 4.1 A, lane 2). In contrast, when hPITPa was absent, and Ni^^-NTA resin alone was sequentially incubated with Ptdlns and then plasma membrane preparations, no PPIs were detected (Fig. 4.1 A, lane 1). The result suggests association of hPITPa with one or more components of the plasma membrane that can phosphorylate Ptdlns and is consistent with the co-immunoprecipitation of the type II PtdIns4K and EGFR with PITPa in A431 cells (Kauffmann-Zeh et a l, 1995). The PtdlnsP generated was unlikely 117 A Ptdlns

PtdIns4P

Ptdlns(4,5)7^2

B 12

10

S

4

o 0 ■2

0 2 4 6 8 Ptdlns 4-kinase activity (cpm X 10'^)

Fig. 4.1 Phosphorylation of PITP-bound Ptdlns by Ptdlns 4-kinase (A) Approx. 20 pg of PITPa bound on 4 pi Ni^^-NTA (lanes 2 and 4) or 4 pi Ni^^-NTA alone (lanes 1 and 3) was incubated with 4 mM Ptdlns micelles for 15 min at room temperature and subsequently with purified A431 membranes for 30 min at 4°C. After washing off unbound membranes, 70 pM [y-^^P]ATP (5 pCi per reaction) were added for 10 min at 37°C. Type II PtdlnsPK (10 pg) was also added (lanes 3 and 4) for a further 10-min incubation at 37°C before lipid extraction and separation by TLC (Sections 2.4.2 and 2.1.3.2.1). This figure is representative of four separate experiments. Arrows mark the positions of the indicated lipid standards. (B) Following SDS-PAGE of purified A431 plasma membranes at 4°C, the gel was sliced into 2-mm pieces and incubated in Renaturing Buffer with 10 pg of PITPa pre-loaded with Ptdlns (Section 2.4.2). Lipids were extracted and counted. Positive gel distance is measured from a 45-kDa marker protein towards the cathode. The data is representative of three independent experiments. 118 to be due to major contamination caused by both micellar Ptdlns and Ptdlns kinase binding non-specifically to the resin and reacting upon ATP addition (see also Fig. 4.6). Nonetheless, it was unclear how hPITPa actually promoted Ptdlns phosphorylation. A

Ptdlns kinase might have bound PITPa and phosphorylated the protein-bound lipid.

Alternatively, PITPa might have increased the concentration of Ptdlns in the purified membranes, thereby stimulating Ptdlns kinase activity. Finally, PITPa might have allosterically activated one or more PI kinases.

A faint radioactive lipid band was observed co-migrating with the Ptdlnsf] standard (Fig. 4.1 A, lane 2). While the lipid might have been Ptdlnsfi, it could alternatively have been lyso-PtdlnsP, which migrates more slowly than Ptdlnsf on the TLC. To obtain more of this lipid for identification, a PtdlnsPK (kindly provided by S. Minogue) was added to the above reaction for a further 10 min incubation before lipid extraction. The PtdlnsPK used was then known as type II PtdlnsPSK but was subsequently redefined as a PtdIns5P4K (Rameh et al., 1997b). The results of independent sets of this experiment were, however, not consistent. In some, the intensity of the putative Ptdlnsfi band remained unchanged compared to reactions without additional incubation with the kinase; in others, there was a visible decrease (e.g.. Fig. 4.1 A, lane 4). Unfortunately, the radioactivity remained too low for accurate identification of the lipid by déacylation and HPLC (Section 2.1.3.2.2).

An interesting observation, especially in the light of subsequent experiments, was the presence of labelled lipid (at the same position as the PtdIns4P standard) in the Ni^"^-NTA resin control sample, which had been sequentially incubated with Ptdlns, plasma membranes and finally PtdIns5P4K (Fig. 4.1 A, lane 3). It was assumed that the kinase (then known as a type II PtdlnsPSK) may have phosphorylated, albeit non-specifically, residual Ptdlns on the resin to produce Ptdlns5f. It was subsequently reported that the PtdIns5P4K can weakly phosphorylate Ptdlns to produce PtdIns4P (Fruman et al., 1998). Hence, the uncharacterized lipid in lane 3 of Fig. 4.1 A was probably Ptdlns4f.

4.2.2 Interaction with a Partially Purified Ptdlns 4-kinase

PITPa had previously been shown to restore EGF-stimulated Ptdlns phosphorylation in permeabilised A431 cells and to co-immunoprecipitate with type II PtdIns4K in an EGF- dependent manner (Kauffmann-Zeh et al., 1995). However this enzyme was then not cloned or purified to homogeneity. Consequently, in order to determine whether type II PtdIns4K from A431 cells was able to interact with PITP in vitro and phosphorylate

119 PITP-bound Ptdlns, the kinase was purified by immunoprécipitation using the monoclonal antibody 4C5G. Although 4C5G inhibits enzyme activity (Endemann et a l, 1991), immunoprecipitated Ptdlns4K can be renatured following SDS-PAGE (Wetzker et a l, 1991).

Using recombinant hPlTPa, which had been pre-loaded with Ptdlns in these assays, PtdlnsP was generated in the region of the gel expected to contain the 5 5-kDa type 11 enzyme (Fig. 4.IB). This suggests that the Ptdlns bound on hPlTPa can be recognised and phosphorylated by type 11 Ptdlns4K in vitro. Whether the PI lipid remained bound to PITP after phosphorylation remains unclear. It was further noted that the presence of vestigial amounts of SDS in these assays might have denatured hPlTPa and thereby released the hPlTPa-bound Ptdlns. In addition, free Ptdlns may have been carried over with hPlTPa following loading with micellar Ptdlns. Thus, it was possible that Ptdlns4K had phosphorylated unbound Ptdlns.

4.3 Co-purification of PI Kinases with PITP

4.3.1 Co-purification of Ptdlns 4-kinase Activity

In order to avoid the possibility of carrying over free Ptdlns with PITPa following Ptdlns loading, bacterially-expressed PITP was replaced with hPlTPa and rPlTPp that were purified from overexpressing Spodoptera frugiperda Sf9 cells. (The original high-titre stocks of baculovirus that encoded hPlTPa or rPlTPp were generated by Dr. Damian Holliman.) Surprisingly however, in these experiments PtdlnsP was produced both in the presence and absence of renatured type 11 Ptdlns4K activity (e.g.. Fig. 4.2). The addition of [y-^^P]ATP to both PITP preparations was sufficient to trigger the generation of Ptdlnsf. This indicates that an endogenous Ptdlns4K had co-purified with PITP and appeared to phosphorylate PITP-bound Ptdlns. It was also inferred that PITP was associated with a Sf9 Ptdlns4K in vivo.

Ptdlns4K isozymes can be distinguished biochemically by their sensitivity to WT, adenosine and the detergent Triton X-100 (TXlOO) (see Section 1.5.3). The activity of the co-purified Ptdlns4K was resistant to WT but sensitive to adenosine, which is characteristic of a type 11 isozyme (Fig. 4.3). This identification is consistent with the experiments described above (Section 4.2) and the co-immunoprecipitation of PITPa with EGF-regulated type 11 Ptdlns4K from A431 cells (Kauffmann-Zeh et a l, 1995).

120 PtdIns4K activity (+ PITPa) (cpm X 1Q-^)

1 2 3 '

12

B £ 0

-6

-9

2.5 7.5 10 12.5

Ptdlns4K activity (+ micellar Ptdlns) (cpm X 10'^)

Figure 4.2 Co-Purification of Ptdlns 4-kinase Activity with PITP from Sf9 Cells Type II PtdIns4K was purified by SDS-PAGE fractionation of A431 plasma membranes and renatured as described in Fig. 4.1. The kinase activity was measured using micellar Ptdlns (solid bars) or PITPa overexpressed in Sf9 cells (open bars). Positive gel distance is measured from a 45-kDa marker protein towards the cathode. The data is representative of three separate experiments. 121 20

15 -

1 =

5 -

+ + Wortmannin + + + DMSO + + Adenosine + + TXlOO

Figure 4.3 Sensitivity of PITP-Associated Ptdlns 4-Kinase Activity from Sf9 Cells 10 |Lig of recombinant PITPa purified from Sf9 cells were assayed for co-purifying PtdIns4K activity against endogenous (i.e., PITP-bound) Ptdlns in the presence or absence of 2 pM wortmannin (in DMSO), 100 pM adenosine and 0.2% Triton X-100 (TXlOO). Lipids were extracted and counted by scintillation. The data is represent­ ative of two independent experiments. The data for PITPp are similar (not shown).

122 Another characteristic of type II PtdIns4K isozymes is their stimulation by TXlOO, but this was not observed with the co-purified PtdIns4K (Fig. 4.3).

4.3.2 Co-purification of Ptdlnsf 5-kinase Activity

Analysis of the extracted lipids on TLC also revealed the production of Ptdlns ? 2 (Fig- 4.4), indicating the simultaneous co-purification of a PtdlnsPK with PITP. Our laboratory had previously reported the co-immunoprecipitation of PITPa with type 11 Ptdlns4K, PLC and EGFR (Kauffmann-Zeh et al., 1995). Consequently, the above demonstration of an association between PITP and PtdlnsPK activity constitutes the last enzyme to complete the cofactor model of PITP function in PI signalling (see Section 1.5.5.4). The production of PtdOH and lyso-PtdOH in these assays will be discussed in Section 4.5.

PtdlnsP] production was often observed in PlTPp but rarely in PITPa preparations. This could be due to the generation of consistently larger amounts of PtdlnsP in PlTPp compared with PITPa preparations (Figs. 4.4 & 4.8). To determine whether PtdlnsP was further phosphorylated by the co-purifying PtdlnsPK, radiolabelling by carrier-free [y-

^^P]ATP was chased with a large excess of unlabelled ATP (Fig. 4.4). In both PITPa and

PITPp preparations, the radiolabel in PtdlnsPi increased. A slight decrease in the radiolabel in PtdlnsP was also detected in PITPa but not PITP p. The lack of a visible decrease in the latter was presumably due to its more intense PtdlnsP labelling. From the passage of the ^^P-label from PtdlnsP to PtdlnsP], it was inferred that PtdlnsP] was produced by the sequential phosphorylation of PITP-bound Ptdlns and not by phosphorylation of any unbound, contaminating PtdlnsP.

4.3.3 Confirmation of Lipid Identity

The identity of each lipid was confirmed by déacylation using methylamine and HPLC analysis using a Partisphere 5-SAX anion-exchange column (Fig. 4.5). (The HPLC analysis was kindly performed by Dr Frank Cooke.) This technique can distinguish between the various PtdlnsP] isomers known thus far: Ptdlns(3,4)P], Ptdlns(4,5)P], and Ptdlns(3,5)P] (Dove et a l, 1997). It can also clearly separate Ptdlns3P from Ptdlns4P. However, Ptdlns5P and Ptdlns4P are poorly separated on a conventional HPLC gradient as deacylated Ptdlns5P elutes only about 0.5 to 1 min later than deacylated Ptdlns4P (Rameh et ah, 1997b; Sbrissa et a l, 1999). The novel PI, Ptdlns5P, was first reported to be found in exponentially growing N1H3T3 fibroblasts and can be produced as a minor

123 PITPa PITPp

+ + + + [y-32p] ATP

+ + unlabel led ATP ^ ---- PtdOH

(lyso-PtdOH)

Ptdlns4P

Ptdlns(4,5)P2

Figure 4.4 TLC Analysis of Endogenous Lipids labelled using in vitro Lipid Kinase Assays 20 jj.g of each PITP preparation was incubated with 70 pM [y-^^P]ATP (10 pCi per reaction) for 20 min at 37®C (Section 2.4.2), after which half of each reaction was further incubated for 30 min at 37°C in the presence of 1 mM unlabelled ATP. Lipids were then extracted and separated by TLC. The figure is representative of four independent experiments.

124 A GroPIns4P 800

700

600

I 500 a Cu 400 rn(N 300

200

100

40 45 50 55 60 65 70 75 80 85 90 95 100 105 110 115 120 125 130 Fraction

B

4000 700

3500 600 3000 500 B I 2500 o 1 400 y T 2000 5 - I 300 1500 200 1000

500 100

1 6 1 1 16 21 26 31 36 41 46 Fraction

Figure 4.5 Identification of Lipids by Déacylation and Liquid Chromatography ^^P-labelled lipids in Fig. 4.4 were individually scraped off the TLC plate, deacylated and analysed using a HPLC Partisphere 5-SAX anion exchange column that was developed with a NaH2P0^ gradient (see Section 2.1.3.2.2). (A) Putative Ptdlnsf, open diamond', putative PtdIns/^2’ square. (B) Putative lyso-PtdOH, so lid triangle', putative PtdOH, open circle. The arrows indicate the elution positions of glycerophosphoinositol-4-phosphate (GroPIns4P), glycerophosphoinositol-4,5- bisphosphate (GroPIns(4,5)?2), and glycerol-3-phosphate (Gro3f). The above elution profiles are representative of two independent déacylation and HPLC experiments. The HPLC analyses were kindly performed by Dr Frank Cooke.

125 product in vitro through dephosphorylation of Ptdlnsfi by the SH2-domain-containing inositol phosphatase, SHIP (Rameh et a l, 1997b), or through phosphorylation of Ptdlns by PIKfyve, a novel PI 5-kinase (Sbrissa et a l, 1999), or type I PtdlnsPK (Tolias et a/., 1998b).

The déacylation product produced from the PtdlnsP shown in Fig. 4.4 co-eluted from the Partisphere column with the glycerophosphoinositol-4-phosphate (GroPIns4f) standard, thereby indicating that the original lipid was Ptdlns4? (Fig. 4.5). Nevertheless, the poor resolving power of this technique may have failed to distinguish between Ptdlns4f and Ptdlns5f.

To help solve this question, the PtdlnsP was treated with a type I PtdlnsPK, STM7, using unlabelled ATP and then analysed by TLC. Some of the ^^P-label co-migrated with the

Ptdlns ? 2 standard (data not shown), indicating that at least some of the Ptdlnsf in Fig. 4.4 was PtdIns4P, although the presence of PtdIns5P could not be ruled out. Unfortunately, PtdIns5f4K was not available when these experiments were performed.

4.3.4 PITP-PI Kinase Interactions are Specific

To determine whether the observed co-purification of Sf9 cell PtdIns4K and PtdIns4P5K was due to specific interactions with PITP, the presence of PI kinase activities was tested in preparations of the unrelated A3 3 protein, which had been expressed and purified using methods similar to those used for the PITPs. The human A33 antigen is found predominantly in human colonic and small bowel epithelium and is a transmembrane glycoprotein containing two putative immunoglobulin-like domains in its extracellular region (Heath et a l, 1997). The results show that no significant kinase activity associated with purified A33 (Fig. 4.6), which indicates that the presence of PI kinases in the PITP preparations was not due to non-specific binding to the Ni^'^-NTA resin by these kinases.

4.4 PITP-bound Ptdlns as a Substrate

4.4.1 Phosphorylation by PI Kinases

In a typical lipid kinase assay of recombinant PITP preparations from Sf9 cells described above, the amount of PPIs produced was estimated to constitute less than 1% of the estimated amount of Ptdlns bound to PITP. Consequently, it was important to establish whether or not the Ptdlns precursor arose from contaminating Ptdlns, which may have bound non-specifically to the Ni^^-NTA resin.

126 175

150 -

125 -

ï 100 -

I I 75 -

î 50

25

+ + + Ptdlns

PITPa PITPp A33

Figure 4.6 Specificity of Co-Purification of PI Kinases A33 protein was expressed and purified using methods similar to those used to prepare PITPs (Section 2.3.2). 5 pg of each protein were assayed for Ptdlns kinase activity in the presence or absence of exogenous Ptdlns (200 pM). Lipids were extracted and then counted. The data is representative of three independent experiments.

127 To answer this question, a point mutation was introduced into each PITP so that they could transfer only PtdCho. The substitution in PlTPa of the threonine residue at position 58 (T58) to an aspartate residue (T58D) had previously been found to abolish its ability to transfer Ptdlns, but not PtdCho, in vitro (Alb et al., 1995). These mutant PITPs were therefore prepared to determine whether PPIs were produced from PITP-bound or contaminating Ptdlns.

Baculoviruses were engineered to express the PlTPa and PlTPp mutants. cDNAs encoding these T58D PITPs with a C-terminal His6-tag were generated using overlapping PGR with primers that introduced the point mutation (Section 2.1.1.1.1). Following DNA sequencing, the cDNAs were spliced into the pVL1392 baculo-transfer vector and co­ transfected into Sf9 cells with BaculoGold DNA (see Section 2.2.1.2). The resultant supernatants were collected for viral-plaque assays and selected plaques were amplified twice to generate high-titre baculovirus stocks. The final titre of virus encoding T58D hPlTPa was 7.85 x 10* plaque-forming unit/ml (pfu/ml) while that for T58D rPlTPp was 3.25 X 10^ pfu/ml.

These mutant PITPs were expressed and purified using the same methods as used for their wildtype counterparts. Unexpectedly, the mutant PITPs were always produced in smaller quantities (up to 10-fold less) than their wildtype counterparts. The lower expression of mutant PITPs was also observed when they were expressed in mammalian cell lines (Section 4.8.2). This may indicate toxicity of the mutant PITPs to cells, because cells may well require the ratio of PITP-Ptdlns:PlTP-PtdCho to be within a viable range (see Discussion below). In order to affirm the loss of ability to bind Ptdlns, lipid extracts of purified T58D PITP purifications were separated by TLC and visualised with iodine staining (Section 2.1.3.2.1). Ptdlns was barely detectable while PtdCho constituted the main lipid (figure not shown). Ptdlns transfer assays (Section 2.4.1.1.2) further demonstrate that both T58D PITPs had no detectable ability to transfer Ptdlns in vitro (Fig. 4.7) and suggest that the sensitivity of Ptdlns transfer activity to mutations at residue T58 may be conserved among PITPs.

Addition of [y-^^P]ATP to preparations of the mutant PITPs show strikingly that there was no detectable production of PtdlnsP or Ptdlns? 2 , in sharp contrast to their wildtype counterparts (Figs. 4.8 and 4,10). This result clearly demonstrates either that the original precursor of PtdlnsP and Ptdlns ? 2 was Ptdlns that was specifically bound by PITPs, or that PI kinases did not bind to mutant PITPs.

128 5000 1 I 4000 “ o(U e o 3000 -

I2 2000 -

1000 -

pz

0 50 75100 12525 150 175 200

PITP concentration (|ig/ml)

♦ wildtype PITPa

o wildtype PITPp

■ T58D PITPa

□ T58D PITPp

Figure. 4.7 Ptdlns Transfer Activity of Mutant PITPs Purified PITPs were assayed for their Ptdlns transfer activity in vitro (mean ± SD; « = 3) (Section 2.4.1.1.2). The data is representative of four separate experiments.

129 (lyso-PtdOH)

Ptdlns4f

PtdIns(4,5)P2

PITPp PITPp PITPa PITPa (T58D) (T58D)

Figure 4.8 Assay of Phosphoinositide Synthesis by Mutant PITP Preparations 6 of each PITP were assayed in duplicate for lipid kinase activity against endogenous substrate (Section 2.4.2). Lipids were then extracted and separated by TLC. The figure is representative of four independent experiments. The positions of lipid standards are indicated.

130 4.4.2 Interactions are Independent of Lipid Type

In order to determine whether the lack of PPI production in mutant PITPs (Fig. 4.8) was due to the absence of PITP-bound Ptdlns or PITP-associated PI kinases, or both, exogenous Ptdlns was added as substrate. Fig. 4.9 shows the generation of PPIs in mutant PITP preparations. Furthermore, the similar amount of PPIs produced by each sample suggest the presence of similar levels of each PI kinase activity in corresponding wildtype and mutant PITP preparations. This result demonstrates the co-purification of PI kinases with mutant PITPs and suggests that the interactions between PITPs and the lipid kinases were not dependent on the type of lipid bound to PITP. These results confirm the inference in Section 4.4.1, namely, that PtdlnsP and Ptdlnsf] were both produced from PITP-bound Ptdlns, rather than that kinases were not present in preparations of mutant PITPs.

4.5 Co-purification of DAG Kinase

Analysis of the lipids extracted following kinase assays of wildtype PITP preparations also revealed two other lipids, which migrated faster than the PPIs on TLC (Fig. 4.4). Their deacylated products were both found to co-elute with a [*"^C]-glycerol-3-phosphate (Gro3f) standard on HPLC (Fig. 4.5). (The HPLC analysis was kindly performed by Dr Frank Cooke.) As these lipids appear to have identical headgroups, the difference in their migration on TLC was probably due to variation in their acyl chains. These lipids were thus tentatively identified as PtdOH and lyso-PtdOH (Fig. 4.4).

PtdOH is generated in vivo via several metabolic pathways. DAG, one of the two products of Ptdlns ? 2 hydrolysis by PLC, can be phosphorylated by DAGK to form PtdOH. In addition, PtdOH can be produced through hydrolysis of PtdCho by PLD, and by the acylation of phosphoglycerol. However, PtdOH produced by the latter two pathways would not have incorporated the radiolabelled y-phosphate of ATP. Thus, the presence of radiolabelled PtdOH (Fig. 4.4) was probably due to the action of a co- purifying DAGK.

Whereas lyso-PtdOH was always produced in both wildtype and mutant PITP preparations, PtdOH was only observed in assays of wildtype PITP preparations. However, the amount produced varied with different preparations. Indeed, PtdOH was not always detected (e.g.. Fig. 4.8), and was never observed if Ptdlnsf 2 was absent. Similar to the associated PI kinase activities, more PtdOH was always produced in assays of PITPp

131 PITPa PITPa (T58D) PITPp PITPp (T58D)

PtdIns4P

PtdIns(4,5)P2

Figure 4.9 Co-purification of Ptdlns kinase and PtdlnsP kinase with Wildtype and Mutant PITPs 6 pg of each PITP was assayed in triplicate for PI kinase activity using 200 pM exogenous Ptdlns and 70 pM [y-^^P]ATP (5 pCi per reaction) for 20 min at 37°C (Section 2.4.2). Lipids were then extracted and separated by TLC. The figure is representative U) of three separate experiments. The positions of lipid standards are indicated. to preparations than PITPa preparations (Fig. 4.4). This was probably due to the smaller amount of Ptdlnsfz formation in the latter (Fig. 4.4) because Ptdlnsfi may be the precursor for PtdOH formation (see Section 4.6).

To test this inference, the presence of DAGK activity was assayed using exogenous substrate. Initial assays using exogenous [^"^C]-DAG as substrate (Section 2.4.4) were not sufficiently sensitive to detect any kinase activity (figure not shown). This could have been due to low DAGK activity in the preparations and the low specific activity of the [“*C]-DAG (53 mCi/mmol).

Subsequently, an alternative, more sensitive method (Section 2.4.4) using unlabelled DAG and [y-^^P]ATP (specific activity: >5000 Ci/mmol) clearly revealed the presence of DAGK activity not only in wildtype PITP preparations but also in the mutant preparations (Fig. 4.10), albeit substantially lower in mutant PITPa preparations. Thus, the association of DAGK with PITPs appeared to parallel that of PI kinases and was independent of the type of lipid bound to PITP.

The origin of lyso-PtdOH is uncertain. There are four known synthesis pathways for lyso- PtdOH, but only two involve ATP (Gaits et a l, 1997). One is the déacylation of PtdOH by phospholipase Ai or A] - this pathway is unlikely in this case. The other is the phosphorylation of monoacyl glycerol (MG) by MG kinase. More work is required to clarify this issue.

In summary, the above results demonstrated unambiguously that PtdIns4K, PtdIns4f5K and DAGK activity had co-purified with both PITP isoforms, and that substrate bound to PITPs was used by these enzymes.

4.6 Detection of Phospholipase Activity

As stated above, Ptdlnsf] could be detected in the absence of PtdOH, but never \ice versa. This would suggest that PtdOH was produced in a reaction that required the generation of Ptdlns? 2 . This notion is supported by the observation that PtdOH was never produced by mutant PITP preparations, which did not generate Ptdlns ? 2 (Section 4.4.1). Furthermore, a simple explanation for the observed lipid phosphorylations was feasible: hydrolysis of Ptdlns ? 2 by PLC generated DAG, which was subsequently phosphorylated to PtdOH by DAGK. These reactions could also account for the apparent lack of increase in ^^P-labelling of PtdOH after further incubation with excess unlabelled ATP (Fig. 4.4):

133 PITPp (T58D) PITPp PITPa (T58D) PITPa

## PtdOH

Ptdlns4f

Figure 4.10 DAG Kinase Assays using unlabelled DAG and [y-^^P]ATP 6 jig of each PITP were assayed for DAGK activity in the presence or absence of exogenous DAG (Section 2.4.4). Lipids were then extracted and separated by TLC. The figure is representative of three independent experiments. The positions of lipid standards are indicated. the intensity of the PtdOH spot on the autoradiograph remained unchanged despite a large increase in the labelling of PtdlnsPi because Ptdlnsfi hydrolysis by PLC would have stripped all radioactivity from the lipid moiety in the form of [^^PJInsPs, and the subsequent phosphorylation of DAG to PtdOH would have primarily used unlabelled ATP. Consequently, it became important to test for PLC activity in the PITP preparations.

If PLC activity were present in the PITP preparations, hydrolysis of [^^P]PtdIns ? 2 that was produced from [y-^^P]ATP and endogenous substrate would generate [^^Pjlnsf] and unlabelled DAG. Accordingly, the isolation and characterization of inositol phosphates produced by the incubation of PITP preparations with [y-^^P]ATP would provide evidence for the postulated co-purification of PLC activity. A widely used method to extract inositol phosphates involves anion-exchange chromatography in which the various inositol phosphates are eluted separately from Dowex 1-X8 resin using different ammonium formate concentrations (Section 2.1.3.1). However, despite extensive washing of the resin prior to elution, a high background radioactivity due to the presence of [y- ^^P] ATP precluded any reliable interpretation of the data (data not shown).

An alternative approach was to use exogenous ^^P-labelled P tdlnsf] as substrate. This was synthesized in vitro using a preparation of recombinant STM7 (a type I PtdlnsPK; kindly provided by Dr. M. dos Santos), PtdIns4P and [y-^^P]ATP. More than 95% of the

^^P-labelled lipids produced by STM7 migrated to the same final position as the PtdlnsP] standard (M. dos Santos, personal communications). The lipid products were extracted and dried under nitrogen before resuspension and sonication in PLC assay buffer (Section 2.4.3.1). These lipids were incubated with each PITP preparation and the reaction products were then analysed using Dowex 1-X8 resin. Similar, modest increases in radiolabelled inositol phosphates were detected in assays of PITP preparations compared with control samples (Fig. 4.11 A). This indicates that PLC activity was indeed present in the PITP preparations.

Regarding the low sensitivity of this assay, it should be noted that, in addition to ^^P- labelled Ptdlns? 2, the substrate would also have contained a comparatively larger amount of unlabelled Ptdlns? (the substrate for STM7), as this was not removed prior to the PLC assay. Since PLC also cleaves Ptdlns? in vitro, it is likely that the unlabelled Ptdlns? reduced the sensitivity of the PLC assay by competing with Ptdlns ? 2 as a substrate for PLC.

135 A 3000

I 2500 -

i - 2000 t

a 1500 -

1000 A3 3 antigen PITPa PITPp

B 250 I 200 È 1Pk 2 (U V I 150

ti 100 If 50

0 PITPa PITPp

Figure 4.11 Presence of PLC Activity in PITP Preparations (A) Lipids extracted from a PtdIns4P5K (STM7) reaction were dried under nitrogen and resuspended in PLC assay buffer. Equal aliqouts (-16,000 cpm) of this preparation were incubated with equal amounts of PITPa, PITPp or A3 3 antigen (control) for 10 min at 30°C. Inositol phosphates were then extracted and counted (mean ± SD; n > 2). The data is representative of two independent experiments. See Section 2.4.3.1 for details. (B) Wildtype and T58D PITPs were purified from pH]inositol-labelled Sf9 cells and then incubated with unlabelled ATP for 10 min at 37°C (Section 2.4.3.2). Inositol phosphates were then extracted and counted (mean ± SD; « = 3). Count readings of mutant PITPs were set at 0%. The data represents the mean of two separate experiments. 136 To further investigate whether the co-purified PLC activity was able to cleave Ptdlnsf 2 that was generated from endogenous Ptdlns, the inositol ring was metabolically labelled. Sf9 cells were incubated with [^H]-inositol for six days, including the two days required for recombinant PITP expression following infection by baculovirus (Section 2.3.2.1), to radiolabel PITP-bound Ptdlns. After purification, wildtype and mutant PITPs were incubated with unlabelled ATP for 10 min to trigger the synthesis and turnover of PPIs. Inositol phosphates were then extracted, separated by anion-exchange chromatography and counted.

The results (Fig. 4.1 IB) clearly show a higher production of inositol phosphates in the wildtype PITPs relative to their mutant counterparts. This confirms the presence of PLC activity and the availability of endogenous substrate in preparations of wildtype PITPs.

4.7 Sf9 Cells as a Model of PITP Function

The co-purification of PI kinases, DAGK and probably PLC with PITP, as described above, followed overexpression of PITPa and PITPp in insect cells. Despite their popularity as hosts for the baculovirus expression system, Sf9 cells have not been extensively studied in their own right, that is, beyond their respective strengths in protein expression. In particular, the scarcity of reports on PI signalling in Sf9 cells raised the question of whether the observed co-purifications represent both genuine physiological associations and the role of PITP in mammalian cells.

This question was addressed, at least in part, by the detection of PITP in Sf9 cell extracts. Western blotting using polyclonal anti-PITPa antibodies recognised an endogenous 36- kDa protein (J. Hsuan, unpublished data), which may be a homologue of PITP or rdgBp in S. frugiperda (Fullwood et a l, 1999). In addition, a PITP homologue has been found in the Drosophila genome. Thus, the results and conclusions described in previous sections may be common to insect and mammalian systems.

4.8 Expression of Recombinant PITPs in Cultured Mammalian Cells

To further address the question of the validity of the Sf9 cell results as a bona fide model of PITP function, the expression of human PITPs in cultured mammalian cells was studied. Both PITPa and PITPp isoforms were overexpressed as wildtype and mutant (T58D) forms and then purified for biochemical characterization.

137 In these experiments, the His6 tag at the C-terminal end of PITPs was replaced by the 8- amino-acid FLAG tag for the following reasons. First, reagents to detect His6-tagged proteins in confocal microscopy were not available when these experiments were performed. Second, many mammalian intracellular proteins bind non-specifically to the Ni^'^-NTA resin (data not shown), which could have made interpretation of co­ purification experiment results difficult.

4.8.1 Generation of New cDNAs and Preparations of Pure Plasmids

Firstly, cDNAs encoding human PITPp (hPITPP) was amplified from a human Jurkat cDNA library by PGR (Section 2.2.1.3) and spliced into the pGEM vector.

The human PITPa (hPITPa) cDNA was synthesized by PGR using the hPITPa cDNA made in earlier experiments (see Chapter 3). In addition, the cDNA encoding the A33 glycoprotein was amplified by PGR from a full-length cDNA clone (a gift from Dr. Ed Nice, Ludwig Institute for Cancer Research, Melbourne). The hPITPa and A33 cDNAs were then subcloned into the pGEM vector.

All ORFs were then sequenced and those devoid of mutations were selected for subsequent manipulations. The hPITPa and hPITPp plasmids were then used as templates in overlapping PGRs designed to generate cDNAs encoding their mutant (T58D) counterparts (Section 2.1.1.1). The mutant cDNAs were also subcloned into pGEM and then sequenced to ensure that only mutations at the desired residues were present.

The five different cDNAs, encoding A3 3 and both a and p wildtype and mutant isoforms, were spliced into pBS-FLAG, immediately 5' of the ORF encoding the FLAG tag. Complete ORFs of PITP or A3 3 were thereby separated by two codons from the FLAG tag.

These PITP-FLAG and A3 3-FLAG ORFs were finally subcloned into the pcDNA3 mammalian expression vector and their nucleotide sequences were again checked for undesired mutations. Selected plasmids were then prepared in highly purified form (QIAGEN EndoFree) for mammalian cell transfection.

4.8.2 Mammalian Cell Transfection

4.8.2.1 COS-7 and A431 Cells

138 Two mammalian cell lines, A431 and COS-7, were initially chosen as expression hosts for the following reasons. It was from A431 cells that immuno-complexes of PITPa with PtdIns4K, PLC and EGFR were first purified (Kauffmann-Zeh et a/., 1995). Hence, it was hypothesised that a clear difference in PPI production would exist between wildtype and mutant PITP preparations from transfected A431 cells. If true, many of the implications of the observations and conclusions reached using insect cells would be extended to mammalian cells. COS-7 is a fibroblast-like simian cell line which expresses the SV40 large T antigen and thus allows episomal replication of the pcDNA3 vector, which contains the SV40 ori (origin of replication), and consequently, gross overexpression of any encoded protein.

All five plasmids (A33, wildtype and mutant PITPa, wildtype and mutant PITPp) were separately transfected into 40% confluent dishes of A431 and COS-7 cells using SuperFect (SF) (see Section 2.3.3.1). Cell lysates were prepared two days later for immunoprécipitation and Western blotting with the M2 anti-FLAG monoclonal antibody.

In A431 cells, expression of recombinant PITPs and A33 were hardly detectable by Western blotting (data not shown). In COS-7 cells, proteins of approximately 35 kDa and 50 kDa were immunoprecipitated from cells transfected with pcDNA3 vectors containing wildtype PITPa and A33 cDNAs, respectively (Fig. 4.12). This is consistent with the expected molecular masses of recombinant PITPa-FLAG and A33-FLAG proteins. However, there was no detectable expression in the other three transfection experiments (mutant PITPa, and wildtype and mutant PITPP). As PITPp has previously been found to mainly localize to the Golgi apparatus (de Vries et a l, 1995; de Vries et a l, 1996), the insoluble fraction in each cell lysate was also probed with anti-FLAG antibody. There was no apparent difference in immuno-staining compared with control cells transfected with pcDNA3 vector alone (Fig. 4.12), although recombinant PITPs may have been obscured by endogenous proteins of similar molecular weight.

The above lack of expression of PITPp and mutant PITPa was unlikely to be due to mutations in the cDNA reading frames, resulting in the synthesis of out-of-frame proteins or premature termination of protein synthesis, because when these plasmids were transfected into another cell line (see Section 4.8.4), proteins were expressed and detected using immunofluorescence microscopy.

4.8.2.2 Human Embryonic Kidney Cells

139 control (BAP) PITPa mutant PITPa

1 2 1 2 1 2 P FT E P FT E P FT E P FT E P FT E P FT E

BAP

PITPa f f

PITPP mutant PITPP A33

1 1 1 P FT E P FT E P FT E P FT E P FT E P FT E

A33

Fig. 4.12 Expression of PITPs and A33 in COS-7 ceils Two dishes (1,2) of COS-7 cells were transfected with the pcDNA3 vector (control) or pcDNA3 encoding A33 or each PITP as described in Section 2.3.3.1. After two days, the cells were lysed, cleared and immunoprecipitated with anti-FLAG M2 affinity gel (Section 2.3.3.2). For control cells, FLAG-Bacterial Alkaline Phosphatase (BAP, a 49.1-kDa protein) was added prior to immunoprécipitation. Immunoprecipitates were analysed by SDS-PAGE and Western blotting with anti-FLAG M2 antibody. P, pellet; FT, flowthrough (cell lysate that did not bind to the affinity gel); E, eluate fi*om affinity gel. 140 Because of the poor expression of most PITPs in both A431 and COS-7 ceils, other cell lines were sought which might allow a higher expression level. The human embryonic kidney (HEK) 293 cell line is a stable cell line transformed by sheared human adenovirus type 5 DNA and has been widely used for the expression of recombinant proteins.

HEK cells were transfected with all 5 plasmids using SF and harvested after two days as before. A visible increase in the expression of all 5 proteins was detected by Western blotting (Fig. 4.13A). Wildtype PITPa and A33 antigen were again the most highly expressed, and the expression level of mutant PITPa was lower than the wildtype. Both wildtype and mutant PlTPp were still poorly expressed; although the expression levels were higher than those in COS-7 cells, they were still too low and variable to allow the desired comparative biochemical studies to be performed. Varying the quantity of plasmid DNA and the DNA:SF ratio had little effect on PITPa (Fig. 4.13B) and PITPp (data not shown) expression in HEK cells.

Since the expression of PITPp could not be improved, it was not feasible to study its association with PI kinases. Thus, all subsequent transfection studies were restricted to PITPa and A3 3. In addition, since varying the amount of transfection reagents did not bring about any significant difference in the quantity of immunoprecipitated PITPa, the minimal amount of DNA (1 pg) and lowest DNA:SF ratio (1:2) were used, in order to conserve reagents and to minimise any cellular toxicity.

4.8.3 Immunoprécipitation and PI Kinase Assays

15-cm dishes of HEK cells at 40% confluence were transfected with 15 pg of plasmid

DNA and 30 pi of SF (DNA:SF ratio = 1:2). A day later, the Dulbecco's Modified Eagle Medium (DMEM), supplemented with 10% FCS, was replaced for 24 h with DMEM supplemented with 10 mg/1 insulin, 6.7 pg/1 sodium selenite and 5.5 mg/1 transferrin (but no FCS). The serum-starved cells were then treated with 10% FCS for 5 min, lysed and immunoprecipitated with anti-FLAG M2 affinity gel (Section 2.3.3.2). As the proteins were expressed in different quantities, the amount of affinity gel used was adjusted according to the amount of protein that was expected to be expressed.

An equal volume of each immunoprecipitate was incubated with [y-^^P]ATP and the lipids subsequently extracted, separated by TLC and their radioactivity quantified (Fig. 4.14). Wildtype PITPa was found to produce more than four times the amount of PtdlnsP

141 A PITPa PITPp W M W M A33 kDa

A33

— 46

PITP

B

plasmid DNA (jig) DNA:SF ratio

PITPa — ►

Fig. 4.13 Expression of PITPs and A33 in HEK cells (A) HEK cells were transfected and proteins harvested as described in Fig. 4.12, except that no BAP was added to control cells. The figure is representative of two separate transfection experiments. (B) 6-well dishes of HEK cells were transfected with pcDNA3 vector encoding PITPa. For each set of three transfections, the amount of plasmid DNA was kept constant while the DNAiSF ratio varied. C, control (cells); W, wildtype; M, mutant.

142 50000

^ 40000 — I I 30000 — I

1 20000 —

10000 —

wildtype T58D A33 PITPa PITPa

Fig. 4.14 Association of PI Kinase Activities with Mammalian- Expressed PITPa 15-cm dishes of HEK cells (40% confluency) were transfected as described in Section 2.3,3.1, except that the cells were serum-starved for 24 h and then treated with 10% FCS for 5 min before lysis. The expressed proteins (wildtype and T58D PITPa, and A33) were immunoprecipitated as described (Section 2.3.3.2), after which 4 pi of the affinity gel binding the respective protein was incubated with [y-^^P]ATP for 20 min in duplicates or triplicates. Lipids were extracted, separated by TLC and quantified using a phosphoimager (mean ± SD; n > 2). The PtdlnsP values for PITPs were averaged and compared with that of A33 (control). The data is representative of three independent experiments.

143 compared to T58D PITPa. This preliminary result hinted that human PI kinase(s) also associated and co-purified with PITP and that wildtype PITPa was providing substrate for the co-purified PI kinase(s).

It is noted that lipids extracted from A3 3 and mutant PITPa after ATP incubation also contained comparable amounts of radioactive lipids that co-migrated with the PtdlnsP standard in TLC. This indicates a higher background of non-specific binding of Ptdlns and/or PI kinase(s) from HEK cells to the agarose-based anti-FLAG M2 affinity gel, in comparison with the Sepharose-based Ni^^-affmity purification from Sf9 cells. This however should not dilute the above indication of substrate presentation by PITP that was expressed in mammalian cells, given the significant difference in Ptdlnsf production.

4.8.4 Immunofluorescence Microscopy

Overexpressed proteins may be misplaced compared with their normal subcellular locations and form associations or complexes with other cellular components. To determine whether overexpression had altered the localization of PITPs in the above experiments, immunofluorescence studies were performed on the transfected HEK cells (Section 2.22.2). Two days following transfection, cells were incubated with or without SLO for 30 min such that cytosolic contents leaked out. This was performed to reduce the amount of cytosolic label and thereby allow any membrane-bound PITP to be more clearly observed. Cells were subsequently fixed with paraformaldehyde, stained with anti- FLAG M2 monoclonal antibody, TRlTC-phalloidin and, where indicated, 4',6-diamidino- 2-phenylindole (DAPl), and viewed using the Zeiss 510 scanning confocal microscope (operated by Drs. J. Hsuan and M. dos Santos).

4.8.4.1 Wildtype PITPa

Wildtype PITPa was found in the nucleus and cytoplasm (Fig. 4.15A), consistent with previous reports (de Vries et al., 1995; de Vries et al., 1996). However, PITPa appeared to be excluded from the nucleolus, suggesting that PI metabolism is not involved in ribosomal synthesis and processing.

Two papers reporting the intracellular distribution of PITPa (de Vries et al., 1995; de Vries et al., 1996) did not observe any association of the protein with the plasma membrane. Fig. 4.15A, however, shows a prominent edge pattern on several transfected cells, which may indicate the localization of recombinant PlTPa-FLAG at the plasma

144 AB

D

Fig. 4.15 Expression of PITPa in HEK Cells Wildtype (A, B, D) and mutant (C, E) hPITPa were expressed by transfection of HEK cells. After 48 h, the cells were incubated with (D, E) or without (A-C) SLO prior to fixing and staining (Section 222.2). The cells were stained for PITPa using FITC-conjugated anti-mouse IgG (green) and for F-actin using Phalloidin-TRITC (red). Each bar represents 10 pm. LA membrane. Additional experiments (e.g., co-fluorescent staining with a plasma membrane-bound protein at a higher magnification) are necessary to verify the above observation. If true, it lends support to the current view that PITPa interacts with phospholipase-mediated PI signalling enzymes at the plasma membrane.

Permeabilisation by SLO caused rounding of the cells and a noticeably dimmer fluorescence labeling of PITPa (Fig. 4.15C). This is consistent with the reported leakage of endogenous PITPa from SLO-permeabilised cells (Cunningham et al., 1995; Kauffmann-Zeh et a l, 1995). Fluorescent staining can again be observed at the plasma membrane (contrast with T58D PITPa in Fig. 4.15D).

4.8.4.2 Mutant PITPa

T58D PITPa was expressed in a visibly fewer number of HEK cells and its intracellular distribution appeared to be restricted to the cytoplasm (Fig. 4.15B). Association of T58D PITPa with the plasma membrane was not detected. This may be due to weaker interactions of PtdCho-bound mutant PITPa with signalling complexes at the plasma membrane or as a result of the generally lower level of protein expression compared to wildtype PITPa. As in the case of wildtype PITPa, partial retention of T58D PITPa was observed in permeabilised cells, but the protein remained excluded from the nucleus. Rounding of the cells was also observed.

4.8.4.3 PITPp

Compared to PITPa, fewer PITPp-overexpressing HEK cells were apparent (Fig. 4.16A), which is consistent with the low yield of recombinant PITPp that was found in transfected cell lysates (Fig. 4.13). One possible reason for this is that PITPp overexpression may be toxic to cells. However, stable NIH3T3 cell lines with up to 16-fold higher PITPp levels have been isolated recently (van Tiel et al., 2000a). Both wildtype and mutant PITPp were confined primarily to the cytoplasm. In the case of the former, staining appeared to be punctate, in the shape of a crescent around the nucleus (Fig. 4.16A). This is consistent with observations that PITPp is localized at the Golgi apparatus (de Vries et al., 1995; de

Vries et a l, 1996). On the other hand, mutant PITPp appeared to be more evenly distributed (Fig. 4.16B).

146 B

Fig. 4.16 Expression of PITPp and A33 in HEK Cells Wildtype (A) and mutant (B) PITPp and A33 (C) were expressed by transfection of HEK cells. After 48 h, the cells were fixed and stained as described in Fig. 4.15. For mutant PITPp and A33, cell nuclei were also stained using DAPl (blue). Each bar represents 10 pm.

147 SLO-treatment of HEK cells overexpressing PITPp caused detachment of the cells from the glass coverslips during fixation. This made quality confocal micrograph difficult to obtain.

4.S.4.4 A33

The A3 3 protein was observed to localize to the plasma membrane when overexpressed in HEK cells (Fig. 4.16C). This is in agreement with reports that the A33 antigen is a transmembrane molecule at the surface of intestinal epithelial cells (Heath et al., 1997; Johnstone et al., 2000).

4.9 Discussion

The central tenet of the ‘co-factor model’ for PITP function in PI signalling is that PITP- bound Ptdlns is the preferred substrate for PI kinases and PLC. This requires the recognition of the Ptdlns lipid by PI kinases and involves the association of PITP and PI kinases. The data presented above indicate the occurrence of such associations under different experimental conditions. PITP-Ptdlns was shown to interact with partially purified, renatured type II PtdIns4K and trigger Ptdlns phosphorylation. PITPa was also able to associate with PtdIns4K activity from purified A431 plasma membranes.

Both PITPa and PITPp also interacted with endogenous type II PtdIns4K, PtdlnsPK, PLC and DAGK in Sf9 cells, resulting in co-purification. Other studies have shown or suggested that PITP or Secl4p interacted individually with the EGFR, PtdIns4K, PI3K or PLC (Aikawa et al., 1999; Jones et al., 1998; Kauffmann-Zeh et al., 1995; Panaretou et al., 1997). The data presented in this chapter are, to our knowledge, the first report of PITP associating and co-purifying with all four of the abovementioned PI signalling enzymes in a single experiment. Such simultaneous co-purification gives strong support to a crucial aspect of the ‘co-factor model’, namely that PITP interacts with the multi­ enzyme PI signalling complex in vivo.

More significantly, it was also demonstrated that the abovementioned enzymes utilised substrates that originated from PITP-bound Ptdlns. The mere addition of ATP to PITP preparations triggered the spontaneous phosphorylation of Ptdlns by the co-purifying type II PtdIns4K, suggesting that the kinase is able to recognise and use PITP-bound Ptdlns as substrate. The [^^PJphosphate chase experiment indicates further phosphorylation by PtdlnsPK to produce PtdlnsP], which was subsequently hydrolysed by PLC to generate

148 InsPs and DAG, the latter being further phosphorylated by DAGK to produce PtdOH. The channelling of DAG from PLC to an associated DAGK was also previously observed by (van der Bend et al., 1994). Hence, it appears that PITP-bound Ptdlns was sequentially metabolised by the co-purifying lipid kinases and lipase. This is in agreement with the proposed model, based on the crystal structure of apo-PITPa, that Pl-modifying enzymes can have direct access to the inositol moiety of PITP-bound Ptdlns (Schouten et al.,

2002).

The origin of the Ptdlns substrate was verified by a specific mutation at T58, which abolished the ability of PITPa and PITPp to bind and transfer Ptdlns. Lipid extraction analyses indicate that the T58D mutation resulted in the apparent absence of PITP-Ptdlns in the PITP preparations. The absence of PITP-Ptdlns in turn abolished the activity of the co-purifying kinases, from which it was inferred that PITP-bound Ptdlns is a direct substrate for the abovementioned enzymes. These data are corroborated by independent reports showing that mutated Secl4p, which is incapable of binding and transfer Ptdlns, not only failed to reconstitute PITP-dependent pathways that involve PPI production but was also unable to stimulate PtdIns4P production in vivo (Phillips et al., 1999).

Surprisingly, the interaction of PITP with the PI signalling enzymes was not affected by the T58D mutation and therefore appeared to be independent of the phospholipid bound on PITP. This would suggest that the phospholipid headgroup and/or its recognition by the signalling complex are not required for the association of PITP with the complex.

No generation of PI 3-phosphates was detected for both PITPa and PITPp following incubation with ATP. This is in agreement with the weak PI3K activity found in the putative PI signalling complex that is associated with the cytoskeleton in A431 cells (Payrastre et al., 1991). Several papers had reported the synergistic cooperation of PITPa or Secl4p and PI3Ks in PI 3-phosphate production and the formation of constitutive transport vesicles from the TGN (Jones et al., 1998; Kular et al., 1997) as well as the co- immunoprecipitation of a Ptdlns 3-kinase activity using an anti-PITP antibody (Panaretou et al., 1997). However, none of the above papers provided direct and unambiguous evidence that the PI 3-phosphates generated were derived from PITP-bound Ptdlns. Indeed, the Ptdlns 3-kinase that co-immunoprecipitated with PITPa/PITPp did not appear to generate Ptdlns3f when the exogenous Ptdlns substrate was replaced with Ptdlnsf or

Ptdlns ? 2 (Panaretou et a l, 1997). This could be interpreted as an apparent inability of the Ptdlns 3-kinase to recognise and phosphorylate PITP-bound Ptdlns, assuming that there

149 was sufficient PITP-Ptdlns in the immunoprecipitate for Ptdlns 3-kinase to infer substrate specificity with respect to PITP.

The lack of any detectable PI 3-phosphate production may also be due to the generation of PtdOH and lyso-PtdOH during incubation with ATP. PtdOH and lyso-PtdOH have been shown to inhibit PI3Ka activity (Lauener et a l, 1995). Lyso-PtdOH is produced and released by a number of agonist-stimulated cells, such as platelets (Billah et al., 1981; Eichholtzet al., 1993) and fibroblasts (Fukami and Takenawa, 1992), and functions as an intercellular signalling molecule which targets cells by binding to a specific cell-surface receptor (Moolenaar, 1994). Among other things, lyso-PtdOH can trigger platelet aggregation, stimulate growth of fibroblasts, vascular smooth muscle cells and kératinocytes, and promote cellular tension and cell-surface fibronectin binding. The pathway by which lyso-PtdOH was synthesized in ATP-treated PITP preparations is presently unclear but it is unlikely to have been formed via phospholipase A 2-mediated déacylation of PtdOH since lyso-PtdOH was also generated in T58D PITPs which produced no detectable PtdOH on addition of ATP (Fig. 4.8).

PITPa has been shown to localize to the nucleus and cytoplasm while PITPp is found associated with Golgi-like membranes (de Vries et a l, 1995; de Vries et al., 1996; Segui et al., 2002). Wildtype PITPa and PITPp, when overexpressed in HEK cells, were also observed to localize to these respective subcellular locations (Figs. 4.15 & 4.16), suggesting that the PITPs were not misplaced from their intracellular distribution. This is consistent with other reports which observed that the concentration of PITPa or PITPp did not affect their distribution in the cell (de Vries et al., 1996; van Tiel et al., 2000a). Attempts to visualise the distribution of recombinant PITPs in Sf9 cells by immunofluorescence were however unsuccessful, because the cells were unable to attach firmly to the glass coverslips.

The role of PITP, particularly PITPa, in receptor-initiated PI signalling at the plasma membrane was called into question when several reports did not detect PITP at the plasma membrane (de Vries et a l, 1995; de Vries et al., 1996). The methods used included indirect immunofluorescence and fluorescent-labelling of PITP isoforms. However, Fig. 4.15 shows more intense staining of wildtype PITPa at the periphery of several HEK cells. The reasons why other laboratories had failed to make this observation could be manifold. Firstly, the proportion of total cellular PITPa associating with the plasma membrane may be small, and may only be detectable by overexpressing the 150 protein. Secondly, epitopes of PITPa may be blocked by interactions with the plasma membrane or membrane proteins and thus rendered inaccessible to antibodies. This would suggest that anti-PITP antibody could inhibit PITPa function. Indeed, an anti-PITPa polyclonal antibody inhibited, in a concentration-dependent manner, the budding of exocytic vesicles, of which PITP is an essential component (Jones et al., 1998). Likewise, fluorescent labels covalently attached to PITPa and PITPp may have caused steric hindrance to association with plasma membrane lipids and/or proteins (de Vries et al., 1996). Lastly, the different types of cell lines used in the experiments may display quantitative differences in the intracellular distribution of PITPs.

The role of PITP in PI signalling was again called into doubt when the overexpression of PITPa in stable mouse NIH3T3 fibroblast cell lines, denoted SPI cells, did not lead to detectable changes in intracellular levels of PtdlnsP and Ptdlnsfi- In SPI cells, the relative incorporation of inositol in PtdlnsP, Ptdlnsfz, inositol-4-phosphate, inositol-1,4-bisphosphate and Insf] were (i) similar to that in wildtype NIH3T3 cells and (ii) essentially unchanged in the presence of such agonists as bombesin or PDGF (Snoek et a l, 1999). Hence, it appears that the increased amount of PITPa not only did not trigger increased levels of PPIs and ‘downstream’ inositol phosphates in unstimulated SPI cells, but also led to an apparent loss of growth factor-stimulated Ptdlnsfz degradation. On the other hand, the relative incorporation of inositol in lyso-Ptdlns and inositol derivatives including glycerophosphoinositol, inositol-1-phosphate and inositol- 2-phosphate were increased significantly in SPI cells, compared to wildtype cells, but fell to a lower level on stimulation. A Ptdlns-specific phospholipase A (PLA) was proposed to have been constitutively activated to generate lyso-Ptdlns and this may help explain the enhanced growth rate of SPI cells (Snoek et a l, 1999, and references therein). Thus, the authors suggested that PITPa overexpression had no effect on PtdInsP 2-specific PLC activity in NIH3T3 cells and that PLA-mediated degradation of Ptdlns was activated instead.

While it is clear that PITPa may play a role in regulating PLA activity, the lack of increased inositol incorporation in Ptdlnsf and Ptdlns ? 2 and the apparent desensitization of PLC-mediated PI signalling in growth factor-stimulated SPI cells could be due to several factors. Firstly, the generation of PPIs may require prior association of PITPa with PtdIns4K, which in A431 cells is dependent on growth factor stimulation

(Kauffmann-Zeh et a l, 1995). Accordingly, the mere increase in intracellular PITPa level

151 in unstimulated cells would not be expected to lead to corresponding increases in PtdlnsP and PtdInsP 2 levels without concurrent activation of the respective PI kinases by an agonist. Secondly, PITPa may only regulate the rate of PPI turnover rather than control the rate of de novo PPI biosynthesis or alter the level of each PPI. This would be analogous to observations in (i) PITPp-overexpressing NIH3T3 cells, where PITPp is involved in restoring, rather than changing, the steady-state level of SM (van Tiel et al., 2000a) and in (ii) COS-7 cells, where overexpression of CDP-diacylglycerol synthase and/or Ptdlns synthase did not significantly affect the rate of Ptdlns biosynthesis nor lead to higher cellular levels of Ptdlns and PPIs (Lykidis et al., 1997). Likewise, no differences in the incorporation of the [^^P]phosphate precursor into Ptdlns, Ptdlns? and

Ptdlns ? 2 in the absence or presence of vasopressin were found between wildtype WRK-1 tumour cells and cells with a lowered level of PITPa (Monaco et al., 1998). These data further support the current consensus that the metabolism of inositol phospholipids is tightly regulated, both spatially and temporally. Thirdly, the aminoglycosidic antibiotic geneticin G418 has been shown to affect Ptdlns ? 2 turnover in vivo (Gabev et al., 1989), apparently by inhibiting PLC activity or PLC/PtdIns? 2-mediated cellular processes (Hu, 1998; Jones and Wessling-Resnick, 1998; Takahashi et a l, 2001; Walz et a l, 2000). The G418 antibiotic that was used to select SPI cells might thus have interfered with the PLC- mediated signalling pathway. However, externally-applied G418 is not equally effective in inhibiting PLC activity in all cell types (for example, see Hildebrandt et a l, 1997). Therefore, the data reported by Snoek et al. (1999) need to be interpreted with caution.

In contrast, stable NIH3T3 cell lines that overexpressed PITPp, denoted SPip cells, acquired a decreased growth rate (van Tiel et a l, 2000a). It was shown that PITPp induced a more rapid resynthesis of SM in SPip cells that were treated with an exogenous sphingomyelinase, such that the steady-state level of SM was able to be restored more quickly. The effect of an increased intracellular level of PITPp on PI metabolism was not reported, perhaps because the antibiotic G418 was also used in selecting stable SPip cell lines. These findings remain controversial because Cockcroft and co-workers reported no similar effects in COS-7 cells overexpressing PITPp (Segui et a l, 2002).

Type II PtdIns4K isozymes are stimulated by TXlOO, but addition of TXlOO to the PITP preparations did not further stimulate the co-purified type II PtdIns4K activity (Fig. 4.3). This could be due to differences in substrate presentation in these assays. Although the mechanism of PtdIns4K stimulation by TXlOO has not been described, it could be caused

152 by changing the physical properties of lipid micelles and/or by an allosteric effect on the PtdIns4K. For example, the enzymatic activity of recombinant PtdIns4K-IIIp can be enhanced by a decrease in diameter of Ptdlns vesicles (Hubner et a l, 1998). The co­ purified Sf9 cell PtdIns4K may be directly presented with its substrate by PITP, which would in accordance with the co-factor model be far more efficient than presentation with free micellar substrate. The addition of TXlOO to the PITP preparations would then not be expected to affect substrate binding and thus kinase activity. In this regard, the addition of exogenous micellar Ptdlns to the PITP preparations had no significant effect on the total lipid kinase activity (Fig. 4.6). This supports the notion that the preferred substrate of the PITP-associated PtdIns4K is PITP-bound Ptdlns.

It was noted with interest that PITPp always co-purified with relatively higher lipid kinase activities than PITPa. The reasons for this phenomenon are unclear but it may indicate a greater degree of molecular association of PITPp with activated PI signalling complexes on a per molar basis. This could be related to the observation that, compared to PITPa, PITPp is primarily localized at membranes, particularly the Golgi apparatus in cells (Fig. 4.16; de Vries et a/., 1995; de Vries et a l, 1996; Segui et a l, 2002), possibly via interactions with membrane-associated lipid kinases. An alternative possibility is that both PITPa and PITPp isoforms co-purified with similar amounts of PI kinases (Fig. 4.9), but PITPp presented the Ptdlns substrate more efficiently, thereby increasing the rate of

PPI generation. PITPp is apparently capable of adopting at least three distinct organizational forms, each with a different apparent molecular mass (see Chapters 3 and 5). Although one of these forms may be more efficient at presenting substrate, the efficiency of PITPa and the three forms of PITPp in in vitro Ptdlns transfer (Fig. 3.5), Ptdlns phosphorylation (Fig. 3.7) and PLC activity reconstitution (Fig. 3.8) were similar.

T58D PITPs were always expressed in smaller quantities than their wildtype counterparts in both Sf9 and mammalian cells. This appears to be due to a combination of fewer cells expressing T58D PITPs as well as comparatively lower expression of T58D PITPs in transfected cells (see Figs. 4.15 and 4.16). This may indicate that a high cellular level of either T58D PITP is toxic. PITPa occurs intracellularly as two charge isomers, each with a different isoelectric point and binding either a Ptdlns or PtdCho molecule (Wirtz, 1991, and references therein, see also Chapter 5). Charge isomers of PITPp had also been reported (van Tiel et a i, 2002, see also Chapter 5). It is possible that the ratio of PITP charge isomers within a cell needs to be delicately balanced and controlled because a

153 disruption of this ratio (for example, in favour of PITP-PtdCho through the expression of T58D PITP) would be expected to perturb critical metabolic pathways such as PtdCho biosynthesis (Monaco et aL, 1998) and induce cell toxicity (Xie et a l, 2001).

The threonine residue at position 58 (for example in human PITPa) or 59 (for example in rat PITPa) is highly conserved among various isoforms of PITPa, PITPp and RdgB, and its mutation in PITPa has been shown to affect the ability to transfer Ptdlns but not PtdCho (Alb et aL, 1995). (This conserved Thr residue will hereafter be denoted as T58 in general and as T59 where appropriate.) The importance of this residue is corroborated by its positional proximity to, and its potential to form hydrogen bonds with, the inositol moiety of Ptdlns in the tertiary structure of PITPa (Schouten et aL, 2002; Yoder et aL,

2001). However, while the Ptdlns transfer activities of both PITPa and the PITP domain of RdgB (RdgB-PITP) are sensitive to substitutions at T58, the effects of different amino acid substitutions for T58 that affect Ptdlns transfer clearly differ between RdgB-PITP and PITPa (Milligan et aL, 1997). The T59D or T59E mutation virtually abolished Ptdlns transfer activity of rat PITPa while the T59A mutation decreased the specific Ptdlns transfer activity by half (Alb et aL, 1995). Data presented here indicated that the T58D mutation also has the same effect on human PITPa and rat PITPp. However, in the case of the RdgB-PITP domain, others have shown that Ptdlns transfer activity was inactivated by a T59A mutation, while the T59E mutation had no detectable effect (Milligan et aL, 1997). Taken together, these results suggest that PITPa/PITPp and the RdgB-PITP domain employ subtly different mechanisms for phospholipid transfer, which may reflect differences in physiological function, despite >40% identity in their primary sequence and having similar in vitro phospholipid transfer activity. In this regard, a full-length, chimeric RdgB containing rat PITPa in place of the RdgB-PITP domain, not only failed to restore the wildtype light response to rdgB^ null mutants but also caused additional adverse effects on photoreceptors (Milligan et aL, 1997).

The T58D mutation also led to the exclusion of PITPa from the cell nucleus, hinting that

PITPa-PtdCho may be restricted to the cytoplasm. PITPa-Ptdlns, as the main (if not only) PITPa charge isomer in the nucleus, may be involved in nuclear PI signalling as previously proposed (Irvine, 2000). It is unclear how PITPa-PtdCho is preferentially localized to the cytoplasm. One possibility is that the nuclear pore complex directly or indirectly recognises and transports only PITPa-Ptdlns into the nucleus and/or expels

PITPa-PtdCho from the nucleus. It is tempting to speculate that the latter process may be 154 regulated by a nuclear PKC that preferentially phosphorylates PITPa-PtdCho (van Tiel et aL, 2000b), which is subsequently exported. This would be analogous to the regulated nuclear export of the protein phosphatase Cdc25 by the sequential action of the serine/threonine kinase Chkl and the 14-3-3 protein Rad24 (Lopez-Girona et aL, 1999). It will be interesting to see whether PITPa-PtdCho also requires a ‘transport factor’, such as a 14-3-3 protein, for its nucleocytoplasmic shuttling. Another possibility is that the T58D mutation resulted in PITPa-PtdCho adopting a different structural or organizational state, such that the mutant form is differentially localized to the cytoplasm.

155 Chapter 5

Chromatographic and Electrophoretic Analyses of PITP and Co- purifying Enzymes

5.1 Introduction

The ‘co-factor model’ of PITP function in PI signalling postulates that PITP presents Ptdlns as a preferred substrate to lipid kinases. PITP may act as a co-factor via at least two alternative mechanisms, as depicted in Figure 1.7. The two mechanisms differ primarily in the role of PITP in the presentation of Ptdlns to PtdIns4K and whether the phosphorylated Ptdlns lipid remains bound to the transfer protein. The data that were discussed in Chapter 4 address, at least in part, the viability of PITP-bound Ptdlns as a substrate. However, it remains unclear whether PITP continues to play a role in the subsequent phosphorylation and hydrolysis of the PPI by PtdlnsfK and PLC, respectively. This chapter will present data that points to the formation of ternary complexes of PITP, Pis and lipid kinases, before and during PI phosphorylation.

5.2 Production of PPIs on PITP

In addition to the presentation of the Ptdlns substrate by PITP to PI kinases, it may be possible that the resultant PPIs (PtdlnsP and PtdlnsP]) remain bound to PITP through the phosphorylation process. This would entail that PITP also presents substrates to PtdlnsPK and PLC for enzymatic action.

To investigate whether PI lipids remain bound to PITPa and PITPp following phosphorylation by PI kinases, Ptdlnsf and Ptdlns ? 2 were first generated by adding [y- ^^P]ATP to the purified PITP preparations (as described in Section 4.3). Instead of extracting the PPIs, the entire reaction was analysed and resolved by gel filtration chromatography and lEF.

5.2.1 Resolution by Gel Filtration Chromatography

Gel filtration chromatography can be used as a tool to investigate whether multiple biomolecules associate and thus co-elute as a multi-component complex. The reaction mixture was resolved by gel filtration using a Superdex 75 column and the eluted fractions were analysed for the presence of PPIs. Surprisingly, no labelled PPIs were detected in any of the eluted fractions (Fig. 5.1). Running further bed-volumes of solvent

156 80

70

60

I 50 I 40 30 h-1i 20

10

10 15 20 24 Fraction

Fig. 5.1 Gel Filtration Analysis of ATP-Treated PITP After incubation with [y-^^P]ATP, PITPa and PITPp were separately gel-filtered through Superdex 75 on the SMART Liquid Chromatography System (Pharmacia). All fractions were subjected to lipid extraction (Section 2.4.2) and counted (PITPa, open squares; PITPp, solid diamonds). The data is representative of two separate gel filtration analyses.

157 through the column did not succeed in retrieving PPIs from the column. This indicates that the PPIs generated during incubation with [y-^^P]ATP had not eluted from the Superdex column.

5.2.2 Resolution by Iso-Electric Focusing

Similar to gel filtration chromatography, lEF can provide evidence for molecular interactions through co-migration of different molecules. Co-migration of the labelled

Ptdlnsf and Ptdlns ? 2 to the same iso-electric position as PITP would suggest the continued residence of the PPIs on the transfer protein, especially if no other protein could be detected at the same position.

5.2.2.1 Analysis of PITP - Before Incubation with ATP

The iso-electric points of purified recombinant PITPs were investigated first. PITPa-His6 and PITPp-His6 were purified following overexpression in Sf9 cells (Section 2.3.2) and analysed by lEF on gels with a pH 3-10 gradient (Section 2.1.2.1.3). No prior incubation with ATP was performed.

5.2.2.1.1 PITPa

When stained with Coomassie Blue, purified PITPa-His6 was observed to have separated into two main bands with pi values of 6.2 and 6.4, respectively, and one minor band with pi 6.1 (Fig. 5.2A). Western blotting with anti-PITPa polyclonal antibodies revealed no other significant protein bands (Fig. 5.2B). The minor band was not further characterized due to difficulties in isolation and purification (discussed in Section 5.4). The migration pattern was identical in separate lEF gel lanes where the PITPa samples were applied at different points between the anode and cathode (figure not shown).

Several differences between recombinant PITPa-His6 and PITPa purified from native sources were observed. Firstly, native PITPa typically exists as two charge isomers, which resolved on an lEF gel as two protein bands with slightly different pi values (Wirtz, 1991, and references therein). The different pi values were attributed to the presence of either Ptdlns or PtdCho in the lipid-binding site of the protein (van Paridon et aL, 1987b) as well as the one negative charge difference between Ptdlns and PtdCho. The PITPa-Ptdlns isomer therefore has a lower pi value (which ranges from pi 4.9 - 5.6) than the PITPa-PtdCho isomer (pi 5.3 - 5.9). Secondly, the pi values (6.2 and 6.4) of the two

158 A 7.1 —

6.8 —

6.5 —

6.0 —

5.1 - V

4.75 [ #

# /

B 7.8 —

7.5 \ _

7.1 —

6.5 — # 6.0 —

pi / / /

C 7.8 — 7.5 V

7.1 ~

6.5 — I 6.0 —

Fig. 5.2 Iso-Electric Focusing of Recombinant PITPs Recombinant PITPa and PITPp were purified (Section 2.3.2.2) and analysed by lEF (Section 2.1.2.1.3). The PITPs (0.3-0.5 pg) were detected by (A) Coomassie Blue staining or by (B) anti-PITPa or (C) anti-PITPp antibodies. The figures are representative of at least three independent lEF experiments. Std, lEF Standards (Board Range pi 4.45-9.6). 159 main PITPa-Hisô isomers differ significantly from those of native PITPa (Wirtz, 1991). This difference is presumably due to the six imidazole side chains (pA^a = 6.0) of the His6 tag at the C-terminus of recombinant PITPa.

Thirdly, approximately two-thirds of native PITPa bind Ptdlns while the remainder binds PtdCho (Thomas et aL, 1994; van Paridon et aL, 1987b). In contrast, there was a similar distribution of PITPa-His6 between the two main isomers at pi 6.2 and 6.4 (Fig. 5.2). As insufficient amounts of lipid could be extracted from each isomer in the lEF gel, lipids were extracted from the starting PITPa preparation. Analysis of the extracted lipids by TLC and iodine staining (Section 2.1.3.2.1) revealed similar amounts of Ptdlns and PtdCho (figure not shown). It is therefore inferred that the two charge isomers at pi 6.2 and 6.4 were PITPa-Ptdlns and PITPa-PtdCho, respectively.

5.2.2.1.2 PITPp In contrast to PITPa, recombinant PITPp appeared smeared on an lEF gel (Fig. 5.2A). Nevertheless, two main protein bands can still be detected at pi values of 6.5 and 6.7, respectively. When the lEF gels were Western-blotted with anti-PITPp polyclonal antibodies, another minor charge isomer at pi 6.3 was observed (Fig. 5.2C). The migration pattern was independent of the location of the sample origin between the electrodes (figure not shown).

As in the case for PITPa, TLC separation of the lipids extracted from the starting PITPp preparations identified similar amounts of Ptdlns and PtdCho (figure not shown). Thus, the two main PITPp charge isomers could be binding either Ptdlns or PtdCho, as described above for PITPa. This is consistent with the fractionation of native PITPp (purified from chicken liver) into two subforms of different iso-electric points (Westerman et aL, 1995). However, purified bovine brain PITPp was found to have only one charge isomer, with pi of 5.4 (de Vries et aL, 1995). In addition, the pi values of recombinant PITPp-His6 also differ from those of native PITPp. This is presumably due to the His6 tag, as in the case for PITPa (see previous Section). It is further noted that the more basic pi values for PITPp isomers are consistent with the relatively lower proportion of strongly acidic amino acid residues in PITPp (Tanaka and Hosaka, 1994).

5.2.2.2 Analysis of PITP - Post Incubation with ATP

160 Having established the apparent pi values of the starting PITP preparations, the effects of Ptdlns phosphorylation were investigated. Purified PITPs were incubated with radiolabelled or unlabelled ATP for 10 min as previously described (Section 4.3), and the products were separated by lEF. lEF gels were then analysed using the different methods described below to identify the positions to which PITPs, the co-purified lipid kinases and the labelled PPIs had migrated.

5.2.2.2.1 Western Blots

The separated proteins in lEF gels were blotted onto PVDF membranes and probed with anti-PITPa or anti-PITPp polyclonal antibodies, to determine (i) whether the production of PPIs had altered the relative amounts of each PITP charge isomer and (ii) whether any new PITP subforms had been generated (Section 2.1.2.1.3).

ATP-treated PITPa resolved again into three protein bands (two major and one minor) at the same iso-electric positions described above (Fig. 5.3A). There was no detectable change in the relative amount of each form.

In the case of ATP-treated PITPp, two observations were made (Fig. 5.3B). First, the minor band at pi 6.3 decreased in relative intensity while the two major bands remained essentially unchanged. Second, a small amount of the ATP-treated, but not untreated, PITPp remained near the sample origin.

5.2.2.2.2 SDS-PAGE

At the end of the lEF run, each sample lane was cut along the pH gradient into 2 mm- wide slices, which were further electrophoresed on a SDS-polyacrylamide gel. Subsequent Coomassie Blue staining confirmed the localization of each PITP at their respective pi values mentioned above (figure not shown). Silver staining further revealed traces of each PITP between the sample origin (4 mm wide) and at up to 14 mm (PITPa) or 16 mm (PITPp) from the origin (see for example Fig. 5.4). By contrast, no silver- stained protein was detected at or around the origin if the PITP sample was not pre­ treated with ATP (figure not shown). Such traces of PITP suggest slow mobility, multiple forms or a dynamic equilibrium between forms of different mobility.

The direction and distance migrated by these minute amounts of PITP varied in different gel lanes and appeared to correlate with the location and distance of the sample origin

161 A 7.5 ^ 7.1 — 6.8 - ^ ------sample origin

6.5 -

6.0 -

pi + + - ATP

B 7.5 - ^ M------sample origin

6.5 minor band 6.0 “ • M ^ ^ pi 6.3 pi + + _ ATP

Fig. 5.3 IFF Analysis of ATP-Treated PITPs 0.3-0.5 pg of purified (A) PITPa and (B) PITPp were incubated with 200 pM unlabelled ATP for 10 min at 37°C and then analysed by lEF (Section 2.1.2.1.3). The gels were then blotted onto PVDF membranes and probed with (A) anti-PITPa or (B) anti-PITPp antibodies. The figures are representative of three independent IFF experiments.

162 Distance from sample origin A towards anode (mm) 16 14 12 10 8 6 4 2 0 MW kDa

— — 66

— 46 ##

PITPa — ►

— 21.5

Distance from sample origin g towards anode (mm) 16 14 12 10 8 6 4 2 0 MW kDa

— 4 — 97.4

II « — 6 6

PITPp — ►

Fig. 5.4 Localization of ATP-Treated PITPa and PITPP on lEF Gels PITPa (1.7 pg) and PITPp (0.5 pg) were treated with 70 pM [y-32p]ATP (5 pCi) for 10 min at 37°C and then analysed by lEF (Section 2.1.2.1.3). Both samples were applied on separate lEF gel lanes near the cathode. After the lEF run, each sample lane was cut along the pH gradient into 2 mm-wide slices, each of which was further analysed by SDS-PAGE (Section 2.1.2.1.1). The SDS-PAGE gels of (A) PITPa and (B) PITPp were subsequently silver stained. The figures are representative of at least three independent experiments. MW, protein molecular weight standards. 1^3 from the final migration positions of most of the PITP (i.e., at pH 6.1 - 6.7). For example, in experiments depicted in Fig. 5.4, the samples were applied near to the cathode. The migration direction was found to skew towards the anode and the distance migrated was longer than that when the sample origin was further away from the electrodes (e.g., in Fig. 5.3B).

It was also noted that, in comparison with PITPp, detection of such trace amounts of

PITPa by silver staining required about three times more protein to be applied on the lEF gel (see details in the legend to Fig. 5.4). This may explain the inability to detect PITPa at the origin by Western blotting in experiments where a smaller amount of sample was used (Fig. 5.3A).

To investigate the migration positions of phosphorylated proteins and Pis, the lEF gel slices were also analysed individually in a scintillation counter. For both PITPs, counts were found to be concentrated in two areas of the lEF gel: at the anode (to which unutilised [y-^^P]ATP had migrated; not shown) and at the sample origin (which was up to 6 mm wide) (Fig. 5.5A and Fig. 5.6A). In order to determine whether these counts corresponded to lipid or protein phosphorylation, gel slices with counts significantly above background were subsequently electrophoresed on a 18% SDS-polyacrylamide gel and autoradiographed (Fig. 5.5B and Fig. 5.6B). Several phosphorylated proteins were found at or near the sample origin, but none has an apparent molecular mass identical to PITPa and PITPp. However, these phosphoproteins were not detected using silver staining (Fig. 5.4) and were not analysed further.

Gel slices from the sample origin also had a phosphorylated band which migrated ahead of the 14.3-kDa protein standard (Fig. 5.5B and Fig. 5.6B). This band appeared to correlate with the relative abundance of the phosphoproteins. Thompson et al. (1985) had previously identified a similar band in their experiments to be PtdIns4P. However, our attempts to extract lipids from these bands using methods similar to those described in their report were not successful.

5.2.2.2.3 Lipid Characterization

To circumvent the abovementioned difficulties encountered in extracting and identifying lipids from SDS-polyacrylamide gels, lipids were extracted directly from lEF gel slices instead. The gel slices were mixed vigorously with TLC solvent and the extracts resolved by TLC (Fig. 5.7). PtdlnsP was found only in slices at and adjacent to the sample origin 164 incorporated (cpm) N) U> towards O anode 8 § O

- 30 s recombinant ll PITPa - 20 g 3

- 10 0" I g ' S 2 o ^ §. 93. sample origin ]0 3 - -4 towards cathode

Distance from sample origin B towards anode (mm) 8 10 12 14 16 18

Stacking Gel

Resolving Gel

/ PtdlnsP \ V PtdlnsP, /

Fig. 5.5 SDS-PAGE Analysis of lEF Gel Slices of ATP-Treated PITPa Recombinant PITPa (0.5 pg) was treated with 70 pM [y-^^PJATP (5 pCi) for 10 min at 37°C and then analysed by lEF (Section 2.1.2.1.3). The lEF gel lane was then cut into 2 mm-wide slices and (A) counted or (B) further analysed by SDS-PAGE (18% gel) and autoradiography. The data and figure are representative of three separate experiments. 165 incorporated (cpm) ^ to u> LA o\ A towards o anode § § 8 8 8

- 30 2 recombinant PITPp -20 if § g-g*

- 10 Ii I " ^ §. sample origin ]« 1- - -4 towards cathode

Distance from sample origin B towards anode (mm)

Stacking Gel

Resolving Gel

Ptdlnsf ( Ptdlnsf

Fig. 5.6 SDS-PAGE Analysis of lEF Gel Slices of ATP-Treated PITPp Recombinant PITPp (0.5 pg) was treated as described for PITPa in Fig. 5.5. The data and figure are representative of three separate experiments. 166 A Ptdlns -

Ptdlns/*

PtdlnsP

18 20 22 24 Distance from sample origin towards anode (mm)

B Ptdlns -

Ptdlns/* -

PtdlnsPg - 0 2 4 6 8 12 14 16 18 20 22 24 Distance from sample origin towards anode (mm)

Fig. 5.7 Distribution of ^^P-labelled Phosphoinositides on lEF Gels 0.3-0.5 pg of purified (A) PITPa and (B) PITPp were incubated with 70 pM [y- ^^P]ATP (5 pCi) for 10 min at 37°C and then analysed by lEF (Seetion 2.1.2.1.3). Each gel lane was sliced into 2 mm-wide pieces, which were then individually sonicated overnight in TLC Buffer to extract lipids. Lipid extracts containing counts above background were further analysed by TLC (Section 2.1.3.2.1). The figures are representative of three independent experiments. The migration positions of the PI lipid standards are indicated.

167 for both PITPs. In the case of PITPp, Ptdlnsf] was also detected, but the amount varied with different PITPp preparations. These observations support the earlier suggestion that the radiolabelled material which migrated ahead of the 14.3-kDa protein standard in SDS polyacrylamide gels (Figs. 5.5B and 5.6B) was composed of Ptdlnsf and, in the case of

PITPp, also Ptdlns? 2 .

5.2.22A PI Kinase Assay

As PI 4- and 5-kinase activities had co-purified with PITPs and could employ PITP- bound substrate (Sections 4.3 and 4.4), their migration positions were also investigated. ATP-treated PITP samples were applied at different points between the electrodes in various lEF gel lanes. Each lane was then sliced into 2-mm pieces as before and assayed. For both PITPs, Ptdlns kinase activity was always detected exclusively in lEF gel slices near or at the sample origin. This is somewhat similar to the case of the PPIs (Section 5.2.2.2.3). Fig. 5.8 shows the data of a typical gel lane for each PITP.

In summary, the above data indicate that ATP treatment generated new form(s) of PITP which has relatively low mobility in lEF gels. In addition, all detectable PPIs and PI kinase activities were found to co-localize with the new PITP form(s) around the sample origin. The majority of each PITP that contain no detectable PI kinases or labelled Pis focused at their respective pi.

The co-localization of ATP-treated PITP and PPIs on lEF gels is consistent with the idea that the PI lipid remains bound to the transfer protein after phosphorylation by PI kinases. Their low mobility in lEF gels can be explained in the context of one or more types of large, multi-component signalling complex which comprise PPl-binding PITP, PI kinases and other undefined activities, such as PLC and DAGK (see Chapter 4). Such complexes may be too hydrophobic or too large to move through the lEF gel medium to focus at their pi. Enzymatic activities within such complexes would therefore be expected to remain near the origin, which was indeed observed. It is however noted with caution that the sensitivity of the PI kinase activity assay and the efficiency of the lipid extraction methods may have been affected by the different pH and charge properties of the gel slices.

5.3 Characterization of the Putative Complexes

168 A PtdlnsP —

Ptdlnsf^ ~

4 23 24 25 26 27 Distance from sample origin towards anode (mm)

B PtdlnsP —

Ptdlns?2 ~

8 12 14 20 22 24 26 28 30

Distance from sample origin towards anode (mm)

Fig. 5.8 Position of Co-Purified Phosphoinositide Kinase Activities on IFF Gels 0.3-0.5 pg of purified (A) PITPa and (B) PITPp were incubated with 200 pM unlabelled ATP for 10 min at 37°C and then analysed by lEF (Section 2.1.2.1.3). Each gel lane was then sliced into 2-mm pieces and assayed for PI kinase activity (Section 2.4.2). Lipid extracts containing counts above background were subsequently analysed by TLC (Section 2.1.3.2.1). The figures are representative of three independent experiments. The migration positions of the PI lipid standards are indicated.

169 An additional inference from the above results is that PITPs appeared to become far less mobile in lEF gels only after phosphorylation of the Ptdlns within their lipid-binding site. However, it was still unclear whether PITPs and the co-purified enzymes existed in a single or multiple types of complex, or within heterogeneous and artefactual aggregates, possibly containing insoluble membrane remnants. To help distinguish between these possibilities, the relative distribution of PITPs, PI kinase and DAGK activities following gel filtration chromatography of PITP preparations were investigated.

5.3.1 Gel Filtration Chromatography of PITP Preparations

A Superdex 200 column was used in place of a Superdex 75 column to achieve better separation resolution at the higher molecular mass range. It was first calibrated using molecular weight standards and a plot of relative elution volume versus the logarithm of molecular mass of the standards was constructed.

The elution profiles of PITPs overexpressed in Sf9 cells were essentially similar to those overexpressed in bacterial cells, although different gel filtration columns were used (compare Fig. 5.9 with Fig. 3.3, and Fig. 5.10 with Fig. 3.4).

5.3.1.1 PITPa

Recombinant PITPa eluted as a single main peak (peak Al) with a small shoulder (peak A2) at the leading edge, with apparent molecular masses of 25 kDa and 66 kDa, respectively (Fig. 5.9A). These values differ significantly from the calculated molecular mass (32 kDa) of human PITPa. However, on a Coomassie-stained SDS-polyacrylamide gel, both peaks Al and A2 appeared to be constituted primarily of PITPa (Fig. 5.9B).

In order to ascertain whether Al and A2 contained different oligomers of PITPa, both peaks were further analysed using a native polyacrylamide gradient gel (Fig. 5.9C). Surprisingly, not only did they each resolve similarly into two protein bands, but the apparent molecular masses of these bands, approximately 100 kDa and 120 kDa, did not match that predicted from their elution from the Superdex column. Thus, using different analytical methods, PITPa seemed to exhibit different apparent molecular masses.

5.3.1.2 PITPp

The phenomenon of inconsistent molecular mass was also observed in the case of PITPp. Its gel filtration profile consisted of three peaks, at apparent molecular masses of 290 kDa 170 Ài peak Al 0.4 -

280 peak A2 0.2 -

TTTT Vo 60 80 100 120 (42 ml) Elution volume Fe (ml)

B kDa C kDa 97.4

— 66.2 669 — 45 440 232 — 31 140

— 21.5 67 14.4 43 30 prep A2 Al MW

Al A2

Fig 5.9 Gel Filtration Analysis of Recombinant PITPa (A) Recombinant PITPa was purified from Sf9 cells and then analysed (1 mg in 1.2 ml) using a Superdex 200 16/60 gel filtration column (flow rate = 1 ml/min). (B) 10 pg of PITPa from purified preparations (prep) and peaks Al and A2 were analysed by SDS-PAGE. (C) Peaks Al and A2 were separately pooled and concentrated using Microsep Microconcentrators. Equal proportions of non-concentrated (none) and concentrated (cone) eluents were then analysed using native Tris-Glycine gels (8-16%). The figures are representative of at least three independent experiments. Fo, void volume; MW/LMW/HMW, low/high molecular weight standards. 171 à i peak B3 0 .3 - peak B1

0.2 - ■280 peak B2

Vo 60 80 100 120

Elution volume Fe (ml)

B kDa 9 7 .4 ^ kDa

66.2 — 669 — 45 440 — 232 — 31 — 140 — 21.5 — 67 ê é 14.4 — 43 30 MW prep B3 B2 B1

B3 B1

Fig. 5.10 Gel Filtration Analysis of Recombinant PITPp (A) Recombinant PITPp was purified from Sf9 cells and then analysed (0.8 mg in 1.5 ml) using a Superdex 200 16/60 gel filtration column (flow rate = 1 ml/min). (B) 10 pg of PITPp from purified preparations (prep) and peaks B1-B3 were analysed by SDS- PAGE. (C) Peaks B1 and B3 were separately pooled and concentrated using Microsep Microconcentrators. Equal proportions of non-concentrated (none) and concentrated (cone) eluents were then analysed using native Tris-Glycine gels (8-16%). The figures are representative of at least two independent experiments. Fo, void volume; MW/LMW/HMW, low/high molecular weight standards.

172 (peak B3), 60 kDa (peak B2) and 23 kDa (peak Bl), respectively (Fig. 5.10A). By virtue of having similar elution volumes, peaks Bl and B2 were thought to contain the PITPp counterparts of the PITPa subforms in peaks Al and A2, respectively. Analyses by native and denaturing polyacrylamide gel electrophoresis, however, showed that only peaks Bl and B3 contained PITPp while peak B2 was devoid of detectable levels of protein. The latter observation was made in at least two independent experiments and unlikely to be an experimental error. This is in contrast to bacterially-expressed PITPp which also eluted in three peaks, all of which have detectable PITP (see Section 3.2.2). This also raised the question of whether peak A2 (Fig. 5.9) is also devoid of PITPa and that the two protein bands detected through native-PAGE of peak A2 (Fig. 5.9C) were derived from the leading edge of peak Al.

In addition to having a similar elution volume to peak Al, peak Bl also resolved into two bands on native gel electrophoresis (Fig. 5.IOC). Surprisingly, the apparent molecular masses of these bands (165 kDa and 230 kDa) not only did not correspond to their elution volumes (as previously observed with peak Al), but also differed from the two PITPa protein bands (100 kDa and 120 kDa) of peak A l.

Peak B3 was also mostly made up of PITPp (Fig. 5.1 OB), but when Coomassie-stained on a native gel, it appeared as a smear with a leading band having an apparent molecular mass of 170 kDa (Fig. 5.IOC). Purified PITPp also produced a smear on lEF gels (Fig. 5.2A). It is unclear whether the smearing of peak B3 on a native gel indicates variations in oligomerisation, solubility, charge, or associated enzymes. Hence, the number of discrete constituents in peak B3 is difficult to ascertain accurately.

5.3.2 Distribution of Co-purified Kinase Activities

To investigate whether PI kinase and DAGK activities remained associated with PITPs during gel filtration, eluent fractions were collected and assayed (Sections 2.4.2 and

2.4.4). Fraction collection began just before the void volume (Vq = 42 ml) and continued till the bed volume (Ft = 120 ml) was reached.

5.3.2.1 PITPa

Weak PtdIns4K activity was detected in Fractions 5 to 36 (Fig. 5.11), which spanned a wide molecular mass range from 190 kDa to about 2,000 kDa. This broad peak of

173 25 kDa A t

66 kDa ■280

Fraction

O h

10 15 20 25 30 35 Fraction c Ptdlns

Ptdlnsf

Ptdlnsfn

9 10 11 12 13 14 15 16 17 18 19 20

Ptdlns

m Ptdlnsf 0

Ptdlns/*n 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36

Fig. 5.11 Phosphoinositide Kinase Activities in Gel-Filtered PITPa After (A) gel filtration of recombinant PITPa through Superdex 200 (as described in Fig, 5.9), (B) all fractions were assayed for PI kinase activity using exogenous Ptdlns (Section 2.4.2). Lipids were extracted, separated by TLC, and quantified using a phosphoimager. (C) Autoradiographs were developed after two weeks of exposure to the TLC plates. The data are representative of two independent experiments.

174 PtdIns4K activity reached a maximum at approximately 680 kDa (fractions 18-20). PLC and DAGK activities were not detectable (not shown).

5.3.2.2 PITPP

PtdIns4K, PtdlnsPSK and DAGK activities were found to co-elute in a single peak with an apparent molecular mass of 450 kDa, at the leading edge of peak B3 (Fig. 5.12). The fractions were also assayed for the presence of PLC, but no enzyme activity was detected, presumably because it was too dilute following chromatography. The co-elution of these three kinase activities in a single peak lying well within the resolving range of the Superdex column provides vital support for the notion of an integrated PLC signalling complex.

Although the kinase activities eluted within peak B3, it was unclear whether any PITPp was associated with the lipid kinases in the putative signalling complex. To investigate this question, selected kinase-containing fractions were treated with Ni^"^-NTA resin to extract the Hisg-tagged PITPp. The volume of Ni^'^-NTA added was calculated to have just sufficient capacity to only bind all of the PITPp within each fraction. The amount of

PITPp in each fraction was estimated by comparing the absorbance (A 280 ) of each fraction

with a standard curve constructed from known amounts of PITPp, as well as assuming

that at least 95% of each fraction was PITPp.

When subsequently assayed, the supernatants of resin-treated fractions no longer contain­ ed detectable amounts of PI kinase (Fig. 5.13A) and DAGK (Fig. 5.13B) activities. This

strongly suggests that PITPp was associated with these kinases and supports the ‘co­ factor model’ in which PITP acts as a co-factor in a PI signalling complex.

5.4 Discussion

PITPa purified from native sources typically has two-thirds of the protein binding Ptdlns and the remaining one-third binding PtdCho (Thomas et al., 1994; van Paridon et al., 1987b). In contrast, recombinant PITPs that were overexpressed and purified from Sf9 cells appeared to have equal amounts of the PITP-Ptdlns and PITP-PtdCho subforms. This is surprising because (i) PITPs have shown much higher affinity for and/or activity towards Ptdlns than PtdCho (de Vries et al., 1995; van Paridon et al., 1987a; Westerman et al., 1995; Wirtz, 1991, and references therein) and (ii) Sf9 cells have a higher Ptdlns content in comparison to mammalian tissues (Marheineke et al., 1998). This difference in

175 290 kDa A à i 23 kDa

60 kDa ^280

Fraction B 5 H

3> 4 -

X I 3 -

I 2 - I

1 -

0 mini- 0

Fraction C 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32

PtdlnsP

PtdlnsA

D 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32

ptdOH -►

Fig. 5.12 Phosphoinositide Kinase and DAG Kinase Activities in Gel-Filtered PITPp After (A) gel filtration of recombinant PITPp through Superdex 200 (as described in Fig. 5.10), (B) all fractions were assayed for lipid kinase activities using exogenous Ptdlns (open squares) or DAG (solid diamonds) (Sections 2.4.2 and 2.4.4). (C) PPIs and (D) PtdOH were extracted, separated by TLC, and quantified using a phosphoimager. In longer exposures, the distribution of Ptdlns ? 2 was clearly similar to PtdlnsP and PtdOH (not shown). The data are representative of two independent experiments. 21 24 Fraction

” “ H- " “ -I- H- NP^-NTA treatment

Ptdlns

PtdInsP

PtdInsPj

engin

25 26 25 26 Fraction B H- + NP^-NTA treatment

PtdOH

Fig. 5.13 Association of Lipid Kinases with PITPp 250 |il of selected gel filtration fractions from the separation shown in Fig. 5.12 were treated with or without Ni^^-NTA resin for 20 min at 4°C. The volume of resin (50% suspension) used for fractions 21,24, 25, and 26 were 20 pi, 40 pi, 50 pi, and 75 pi, respectively. The supernatants of resin-treated fractions were subsequently assayed for (A) PI kinase and (B) DAG kinase activities remaining in solution (Sections 2.4.2 and 2.4.4). The figures are representative of two separate experiments. The positions of lipid standards were indicated.

177 the ratio of PITP subforms may be due to the significant differences in phospholipid composition betweens Sf9 and mammalian cells (Renkonen et a l, 1972). Specifically, Sf9 cells lack phosphatidylserine, glyco- and sphingolipids (Marheineke et a l, 1998), and its membranes may hence exhibit physical properties, such as charge and fluidity, that are different from mammalian cells and which would have a large influence on PITPa activity (Helmkamp, 1980a; Helmkamp, 1980b; Somerharju et a l, 1983). Furthermore, the acyl chains of PtdCho in Sf9 cells are generally less saturated than those in mammalian cells (Marheineke et a l, 1998) and this may affect the affinity of PITPa for PtdCho (van Paridon et aL, 1988a).

A minor charge isomer was also detected following lEF of both recombinant PITPs. However it is unlikely to be PITP-PtdEtn, even though PtdEtn is a major constituent lipid of Sf9 cells (Marheineke et aL, 1998) and PITPa has low affinity for PtdEtn (DiCorleto et

al., 1979), because this form of PITPa would be expected to have a pi value close to that

of PITPa-PtdCho, since both the ethanolamine and choline headgroups carry a single positive charge. The minor form is also unlikely to be a SM-binding isomer, since sphingolipids were not detected in Sf9 cells (Marheineke et a l, 1998). As PtdlnsP was not detected among the lipids extracted from the PITP preparations, the minor band is also unlikely to be PITP-Ptdlnsf.

The pi value of the minor protein band, which is slightly lower than that of PITP-Ptdlns, makes it tempting to speculate that it contains a phosphorylated form of PITP (Geijtenbeek et al., 1994; Snoek et al., 1993; van Tiel et al., 2000b). The negative charge of one or more phosphate residues could account for its lower pi value. Since Sf9 cells were incubated in 10% PCS during protein expression, agonist-dependent protein kinases that recognise PITPs, such as PKC, may be active (Snoek et al., 1993; van Tiel et al., 2000b). Indeed, a recent study also reported the separation of recombinant murine PITPg- His6 by two-dimensional PAGE into two charge isomers, with pH (pi) 6.2 and 6.5 (van Tiel et al., 2002). The authors inferred that the isomers were phosphorylated and non- phosphorylated forms of PITPp-Ptdlns, respectively. The similar pi values suggest that they are equivalents of the two charge isomers of pi 6.3 and 6.5 in Fig. 5.2C.

Following incubation with ATP, a small amount of PITP became less mobile in lEF gels and remained near the sample origin. The tapering of PITP proteins away from the sample origin suggests that the low mobility fraction might have been in equilibrium with the bulk of the PITP. In the same experiment, all retrievable ^^P-labelled PPIs were found 178 only near the sample origin, suggesting that PPIs produced during incubation with ATP may have remained bound to PITP. What could have caused the low mobility? One possibility that has already been discussed is that the low mobility was due to association with a large molecular complex. A second possibility is that the presence of a PPI in the lipid-binding site may have rendered PITP relatively insoluble in aqueous media - this would be consistent with studies using PPI micelles which demonstrated that (i) PITPa is able to bind strongly but not transfer PtdlnsP and Ptdlnsfi (Schermoly and Helmkamp, 1983; van Paridon et al., 1988b) and that (ii) PPIs, even at low concentrations, efficiently inhibit the lipid transfer ability of PITPa (van Paridon et al., 1988b). This loss of solubility and/or mobility of a PPI-binding PITP may help explain (i) the lack of migration of ATP-treated PITP from the sample origin and (ii) why PITP-bound PPIs were difficult to retrieve following gel filtration (Section 5.2,1).

PI kinase activities that co-purified with PITP were detected only near the sample origin on the lEF gel following incubation with ATP, and appeared to co-migrate with the labelled PPIs and the minor PITP fraction. Although the iso-electric-focusing pattern of PI kinases in the starting PITP preparations remains unclear, the above data points to the continued association of lipid kinases with PITP following the phosphorylation of the PITP-bound Ptdlns substrate. This is consistent with the co-factor model of PITP function in PI signalling.

While PITPa and PITPp exhibited rather similar lEF resolution patterns, their migration during gel filtration and native gel electrophoresis were different. When applied to a Superdex 200 gel filtration column, PITPa eluted as one main peak, whereas PITPp separated into two peaks. An additional minor peak in each preparation did not appear to contain PITP. In addition, neither PITP eluted at their respective molecular mass.

Such distinctive elution profiles were also observed when bacterially-expressed PITPs migrated through a Superdex 75 column (see Section 3.2.2). However, there are significant differences, especially with the abovementioned minor peaks. When eluting through a Superdex 75 column, PITPa produced a less prominent but broader minor peak

(Fig. 3.3A), whereas in the case of PITPp, the ‘minor’ peak is a major one (Fig. 3.4A). Furthermore, both peaks clearly contained proteins (Figs. 3.313 and 3.4B), in contrast to their counterparts in the elution profiles of PITPs expressed in Sf9 cells. The reasons for these differences remain unclear, but a plausible explanation may be that, in bacteria.

179 PITPs bind proteins that are absent in Sf9 cells, such as the 70 kDa and 28 kDa contaminants in Figs. 3.3B and 3.4B, respectively.

Several reports suggest that such elution profiles are unlikely to be due to the type of expression host, the lipid bound to PITP or the C-terminal Hise tag (see Table 1). Native PITPa purified from bovine brain and heart also eluted later than expected from their molecular masses (DiCorleto et al., 1979; Thomas et al., 1994), while Wirtz and co­ workers detected multiple elution peaks which showed immunoreactivity to an antibody that recognises PITPp during gel filtration of bovine brain cytosol (de Vries et ah, 1995).

It is unclear why PITPs eluted later than expected from their molecular mass. The molecular structure of PITPa is basically globular (Yoder et al., 2001) and not expected to retard its passage through the gel filtration matrix. PITPp, which shares 77% sequence identity with PITPa, would also be expected to adopt a similar structure. Interestingly,

Prosser et al. (1997) reported that while wildtype PITPa eluted with an apparent molecular mass of 20 kDa, putative dimers of PITPa that were truncated at their C- terminus region eluted earlier than a hypothetical dimer of full-length PITPa. It is thus plausible that the C-terminal region of PITP may have limited affinity for the gel filtration matrix.

It is noted with interest that mouse PITPa dimerises in the ‘apo’ (phospholipid-free) state, at least in a crystal (Schouten et al., 2002), and such dimers may constitute peak A2, which has an apparent molecular mass (66 kDa) that is twice that of a PITP monomer. Nevertheless, it remains uncertain whether peak A2 was also devoid of protein, as in the case of peak B2.

In addition to having different gel filtration elution profiles, recombinant PITPs migrated through native gels at apparent molecular masses that are inconsistent with their calculated molecular masses as well as their respective gel filtration elution volumes. Furthermore, the mobility of PITPa (peak Al) and PITPp (peak Bl), which had similar gel filtration elution volumes, were dissimilar through a native gel. One possible reason for this inconsistency in molecular masses could be that PITPs have a low charge density - a protein with lower charge density migrates at a slower rate through a native gradient gel than one with higher charge density. That charge density might have played a role in PITP electrophoretic mobility in native gels is corroborated by the resolution of peaks AI and Bl into two protein bands. These doublets presumably contained different PITP

180 Apparent Native Source/ Apparent Mr PITP Gel Filtration Media Reference Mr (kDa) Expression Host * In Native Gels (kDa)

Thomas et al. (1994) rat brain ND

Superdex 75 19-20 Prosser et al. (1997) E. coli ND

PITPa Fig. 3.3 E. coli ND

Superdex 200 25 Fig. 5.9 100, 120

Sephacryl SI00 23.5 DiCorleto et al. (1979) bovine heart ND

17.5 Superdex 75 60 Fig. 3.4 E. coli ND 200

PlTPp 23 165,230 Superdex 200 290 Fig. 5.10 170 (at the leading edge of protein band)

Sephacryl SI00 NDt de Vries et al. (1995) bovine brain ND

Table 1 Apparent Molecular Masses of PITP In Various Gel Filtration Media PITPa and PITPp eluted anomalously from various gel filtration media with apparent molecular masses that differ from their respective calculated molecular masses. This anomaly was also observed in their electrophoretic mobility in native gels. Mr, molecular mass; ND, not 2 determined, t, three peaks detected (two major, one minor). *, expression hosts are italicised. charge isomers as a consequence of binding either Ptdlns or PtdCho and therefore having different pi values (and thus charge density and mobility). Moreover, the comparatively lower pi values of PITPa compared with PITPB charge isomers suggest a higher net negative charge in the former. This could account for the higher mobility of PITPa compared with PITPp charge isomers in native gels run at pH 8.3.

Peak B3, with an average apparent molecular mass of 290 kDa and composed mostly of PITPp, must have contained PITPp oligomers. The smear produced on the native gel suggests that this peak contained a heterogeneous mixture of PITPp oligomers, either as discrete homo-oligomers or in various complexes with co-purified factors and/or membrane remnants. Native PITPp has been detected in several elution peaks during gel filtration (de Vries et a l, 1995), and the form of PITPB contained in peak B3 might have increased as a result of protein overexpression. The physiological significance of PITPp oligomerisation remains to be established, but the co-elution (Fig. 5.12) and association (Fig. 5.13) of all co-purified PI kinase and DAGK activities with PITPp within peak B3 suggests that oligomerisation may be facilitating interactions with a complex of PI signalling enzymes (see Chapter 4). These deductions support the notion of PITP being part of an integrated PI signalling complex. As suggested in Chapter 4, PITPp oligomers may be able to present Ptdlns more efficiently as substrate to the lipid kinases and thus help explain the higher production of PPIs observed in ATP-treated PITPp preparations.

In contrast, the elution pattern of the smaller amount of PI kinase activities that co- purified with PITPa was more diffuse. It is tempting to speculate that this might be due to the apparent lack of equivalent oligomers in PITPa preparations. It is also possible that, by virtue of its localization at the plasma membrane, PITPa associates with a greater variety of PI signalling complexes, which may also be bound to different types of receptors.

It was noted with interest that the lipid kinase activities which co-purified with PITPp eluted with an apparent molecular mass (450 kDa) that is similar to a protein complex (PEPl; apparent molecular mass 500 kDa) identified by gel filtration of a cytosolic fraction of PC 12 neuroendocrine cells (Hay et al., 1995). PEPl (priming in exocytosis protein 1) has been shown to contain 68-kDa and 90-kDa PtdIns4P5K activities. These enzymes, one or more as yet uncharacterized proteins (PEP2; apparent molecular mass 120 kDa) and PITP (PEP3; apparent molecular mass 20 kDa) are ATP-dependent priming

182 factors which promote Ptdlns?] synthesis that is required for Ca^^-activated noradrenalin secretion (Hay et a l, 1995; Hay and Martin, 1993). The G protein Rac also forms a 500- kDa complex in vivo with PtdIns?5K, DAGK and possibly Rho guanine nucleotide dissociation inhibitor (RhoGDI) (Tolias et a l, 1998a). As Rac and RhoGDI have been implicated in secretion (O'Sullivan et a l, 1996; Price et a l, 1995), the Rac-RhoGDI-lipid kinase complex may be regulating secretion, for example, by raising Ptdlns?] levels to stimulate the activities of both ARF and PLD (see Fensome et a l, 1996, for details; Liscovitch and Cantley, 1995). Incidentally, Rac also associates in vivo with a Ptdlns kinase but the type of kinase was not confirmed (Tolias et a l, 1995). Consequently, it will be interesting to determine whether PEPl, the Rac-RhoGDI-lipid kinase complex, and the PITPp-PLC-lipid kinase complex share other common constituents besides PtdIns?5K and DAGK.

Finally, the gel filtration data provide strong evidence against the possibility that PPIs and the PI kinases were embedded in or associated with membrane remnants that may have co-purified and co-eluted with either PITP. Hence, it would be interesting to further investigate the association of PITP with PPIs and lipid kinases using, for example, a liquid-phase lEF system, which should allow easier retrieval of the putative PITP-PPI-PI kinase complex for more detailed analysis.

183 Chapter 6

Discussion

6.1 Proposed Models of PITP Function

The role of the lipid Ptdlns as a signalling molecule was consolidated in the 1970s when it became clear that the phosphorylinositol group of Ptdlns undergoes chemical modifications in a cyclic series of reactions (reviewed in Michell, 1975). This ‘Ptdlns cycle’ of reactions describes the increased turnover of Ptdlns and PtdOH that are provoked by physiological stimuli and the marked changes in Ptdlns concentrations at different cellular sites. Phospholipid transfer proteins were proposed to shuttle metabolites of the Ptdlns cycle between different cellular membranes where Ptdlns is being consumed (plasma membrane) and re-synthesized (endoplasmic reticulum). This model of action helps rationalize observations that hormonal stimulations can cause hydrolysis of up to half of inositol lipids of some cells (Michell, 1975; Michell et al., 1988, and references therein).

Data from a series of labelling experiments led Downes and Michell (Downes and Michell, 1985) to propose the compartmentation of inositol lipids into two metabolic or functional pools, only one of which is susceptible to hydrolysis by agonist-stimulated PLC. Subsequently, the hydrolysis and resynthesis of Pis were proposed to occur in a ‘closed’ cycle in which the lipid synthesized in response to agonist is preferentially sensitive to subsequent agoinst-induced hydrolysis (Michell et al., 1988; Monaco and Gershengom, 1992). However, progress in elucidating the organization and regulation of the Ptdlns cycle has been slow.

The use of permeabilised or broken cells in experiments to investigate the Ptdlns cycle is likely to compromise or preclude endogenous Ptdlns synthesis. For example, permeabilised neutrophils depleted of cytosol were capable of Ptdlns synthesis provided that CTP and inositol are also present (Whatmore et al., 1999). In addition, studies using blowfly salivary glands showed that only 5% of Ptdlns could be utilised for signalling under conditions in which Ptdlns synthesis was inhibited (Fain and Berridge, 1979). Hence, it is uncertain whether the observed effects of PITP in restoring agonist-induced PI metabolism in permeabilised cells reflects its role in intact cells or simply its ability to compensate for preparational deficiencies in Ptdlns synthase activity. Furthermore, the cytoskeleton in SLO-permeabilised cells were disrupted (Figs. 4.15 and 4.16). As PI

184 kinases, PLC and DAGK have been shovm to associate with the cytoskeleton (Payrastre et al., 1991), it is unclear whether this affects their enzymatic activities and/or their access to substrate. Nevertheless, as up to 90% of cellular Ptdlns and PPIs may be consumed in various agoinst-stimulated cells (for example, see Vamai and Balia, 1998), lipid transport between the plasma membrane and intracellular membranes must be necessary, and membrane transport/sorting is apparently not sufficiently fast to support the observed PI turnover (see for example. Batty et al., 1998). It is therefore likely that PITP contributes significantly towards substrate supply to the PI signalling sites.

The mechanism by which PITP mediates Ptdlns transfer within this closed cycle remains unclear. This is complicated by the observation that the size of the agonist-sensitive Ptdlns pool within a cell type may vary with different types and/or concentrations of agonists (Monaco and Gershengom, 1992, and references therein). Nevertheless, it is clear that PITP, most probably PITPa, plays an important role in PLC-mediated signalling (Allen et a l, 1997; Kauffmann-Zeh et al., 1995; Thomas et al., 1993), by promoting the synthesis of Ptdlnsf 2 (Cunningham et al., 1995). Downes and co-workers proposed that PITP acts only to transfer Ptdlns down a concentration gradient from one or more intracellular sites of Ptdlns synthesis to the depleted Ptdlns pool in the plasma membrane, through specific targeting to PI kinase-associated receptors - the ‘targeted replenishment model’ (Batty et a l, 1998; Currie et al., 1997). Others suggested that PITP plays a more active role as a ‘co-factor’ for Pl-metabolising enzymes. Specifically, PITP may act by presenting the Ptdlns substrate to PtdIns4K, PtdlnsPK and PLC in series within a single complex - the ‘co-factor model’ (Cunningham et al., 1995; Hsuan, 1993). The latter model postulates that PITP-bound Ptdlns is the preferred substrate for the PLC- signalling pathway and accounts, at least in part, for a Ptdlns pool that is sensitive to agonists. Bankaitis and co-workers further proposed that PITP presents Ptdlns by ‘lifting’ the lipid away from the plane of the membrane bilayer such that the inositol headgroup is rendered susceptible to chemical modifications by lipid kinases (Kearns et al., 1998a).

6.2 Interaction with PI Signalling Enzymes

PITPa is shown to localize in the nucleus and throughout the cytoplasm, while PITPp is preferentially associated with perinuclear membrane structures (Figs. 4.15 and 4.16). This is consistent with observations reported elsewhere (de Vries et al., 1995; de Vries et al., 1996; Segui et a l, 2002). Such localizations are apparently independent of the intracellular concentration of PITPs (de Vries et al., 1996; van Tiel et al., 2000a). Wirtz

185 and colleagues did not detect either PITP at the plasma membrane of various cell types, and this result was used to argue against the co-factor model of PITP function in PI signalling. Figs. 4.15 and 4.16, however, show that PITPa, but not PITPp, may be partially localized at the plasma membrane of HEK cells. This makes it tempting to speculate that PITPa, rather than PITPp, is the primary PITP isoform that participates in agonist-stimulated, PLC-mediated PI signalling at the plasma membrane. Furthermore, PITPa co-immunoprecipitated with the type II PtdIns4K, PLCy and EGFR, all of which are associated with the plasma membrane in EGF-treated A431 cells (Kauffmann-Zeh et a l, 1995). Hence, at least in A431 and HEK cells, PITPa is capable of associating with the plasma membrane and/or membrane proteins.

As PITPp is primarily localized to perinuclear structures (Fig. 4.14), it may provide the Ptdlns transfer activity previously implicated in vivo in cellular secretion (Fensome et al., 1996; Hay and Martin, 1993) and/or vesicle formation (Jones et al., 1998; Ohashiet al., 1995; Simon et al., 1998) and transport (Paul et a l, 1998). Several of these processes also require PI phosphorylation (Fensome et a l, 1996; Hay et a l, 1995; Jones et al., 1998; Ohashiet al., 1995), which, taken together with the constitutive membrane localization observed here, may partially explain the greater amount of PI kinase activities that consistently co-purified with PITPp compared to PITPa (Figs. 4.4 and 4.6).

Both PITPs appeared to bind type II Ptdlns 4-kinase, which has been found in several intracellular membranes as well as the plasma membrane (Cantley et al., 1998). This laboratory had previously demonstrated the co-immunoprecipitation of PITPa with

PtdIns4K, PLCy and EGFR (Kauffmann-Zeh et al., 1995), while MrdgBl was recently shown to associate with an endogenous PtdIns4K and recombinant type III PtdIns4K expressed in COS-7 cells (Aikawa et a l, 1999). Preliminary data suggests that PITPa also associated with endogenous PI kinases when overexpressed in HEK cells (Fig. 4.14).

More significantly, recombinant PITPa and PITPp co-purified with endogenous type II PtdIns4K, PtdInsR5K, PLC and DAGK when expressed in Sf9 cells, which suggest they may interact in vivo as described by the cofactor model. This interaction appeared to be specific to PITPs, as no PI kinase and lipase activities co-purified with the antigen A33. PtdIns4K, PtdInsP5K, PLC and DAGK activities had previously been shown to associate with the cytoskeleton of A431 cells where they may be organized into an integrated PI signalling complex (Payrastre et a l, 1991). Indeed, these Pl-metabolising enzymes and

186 PITP behaved as a complex by co-eluting in gel filtration and co-binding to Ni^^-NTA resin (Fig. 5.13). This suggests that PITP may be an integral part of the putative complex.

This complex has an apparent mass of 450 kDa, which is similar to the protein complex PEPl (500 kDa) that contains PtdInsP5K activity and is one of three ATP-dependent priming factors required for Ca^^-activated noradrenalin secretion by promoting Ptdlnsf] synthesis (Hay et ah, 1995). It is also similar to another 500-kDa protein complex that contains Rac, PtdInsP5K, DAGK and possibly RhoGDI (Tolias et a l, 1998a). Cantley and co-workers proposed that DAGK acts upstream of PtdInsf5K within the complex, such that the PtdOH synthesized by DAGK stimulates PtdInsP5K to produce PtdlnsPi, which will further act on Rac and/or regulate actin uncapping or polymerization. However, in purified PITP preparations, Ptdlnsfz can be produced in the absence of PtdOH generation (for example, see Fig. 4.8), suggesting that DAGK is likely to be downstream of PtdInsR5K. As Rac also associated in vivo and co-immunoprecipitated with a Ptdlns kinase (Tolias et al., 1995), it is possible that the Rac-lipid kinase complex may be a PI signalling complex. It will therefore be interesting to see whether PEPl, the Rac-RhoGDI-lipid kinase complex and the PITPp-PLC-lipid kinase complex share other common constitutents and whether they are members of a novel family of signalling complexes.

It remains to be addressed whether the interaction of PITP with the lipid kinases and PLC occurs within the context of scaffold proteins. Some scaffolds act to localize a specific group of molecules that participate in the same signalling pathway to a specific intracellular site, while others serve to catalyse the activities of pathway enzymes by binding them in an orientation that directly enhances signal progression (Burack and Shaw, 2000). Hence, scaffolds may enhance the efficiency of signal propagation. Such scaffold proteins include InaD (see Chapter 1), the cytoplasmic domain of the PDGFR (Rudd, 1999), and the mitogen-activated protein kinase scaffold Ste5 (Choi et a l, 1994; Marcus et al., 1994). The aggregation of pathway enzymes to a particular scaffold is in line with the notion of a discrete cellular complex of PI signalling enzymes. It will be interesting to determine whether the putative PI signalling complex is analogous to the InaD ‘transducisome’ (Tsunoda et al., 1997). New techniques and assays will need to be developed to discern this issue.

6.3 Viability of PITP-bound Ptdlns as a Substrate

187 In the presence of ATP, the Ptdlns bound to PITP was phosphorylated by the co-purifying PtdIns4K and Ptdlns/*5K. PITP-bound Ptdlns appeared to be the preferred substrate and the reactions apparently proceeded efficiently, because addition of exogenous micellar Ptdlns triggered no detectable increase in PPI production (Fig. 4.6) The experiments using a [^^P]-phosphate chase indicated that the generation of Ptdlns ? 2 was due to sequential phosphorylation by the PI kinases of PITP-bound Ptdlns, not contaminant PtdlnsP (Fig. 4.4). The Ptdlnsfi was subsequently hydrolysed by a co-purifying PLC to InsPa and DAG, and the co-purified DAGK further phosphorylated DAG to PtdOH. Generation of PPIs, inositol phosphates and PtdOH were critically dependent on the presence of PITP-bound Ptdlns as these reactions were absent in ATP-treated T58D PITPs (Fig. 4.8), which are unable to bind and transfer Ptdlns (Fig. 4.7). This indicates that PITP-bound Ptdlns served as a direct substrate for the Pl-metabolising enzymes, which acted in series on the lipid, and the spontaneity of the reactions further suggests the easy accessibility of PITP-bound Ptdlns. The simultaneous co-purification of PtdIns4K, PtdInsP5K, PLC and DAGK with PITP as an apparent complex and the viability of PITP- bound Ptdlns as a direct substrate for these enzymes thus provide strong support for the co-factor model of PITP function in PI signalling.

An important issue in the ‘co-factor model’ is whether the PI lipid remains bound to PITP as it is further metabolised. PITPa has extremely high affinity for Ptdlnsf and Ptdlnsf] but is unable to transfer them (Schermoly and Helmkamp, 1983; van Paridon et ah, 1988b). Furthermore, very low concentrations of PPIs completely block lipid transfer by PITPa (van Paridon et ah, 1988b). These data hinted that, following presentation to and phosphorylation by PtdIns4K, the PI lipid may remain strongly bound to PITP for subsequent presentation to PtdlnsPK and PLC for phosphorylation and hydrolysis, respectively. During this process, PITP would be unable to bind or transfer other phospholipids. This hypothesis may explain the reduced mobility in lEF gels of a small amount of ATP-treated PITP and all detectable PPIs and PI kinase activities, which were retained together at the sample origin (see Section 5.2.2.2). This result is consistent with the formation of PITP-polyphosphoinositide-PI kinase complexes. Nevertheless, it can also be argued that in vivo, PITP releases the bound Ptdlns as the lipid is being presented to the signalling complex, and abstracts another phospholipid from the membrane. Such lipid exchange is unlikely to occur in membrane-free preparations of recombinant PITP. PITP-PPI complexes may therefore be generated by default as artefacts in these assays.

188 Attempts to retrieve the PITP, PPIs and PI kinases simultaneously from the lEF gels in non-denaturing conditions were however unsuccessful. This may be resolved in future by using a liquid-based lEF system. Alternatively, the Ptdlns bound to PITP could be replaced with Ptdlns analogues that contain chemically reactive groups on their acyl chains - activatable analogues would allow controlled covalent attachment to PITP after incubation with ATP or PI signalling reconstitution. Any PITP-PPI complexes may then be isolated with greater ease using denaturing conditions for further analysis.

The PtdIns4K that co-purified with PITP was sensitive to adenosine, suggesting that it is a type II enzyme. However, enzymatic activity was not significantly increased in the presence of TXlOO or exogenous Ptdlns (Figs. 4.3 and 4.6). This may indicate that presentation of Ptdlns to the kinase by PITP is much more efficient than free Ptdlns micelles. Comparatively more PPIs, inositol phosphates and PtdOH are always generated in ATP-treated PITPp compared with PITPa preparations (Figs. 4.4, 4.6 and 4.8). As

PITPa and PITPp co-purified with approximately equivalent amounts of PI kinase and DAGK activities (Figs. 4.9 and 4,10), this infers that the lipid kinases that co-purified with PITPp have higher apparent activities than those that co-purified with PITPa. Further studies are needed to address whether the difference in activity is due to co- purification of PITPa and PITPp with different lipid kinase and lipase isozymes, to one or more unidentified co-purifying factors, or to an ability of PITPp to present Ptdlns more efficiently as a substrate than PITPa, possibly, by forming homo-oligomers (see below). The recent cloning of type II PtdIns4Ks will help to address some of the above issues (Barylko et al., 2001 ; Minogue et al., 2001).

6.4 Mechanism of Interaction

Despite their different intracellular distribution, PITPa and PITPp showed equivalent functional capabilities in reconstituting agonist-stimulated PtdIns4K and PLC activities in permeabilised A431 and HL60 cells (Figs. 3.7 and 3.8). This functional redundancy was also demonstrated in the formation of secretory vesicles from the trans-Qo\g\ network (Ohashiet al., 1995) and the stimulation of pl50*PtdIns3K activity in vitro (Panaretou et al., 1997). Surprisingly, the yeast PtdlnsTP, Secl4p, was able to substitute for PITPs in reconstituting PLC signalling (Cunningham et al., 1996) as well as secretory vesicle formation (Ohashi et al., 1995) and priming (Hay and Martin, 1993), despite the lack of similarity in their primary and tertiary structures.

189 The biochemical and structural basis of the apparent redundancy between PITPa, PITPp and Secl4p in these assays remains to be elucidated, but it is clear that the ability to bind and transfer Ptdlns is essential. Secl4p^^^’^^^^, which exhibits no detectable Ptdlns transfer activity, failed to stimulate secretory vesicle priming and PLC activity (Phillips et al., 1999). However, C-terminally truncated PITPa, which can still transfer Ptdlns in vitro, not only was incapable of restoring G-protein-mediated PLCp signalling, but also blocked the signalling function of wildtype PITPa (Kara et al., 1997). This ability to uncouple Ptdlns transfer from PLC reconstitution infers that the former alone is insufficient to enable a PtdlnsTP to restore PLC signalling and supports the view that the function of PITP is not just the passive transfer of Ptdlns from its site of synthesis to the plasma membrane. Similar functional uncoupling was also observed in the RdgB-PITP domain - introduction of the T59E mutation in the RdgB-PITP domain had no apparent effect on its Ptdlns and PtdCho transfer ability but abolished its rescue of retinal degeneration in rdgB^ flies (Milligan et al., 1997). It will therefore be interesting to compare the PLC-reconstituting capability of PITP and the RdgB-PITP domain and their T58D or T58A mutants, which have no Ptdlns transfer activity.

While the ability to bind and transfer Ptdlns is necessary for reconstituting the abovementioned assays, it does not appear to be essential for interactions with PI- modifying enzymes. T58D PITPs, which bind only PtdCho, appeared to co-purify with equivalent amounts of PI kinase activities as wildtype PITPs (Fig. 4.9). Hence, interactions between PITPs and the putative PI signalling complex or its subunits are unlikely to be determined by the type of phospholipid bound on the transfer protein. In contrast, the influence of protein structure on the interaction of PtdlnsTPs with signalling enzymes is less clear. The crystal structures of PITPa and Secl4p were solved recently. Both proteins share no apparent similarity in their overall tertiary structures or their respective motifs that were postulated to be important for association with other proteins (Sha and Luo, 1999; Yoder et al., 2001). This dissimilarity gave the impression that protein structure does not play a significant part in PtdlnsTP’s role in reconstituting signalling.

Nevertheless, it remains possible that PtdlnsTP structure is involved in protein-protein interactions. Comparisons of PITPa, PITPp and RdgB-PITP sequences indicate that most non-conservative amino acid substitutions occur in the regulatory loop (Tanaka and Hosaka, 1994; Vihtelic et al., 1993), which has been postulated to mediate interactions

190 with other proteins, including lipid kinases (Yoder et a l, 2001). This may explain, at least in part, the inability of PITPa or PITPp to completely rescue mutant phenotypes of PITPa (Hamilton et a l, 1997), PITPp (Alb et a l, 2002) and RdgB (Milligan et a/., 1997). It will therefore be of great interest to investigate (for example, by using chimeric PITPs) whether the regulatory loop mediates specific protein-protein interactions and intracellular localizations.

The issue of whether PITP interacts directly with Pl-modifying enzymes remains unclear. Aikawa et al. (1999) demonstrated the in vivo association of MrdgBl with p230 type III PtdIns4K and deletion studies indicated that both the PITP domain and the adjacent acidic domain are required and sufficient for the association with p230. Hence, an unidentified protein may be required by PITP (and probably SecHp as well) in order to interact with the lipid kinases in the putative signalling complex. This protein may correspond to the ‘PITP receptor’ proposed by Hara et al. (1997). The ‘PITP receptor’ can apparently be blocked by C-terminally truncated PITPa, such that the signalling function of wildtype

PITPa is inhibited. These truncated PITPa molecules are able to exchange but not transfer Ptdlns or PtdCho, suggesting that the C-terminal region is required for PITP to dissociate from the membrane surface (Hara et al., 1997). This is consistent with the lipid exchange mechanism proposed by Yoder et al. (2001). By failing to dissociate from membranes, truncated PITPs may block access of wildtype PITPs to the putative receptor.

It can also be argued that a ‘PITP receptor’ is not necessary for PITP to associate with the PI signalling complex. Despite their lack of structural similarity, PITPa and SecHp have been proposed to share similar aspects in the mechanisms by which they mediate phospholipid exchange at the membrane surface - helix A10/T4 of SecHp or the C- terminal region of PITPa swings away from the rest of the protein to expose the bound lipid when it contacts the membrane surface (Sha et a l, 1998; Yoder et al., 2001). The latter is in agreement with the open apo-PITPa conformation, which is postulated to resemble the membrane-bound state of PITP (Schouten et al., 2002). After lipid exchange, the refolding and reassociation of helix A10/T4 or the C-terminal region will exert the driving force to dissociate SecHp or PITPa, respectively, from the membrane surface. It is tempting to speculate that (i) the exposed Ptdlns on PITP or SecHp, instead of being inserted into the membrane bilayer, is recognised and phosphorylated by PI kinases in the signalling complex and that (ii) loss of the C-terminal region deprives PITP of the driving force to dissociate from the signalling complex, thus blocking access of

191 wildtype PITP to the complex. This would also be consistent with the inference drawn earlier in this discussion that the type of phospholipid bound is unlikely to determine the interaction between PITP and the signalling complex.

Another influencing factor may be the formation of PITP oligomers. Analyses by gel filtration, lEF and gel electrophoresis suggest strongly that PITPp is able to organize as non-covalent homo-oligomers. This ability to oligomerise is apparently not dependent on the type of phospholipid bound or the expression host. Following gel filtration, all detectable lipid kinase activities are found to be associated with putative PITPp oligomers. In contrast, the gel filtration elution profile of lipid kinases that co-purified with PITPa is more diffuse. This may indicate that the affinity between PITPa and the signalling complex is comparatively weaker than that in PITPp, and/or that PITPa interacts with different types of signalling complexes, as one may expect to find downstream of different agonist receptors at the plasma membrane. One may even speculate that PITPa and PITPp present the Ptdlns substrate to the PI signalling complex via different mechanisms.

In addition to physical association between PITP and PI kinases, PPI synthesis may also be dependent on the proper localization of PITP within the cell. Aikawa et al. (1999) reported that the in vivo interaction of MrdgBl with type III PtdIns4K in COS-7 cells and the concomitant stimulation of Ptdlnsf synthesis are dependent on MrdgBl localising at the Golgi apparatus. MrdgBl deletion mutants that are ectopic in their subcellular locations with respect to a full-length MrdgBl were poor at inducing Ptdlnsf production using endogenous (MrdgBl-bound) Ptdlns as a substrate, even though the mutants can bind Ptdlns and interact with PtdIns4K. Hence, localization to its bona fide intracellular region may be important in determining the interaction of PITP with PI kinases as well as the viability of the PITP-bound lipid as a kinase substrate.

6.5 Concluding Remarks

In summary, the data presented in the preceding Chapters give a strong indication that PITP is an integral component of the PI signalling complex and that PITP-bound Ptdlns is a viable substrate for sequential enzymatic activity by the lipid kinases and lipases within the complex. Interactions between PITP and these enzymes were not dependent on the ability of PITP to bind Ptdlns, indicating that such interactions do not require prior recognition of the inositol headgroup by the enzymes and may be mediated, at least in

192 part, by a structural motif (such as the regulatory loop) of PITP. The data further suggest that the PI lipid remains bound to PITP following phosphorylation by PI kinases. Collectively, the data give strong support to the co-factor model of PITP function in PI signalling, as well as provide a mechanism to allow metabolic compartmentation in cellular membranes and an explanation to the empirical phenomenon of agonist- responsive and -unresponsive pools of Ptdlns (Monaco and Gershengom, 1992).

By acting as a co-factor in PI signalling, PITP also introduces a potential new point of regulation in substrate availability and accessibility. For example, the phosphorylation state of PITPs has been suggested to affect their lipid transfer capabilities and/or intracellular localizations, and perhaps even their interactions with other proteins (van Tiel et aL, 2000b; van Tiel et al., 2002; Yoder et a l, 2001). This, together with the notion of PITP-bound Ptdlns being the preferred substrate for PI kinases and lipases, will allow a more subtle control of substrate supply. Furthermore, if the resultant PPI remains bound to PITP as hypothesised, PITP may play an additional regulatory role in the competition between the various cellular pathways and processes that utilise the lipid.

The co-purifying lipid kinases include a PtdIns4K (that exhibits type II characteristics) and a PtdIns4P5K. Other laboratories have also reported PITPa associating with or stimulating class I and class III PI3Ks (Jones et a l, 1998; Kular et a l, 1997; Panaretou et a l, 1997) and PITPnm (a mouse MrdgBl homologue) associating with a type III PtdIns4K (Aikawa et al., 1999). In this regard, it will be of interest to determine whether interactions with these and perhaps other PI kinases also exhibit the same mechanisms described herein. If PITP-mediated presentation of Ptdlns proves to be a common characteristic, there may exist distinct metabolic PI pools at each intracellular membrane site where PITP performs a crucial regulatory role. It is therefore necessary to ascertain the specific isoforms of lipid kinases and PLC with which the various PITPs interact, as well as the subcellular locations (and compartmentation) of such interactions. A pertinent issue is whether (and how) each PITP isoform is also segregated to serve the various PI pools. New techniques that have emerged in recent years, such as fluorescence resonance energy transfer microscopy and fluorescence correlation spectroscopy, can help to further investigate the abovementioned issues as well as the dynamics of PITP-mediated PI compartmentation.

Pis and Pl-modifying enzymes have in recent years been implicated in many cellular processes, including signal transduction, vesicle trafficking, neurotransmission, regulation

193 of ion channel activity, cytoskeletal reorganization, cell proliferation and transformation, and apoptosis. Tight regulation is necessary to avoid abnormalities in their intracellular levels and/or activities which have been observed in various diseases such as carcinomas, neurological and immunological disorders, and inflammation. In the case of PITP, a reduction in PITPa protein levels can decrease the growth rate of WRK-1 tumour cells (Monaco et al., 1998) but leads to neurodegeneration in the vibrator mice (Hamilton et al., 1997), while PITPp deficiency results in early embryonic death in mice (Alb et al., 2002). As regulatory co-factors of PI signalling, PITPs and their interactions with other enzymes are thus potential targets for therapeutic intervention.

The data discussed in this thesis can provide insight into the crucial role that PITPs play in PI metabolism, and help to stimulate additional research.

194 References

Acharya, J.K., Labarca, P., Delgado, R., Jalink, K. and Zuker, C.S. (1998) Synaptic defects and compensatory regulation of inositol metabolism in inositol polyphosphate 1 -phosphatase mutants. Neuron, 20, 1219-1229.

Aikawa, Y., Hara, H. and Watanabe, T. (1997) Molecular cloning and characterization of mammalian homologues of the Drosophila retinal degeneration B gene. Biochem. Biophys. Res. Commun., 236, 559- 564.

Aikawa, ¥., Kuraoka, A., Kondo, H., Kawabuchi, M. and Watanabe, T. (1999) Involvement of PITPnm, a mammalian homologue of Drosophila rdgB, in phosphoinositide synthesis on Golgi membranes. J. Biol. Chem., 274, 20569-20577.

Aitken, J.F., van Heusden, G., Temkin, M. and Dowhan, W. (1990) The gene encoding the phosphatidylinositol transfer protein is essential for cell growth. J. Biol. Chem., 265, 4711-4717.

Akgoz, M., Azpiazu, I., Kalyanaraman, V. and Gautam, N. (2002) Role of the G protein y subunit in fiy complex modulation of phospholipase C-B function. J. Biol. Chem., 277, 19573-15978.

Alb, J.G., Phillips, S.E., Rostand, K., Cui, X., Pinxteren, J., Cotlin, L., Manning, T., Guo, S., York, J.D., Sontheimer, H., Collawn, J.F. and Bankaitis, V.A. (2002) Genetic ablation of phosphatidylinositol transfer protein function in murine embryonic stem cells. Mol. Biol. Cell, 13, 739-754.

Alb, J.G.J., Gedvilaite, A., Cartee, R.T., Skinner, H.B. and Bankaitis, V.A. (1995) Mutant rat phosphatidylinositol/phosphatidylcholine transfer proteins specifically defective in phosphatidylinositol transfer: implications for the regulation of phospholipid transfer activity. Proc. Natl. Acad. Soi. USA, 92, 8826-8830.

Allan, D. and Quinn, P. (1989) Membrane phospholipid asymmetry in Semliki Forest virus grown in BHK cells. Biochim. Biophys. Acta, 987, 199-204.

Allen, V., Swigart, P., Cheung, R., Cockcroft, S. and Katan, M. (1997) Regulation of inositol lipid-specific phospholipase CÔ by changes in Ca^+ ion concentrations. Biochem J., 327, 545-552.

Attree, O., Olivos, I.M., Okabe, I., Bailey, L.C., Nelson, D.L., Lewis, R.A., Mclnnes, R.R. and Nussbaum, R.L. (1992) The Lowe's oculocerebrorenal syndrome gene encodes a protein highly homologous to inositol polyphosphate-5-phosphatase. Nature, 358, 239-242.

Audhya, A., Foti, M. and Emr, S.D. (2000) Distinct roles for the yeast phosphatidylinositol 4-kinases, stt4p and piklp, in secretion, cell growth, and organelle membrane dynamics. Mol. Biol. Cell, 11, 2673-2689.

Auger, K.R., Serunian, L.A., Soltoff, S.P., Libby, P. and Cantley, L.C. (1989) PDGF-dependent tyrosine phosphorylation stimulates production of novel polyphosphoinositides in intact cells. Cell, 57, 167-175.

Azuma, T., Koths, K., Flanagan, L. and Kwiatkowski, D. (2000) Gelsolin in complex with phosphatidylinositol 4,5-bisphosphate inhibits caspase-3 and -9 to retard apoptotic progression. J. Biol. Chem., 275, 3761-3766.

Bae, Y.S., Cantley, L.G., Chen, C.-S., Kim, S.-R., Kwon, K.-S. and Rhee, S.G. (1998) Activation of phospholipase C-y by phosphatidylinositol 3,4,5-trisphosphate. J. Biol. Chem., 273, 4465-4469.

Bae, Y.S., Kang, S.W., Seo, M.S., Baines, I.C., Tekle, E., Chock, P.B. and Rhee, S.G. (1997) Epidermal growth factor-induced generation of hydrogen peroxide. J. Biol. Chem., 272, 217-221.

Balia, A., Tuymetova, G., Barshishat, M., Geiszt, M. and Balia, T. (2002) Characterization of type-ll phosphatidylinositol 4-kinase isoforms reveals association of the enzymes with endosomal vesicular compartments. J. Biol. Chem., 277, 20041-20050.

Balia, T., Downing, G.J., Jaffe, H., Kim, S., Zolyomi, A. and Catt, K.J. (1997) Isolation and molecular cloning of wortmannin-sensitive bovine type III phosphatidylinositol 4-kinases. J. Biol. Chem., 272, 18358- 18366.

Bankaitis, V.A., Aitken, J.R., Cleves, A.E. and Dowhan, W. (1990) An essential role for a phospholipid transfer protein in yeast Golgi function. Nature, 347, 561-562. 195 Bankaitis, V.A., Malehom, D.E., Emr, S.D. and Greene, R. (1989) The Saccharomyces cerevisiae SEC 14 gene encodes a cytosolic factor that is required for transport of secretory proteins from the yeast Golgi complex. J. Cell Biol, 108, 1271-1281.

Barbieri, M.A., Heath, C.M., Peters, E.M., Wells, A., Davis, J.N. and Stahl, P.D. (2001) Phosphatidylinositol-4-phosphate 5-kinase-ip is essential for epidermal growth factor receptor-mediated endocytosis. J. Biol Chem., 276, 47212-47216.

Barnard, J.A., Graves, D.R., Pittelkow, M.R., DuBois, R., Cook, P., Ramsey, G.W., Bishop, PR., Damstrup, L. and Coffey, R.J. (1994) Auto- and cross-induction within the mammalian epidermal growth factor-related peptide family. J. Biol Chem., 269, 22817-22822.

Baron, C.L. and Malhotra, V. (2002) Role of diacylglycerol in PKD recruitment to the TGN and protein transport to the plasma membrane. Science, 295, 325-328.

Bar-Sagi, D., Rotin, D., Batzer, A., Mandiyan, V. and Schlessinger, J. (1993) SH3 domains direct cellular localization of signaling molecules. Cell, 74, 83-91.

Barylko, B., Gerber, S.H., Binns, D.D., Grichine, N., Khvotchev, M., Sudhof, T.C. and Albanesi, J.P. (2001) A novel family of phosphatidylinositol 4-kinases conserved from yeast to humans. J. Biol Chem., 276, 7705-7708.

Batty, I.H., Currie, R.A. and Downes, C.P. (1998) Evidence for a model of integrated inositol phospholipid pools implies an essential role for lipid transport in the maintenance of receptor-mediated phospholipase C activity in 132INI cells. Biochem. J., 330, 1069-1077.

Bazenet, C.E., Ruano, A.R., Brockman, J.L. and Anderson, R.A. (1990) The human erythrocyte contains two forms of phosphatidylinositol-4-phosphate 5-kinase which are differentially active toward membranes. J. Biol Chem., 265, 18012-18022.

Berridge, M.J. (1993) Inositol trisphosphate and calcium signalling.Nature, 361, 315-325.

Berstein, G., Blank, J.L., Jhon, D.Y., Exton, J.H., Rhee, S.G. and Ross, E.M. (1992) Phospholipase C-beta 1 is a GTPase-activating protein for Gq/11, its physiologic regulator. Cell, 70, 411-418.

Bertaccini, E. and Trudell, J R. (2001) Molecular modeling of ligand-gated ion channels: progress and challenges. Int. Rev. Neurobiol, 48, 141-166.

Billah, M.M., Lapetina, E.G. and Cuatrecasas, P. (1981) Phospholipase Aj activity specific for phosphatidic acid. J. Biol Chem., 256, 5399-5403.

Borg, J.B., Ooi, J., Levy, E. and Margolis, B. (1996) The phosphotyrosine interaction domains of X I1 and FE65 bind to distinct sites on the YENPTY motif of amyloid precursor protein. Mol Cell. Biol, 16, 6229- 6241.

Boronenkov, I.V. and Anderson, R.A. (1995) The sequence of phosphatidylinositol-4-phosphate 5-kinase defines a novel family of lipid kinases. J. Biol. Chem., 270, 2881-2884.

Boronenkov, I.V., Loijens, J.C., Umeda, M. and Anderson, R.A. (1998) Phosphoinositide signaling pathways in nuclei are associated with nuclear speckles containing pre-mRNA processing factors. Mol Biol Cell, 9, 3547-3560.

Bradford, M.M. (1976) A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal Biochem., 72, 248-254.

Brill, J.A., Hime, G.R., Scharer-Schuksz, M. and Fuller, M.T. (2000) A phospholipid kinase regulates actin organization and intercellular bridge formation during germline cytokinesis. Development, 127, 3855-3864.

Brown, H.A., Gutowski, S., Moomaw, C.R., Slaughter, C. and Stemweis, P.C. (1993) ADP-ribosylation factor, a small GTP-dependent regulatory protein, stimulates phospholipase D activity. Cell, 75, 1137-1144.

Brown, M.T., Andrade, J., Radhakrishna, H., Donaldson, J.G., Cooper, J.A. and Randazzo, P.A. (1998) ASAPl, a phospholipid-dependent Arf GTPase-activating protein that associates with and is phosphorylated by Src. Mol Cell Biol, 18, 7038-7051.

196 Buck, E., Li, J., Chen, Y., Weng, G., Scarlata, S. and Iyengar, R. (1999) Resolution of a signal transfer region from a general binding domain in Op for stimulation of phospholipase C-P2, Science, 283, 1332- 1335.

Buckingham, L. and Duncan, J.L. (1983) Approximate dimensions of membrane lesions produced by streptolysin S and streptolysin O. Biochim. Biophys. Acta, 729, 115-122.

Burack, W.R. and Shaw, A.S. (2000) Signal transduction; hanging on a scaffold.Curr. Opin. Cell Biol, 12, 211-216.

Burd, C.G. and Emr, S.D. (1998) Phosphatidylinositol (3)-phosphate signaling mediated by specific binding to RING FYVE domains. Mol Cell, 2, 157-162.

Burnette, W.N. (1981) "Western blotting": electrophoretic transfer of proteins from sodium dodecyl sulfate- -polyacrylamide gels to unmodified nitrocellulose and radiographic detection with antibody and radioiodinated protein A. Anal Biochem., 112, 195-203.

Cantley, L.C., Fruman, D.A. and Meyers, R.E. (1998) Phosphoinositide kinases. Annu. Rev. Biochem., 67, in press.

Carpenter, G. (1999) Employment of the epidermal growth factor receptor in growth factor-independent signaling pathways. J. Cell Biol, 146, 697-702.

Carpenter, G., Hernandez, S.S., Nishibe, S., Todderud, G., Mumby, M. and Wahl, M. (1992) Growth factor phosphorylation of PLC-yl. Ciba Found. Symp., 164, 223-233.

Carpenter, G. and Ji, Q. (1999) Phospholipase C-y as a signal-transducing element.Exp. Cell Res., 253, 15- 24.

Carter, A.N. and Downes, C.P. (1992) Phosphatidylinositol 3-kinase is activated by nerve growth factor and epidermal growth factor in PC 12 cells. J. Biol. Chem., 267, 14563-14567.

Castellino, A.M., Parker, G.J., Boronenkov, I.V., Anderson, R.A. and Chao, M.V. (1997) A novel interaction between the juxtamembrane region of the p55 tumor necrosis factor receptor and phosphatidylinositol-4-phosphate 5-kinase. J. Biol. Chem., 272, 5861-5870.

Chan, D.C., Bedford, M.T. and Leder, P. (1996) Formin binding proteins bear WWP/WW domains that bind proline-rich peptides and functionally resemble SH3 domains. EMBOJ., 15, 1045-1054.

Chang, J.T., Milligan, S., Li, Y., Chew, C.E., Wiggs, J., Copeland, N.G., Jenkins, N.A., Campochiaro, P.A., Hyde, D R. and Zack, D.J. (1997) Mammalian homolog of Drosophila retinal degeneration B rescues the mutant fly phenotype. J. Neuroscl, 17, 5881-5890.

Chatah, N.-E.-.H. and Abrams, C.S. (2001) G-protein-coupled receptor activation induces the membrane translocation and activation of phosphatidylinositol-4-phosphate 5-kinase la by a Rac- and Rho-dependent pathway. J. Biol Chem., 276, 34059-34065.

Chattopadhyay, A., Vecchi, M., Ji, Q.-S., Memaugh, R. and Carpenter, G. (1999) The role of individual SH2 domains in mediating association of phospholipase C-yl with the activated EGF receptor. J. Biol Chem., 274, 26091-26097.

Chen, D., Waters, S.B., Holt, K.H. and Pessin, J.E. (1996) SOS phosphorylation and disassociation of the Grb2-S0S complex by the ERK and JNK signaling pathways. J. Biol Chem., 271, 6328-6332.

Chen, S.W., Lazar, S.C., Poenie, M., Tsien, R.Y., Gill, G. and Rosenfeld, M.G. (1987) Requirement for intrinsic protein tyrosine kinase in the immediate and late actions of the EGF receptor. Nature, 328, 820- 823.

Cheng, H.F., Jiang, M.J., Chen, C.L., Liu, S.M., Wong, L.P., Lomasney, J.W. and King, K. (1995) Cloning and identification of amino acid residues of human phospholipase C 5 1 essential for catalysis. J. Biol Chem., 270, 5495-5505.

Cho, W. (2001) Membrane targeting by Cl and C2 domains. J. Biol Chem., 276, 32407-32410.

Choi, K.Y., Satterberg, B., Lyons, D.M. and Elion, E.A. (1994) Ste5 tethers multiple protein kinases in the MAP kinase cascade required for mating in S. cerevisiae. Cell, 78, 499-512.

197 Chong, L.D., Traynor, K.A., Bokoch, G.M. and Schwartz, M.A. (1994) The small GTP-binding protein Rho regulates a phosphatidylinositol 4-phosphate 5-kinase in mammalian cells. Cell, 79, 507-513.

Chou, J., Stolz, D.B., Burke, N.A., Watkins, S.C. and Wells, A. (2002) Distribution of gelsolin and phosphoinositol 4,5-bisphosphate in lamellipodia during EGF-induced motility. Int. J. Biochem. Cell Biol, 34, 776-790.

Cifuentes, M.E., Honkanen, L. and Rebecchi, M.J. (1993) Proteolytic fragments of phosphoinositide- specific phospholipase C-delta 1. Catalytic and membrane binding properties. J. Biol Chem., 268, 11586- 11593.

Ciruela, A., Hinchliffe, K.A., Divecha, N. and Irvine, R.F. (2000) Nuclear targeting of the P isoform of type II phosphatidylinositol phosphate kinase (phosphatidylinositol 5-phosphate 4-kinase) by its a-helix 7. Biochem. J., 346, 587-591.

Clarke, N.G. and Dawson, R.M.C. (1981) Alkaline O leads to N-transacylation. A new method for the quantitative déacylation of phospholipids. Biochem. J., 195, 301-306.

Cocco, L., Martelli, A.M., Bamabei, O. and Manzoli, F.A. (2001) Nuclear inositol lipid signaling.Adv. Enzyme Regul, 41, 361-384.

Cocco, L., Martelli, A.M., Mazzotti, G., Bamabei, O. and Manzoli, F.A. (2000) Inositides and the nucleus: phospholipase CP family localization and signaling activity. Adv. Enzyme Regul, 40, 83-95.

Cochet, C., Filhol, O., Payrastre, B., Hunter, T. and Gill, G.N. (1991) Interaction between the epidermal growth factor receptor and phosphoinositide kinases. J. Biol Chem., 266, 637-644.

Cockcroft, S., Thomas, G.M., Fensome, A., Geny, B., Cunningham, E., Gout, I., Hiles, I., Totty, N.F., Truong, O. and Hsuan, J.J. (1994a) Phospholipase D: a downstream effector of ARF in granulocytes. Science, 263, 523-526.

Cockcroft, S., Thomas, G.M.H., Cunningham, E. and Ball, A. (1994b) Use of cytosol-depleted HL-60 cells for reconstitution studies of G-protein-regulated phosphoinositide-specific phospholipase C-P isozymes. Methods Enzymol, 238, 154-168.

Cockcroft, V.B., Osguthorpe, D.J., Barnard, E.A., Friday, A.E. and Lunt, G.G. (1990) Ligand-gated ion channels. Homology and diversity. Mol Neurobiol, 4, 129-169.

Constantinescu, S.N., Keren, T., Socolovsky, M., Nam, H., Henis, Y.I. and Lodish, H.F. (2001) Ligand- independent oligomerization of cell-surface erythropoietin receptor is mediated by the transmembrane domain. Proc. Natl Acad. Scl USA, 98, 4379-4384.

Cooke, F.T., Dove, S.K., McEwen, R.K., Painter, G., Holmes, A.B., Hall, M.N., Michell, R.H. and Parker, P.J. (1998) The stress-activated phosphatidylinositol 3-phosphate 5-kinase Fab Ip is essential for vacuole fimction in S. cerevisiae. Curr. Biol, 8, 1219-1222.

Cunningham, E., Tan, S.K., Swigart, P., Hsuan, J., Bankaitis, V. and Cockcroft, S. (1996) The yeast and mammalian isoforms of phosphatidylinositol transfer protein can all restore phospholipase C-mediated inositol lipid signaling in cytosol-depleted RBL-2H3 and HL-60 cells. Proc. Natl Acad. Scl USA, 93, 6589-6593.

Cunningham, E., Thomas, G.M., Ball, A., Hiles, I. and Cockcroft, S. (1995) Phosphatidylinositol transfer protein dictates the rate of inositol trisphosphate production by promoting the synthesis of PIP 2 . Curr. Biol, 5, 775-783.

Cunningham, E.M. (1994) Identification of phosphatidylinositol transfer protein as an essential requirement for inositol lipid signalling in HL60 cells.. University of London.

Currie, R.A., MacLead, B.M.G. and Downes, C.P. (1997) The lipid transfer activity of phosphatidylinositol transfer protein is sufficient to account for enhanced phospholipase C activity in turkey erythrocyte ghosts. Curr. Biol, 7, 184-190.

Dasgupta, J.D., Granja, C., Druker, B., Lin, L.L., Yunis, E.J. and Relias, V. (1992) Phospholipase C-yl association with CD3 structure in T cells. J. Exp. Med, 175, 285-288.

198 Daub, H., Weiss, F.U., Wallasch, C. and Ullrich, A. (1996) Role of transactivation of the EGF receptor in signalling by G-protein-coupled receptors. Nature, 379, 557-560.

de Camilli, P., Chen, H., Hyman, J., Panepucci, E., Bateman, A. and Brunger, A.T. (2002) The ENTH domain. FEES Lett, 513, 11-18.

de Camilli, P., Emr, S.D., McPherson, P.S. and Novick, P. (1996) Phosphoinositides as regulators in membrane traffic. Science, 271, 1533-1539.

de Graaf, P., Klapisz, E.E., Schulz, T.K., Cremers, A.F., Verkleij, A.J. and van Bergen En Henegouwen, P.M. (2002) Nuclear localization of phosphatidylinositol 4-kinase (3. J. Cell Set., 115, 1769-1775.

de Neef, R.S., Hardy-Dessources, M.D. and Giraud, F. (1996) Relationship between type II phosphatidylinositol 4-kinase activity and protein tyrosine phosphorylation in membranes derived from normal and sickle red cells. Eur. J. Biochem., 235, 549-556.

de Vries, K.J., Heinrichs, A.A., Cunningham, E,, Brunink, F., Westerman, J., Somerharju, P.J., Cockcroft, S., Wirtz, K.W. and Snoek, G.T. (1995) An isoform of the phosphatidylinositol-transfer protein transfers sphingomyelin and is associated with the Golgi system. Biochem. J., 310, 643-649.

de Vries, K.J., Westerman, J., Bastiaens, P.I.H., Jovin, T.M., Wirtz, K.W.A. and Snoek, G.T. (1996) Fluorescently labeled phosphatidylinositol transfer protein isoforms (a and P), microinjected into fetal bovine heart endothelial cells, are targeted to distinct intracellular sites.Exp. Cell Res., 227, 33-39.

Desrivieres, S., Cooke, F.T., Parker, P.J. and Hall, M.N. (1998) MSS4, a phosphatidylinositol-4-phosphate 5-kinase required for organisation of the actin cytoskeleton in Saccharomyces cerevisiae. J. Biol. Chem., 273, 15787-15793.

Dickeson, S.K., Helmkamp, G.J. and Yarbrough, L.R. (1994) Sequence of a human cDNA encoding phosphatidylinositol transfer protein and occurrence of a related sequence in widely divergent eukaryotes. Gene, 142, 301-305.

Dickeson, S.K., Lim, C.N., Schuyler, G.T., Dalton, T.P., Helmkamp, G.M.J. and Yarbrough, L.R. (1989) Isolation and sequence of cDNA clones encoding rat phosphatidylinositol transfer protein. J. Biol. Chem., 264, 16557-16564.

DiCorleto, P.E., Warach, J.B. and Zilversmit, D.B. (1979) Purification and characterization of two phospholipid exchange proteins from bovine heart. J. Biol. Chem., 254, 7795-7802.

Divecha, N., Brooksbank, C.E. and Irvine, R.F. (1992) Purification and characterization of phosphatidylinositol 4-phosphate 5-kinases. Biochem. J.

Divecha, N., Roefs, M., Halstead, J R., D'Andrea, S., Femandez-Borga, M., Oomen, L., Saqib, K.M., Wakelam, M.J.O. and D'Santos, C. (2000) Interaction of the type la PIPkinase with phospholipase D: a role for the local generation of phosphatidylinositol 4,5-bisphosphate in the regulation of PLD2 activity. EMBO J., 19, 5440-5449.

Divecha, N., Truong, O., Hsuan, J.J., Hinchliffe, K.A. and Irvine, R.F. (1995) The cloning and sequence of the C isoform of PtdIns4P 5-kinase. Biochem. J., 309, 715-719.

Donaldson, J.G. and Klausner, R.D. (1994) ARF: a key regulatory switch in membrane traffic and organelle structure. Curr. Opin. Cell Biol, 6, 527-532.

Dove, S.K., Cooke, F.T., Douglas, M.R., Sayers, L.G., Parker, P.J. and Michell, R.H. (1997) Osmotic stress activates phosphatidylinositol-3,5-bisphosphate sythesis. Nature, 390, 187-192.

Downes, C.P. and Batty, I.H. (1993) Inositol lipid signalling extacts a transfer fee. Curr. Biol, 3, 794-796.

Downes, C.P. and Michell, R.H. (1985) Inositol lipid breakdown as a receptor-controlled generator of second messengers. In Cohen, P. and Houslay, M.D. (eds.). Molecular mechanisms of transmembrane signalling. Elsevier, Amsterdam, pp. 3-56.

Downing, G.J., Kim, S., Nakanishi, S., Catt, K.J. and Balia, T. (1996) Characterization of a soluble adrenal phosphatidylinositol 4-kinase reveals wortmannin sensitivity of type III phosphatidylinositol kinases. Biochemistry, 35, 3587-3594.

199 Downward, J., Parker, P. and Waterfield, M.D. (1984) Autophosphorylation sites on the epidermal growth factor receptor. Nature, 311, 483-485.

D'Santos, C.S., Clarke, J.H., Irvine, R.F. and Divecha, N. (1999) Nuclei contain two differentially regulated pools of diacylglycerol. Curr. Biol, 9,437-440.

Eberhard, D.A., Cooper, C.L., Low, M.G. and Holz, R.W. (1990) Evidence that the inositol phospholipids are necessary for exocytosis. Loss of inositol phospholipids and inhibition of secretion in permeabilized cells caused by a bacterial phospholipase C and removal of ATP. Biochem. J., 268, 15-25.

Eck, M.J., Dhe Paganon, S., Trub, T., Nolte, R.T. and Shoelson, S.E. (1996) Structure of the IRS-1 PTB domain bound to the juxtamembrane region of the insulin receptor. Cell, 85, 695-705.

Eguchi, S., Numaguchi, K., Iwasaki, H., Matsumoto, T., Yamakawa, T., Utsunomiya, H., Motley, E.D., Owada, K.M., Hirata, Y., Marumo, F. and Inagami, T. (1998) Calcium-dependent epidermal growth factor receptor transactivation mediates the angiotensin Il-induced mitogen-activated protein kinase activation in vascular smooth muscle cells. J. Biol Chem., 273, 8890-8896.

Eichholtz, T., Jalink, K., Fahrenfort, I. and Moolenaar, W.H. (1993) The bioactive phospholipid lysophosphatidic acid is released from activated platelets. Biochem. J., 291, 677-680.

Elenius, K., Paul, S., Allison, G., Sun, J. and Klagsbrun, M. (1997) Activation of HER4 by heparin-binding EGF-like growth factor stimulates chemotaxis but not proliferation. EMBOJ., 16, 1268-1278.

Ellis, R.J. (1998) Steric chaperones. Trends Biochem. Sci., 23,43-45.

Endemann, G., Dunn, S.N. and Cantley, L.C. (1987) Bovine brain contains two types of phosphatidylinositol kinase. Biochemistry, 26, 6845-6852.

Endemann, G.C., Graziani, A. and Cantley, L.C. (1991) A monoclonal antibody distinguishes two types of phosphatidylinositol 4-kinase. Biochem. J., 273, 63-66.

Essen, L.-O., Perisic, O., Cheung, R., Katan, M. and Williams, R.L. (1996) Crystal structure of a mammalian phosphoinositide-specific phospholipase CÔ. Nature, 380, 595-602.

Evellin, S., Nolte, J., Tysack, K., vom Dorp, F., Thiel, M., Oude Weemink, P.A., Jakobs, K.H., Webb, E., Lomasney, J.W. and Schmidt, M. (2002) Stimulation of phospholipase C-e by the M3 muscarinic acetylcholine receptor mediated by cyclic AMP and the GTPase Rap2B. J. Biol Chem., 277, 16805-16813.

Exton, J.H. (1997) New developments in phospholipase D. J. Biol Chem., 272, 15579-15582.

Fain, J.N. and Berridge, M.J. (1979) Relationship between phosphatidylinositol synthesis and recovery of 5-hydroxytryptamine-responsive Ca^'*’ flux in blowfly salivary glands. Biochem. J., 180, 655-661.

Falasca, M., Logan, S.K., Lehto, V.P., Baccante, G., Lemmon, M.A. and Schlessinger, J. (1998) Activation of phospholipase Cy by PI 3-kinase-induced PH domain-mediated membrane targeting. EMBOJ., 17, 414- 422.

Fantl, W.J., Escobedo, J.A., Martin, G.A., Turck, C.W., Del Rosario, M., McCormick, F. and Williams, L.T. (1992) Distinct phosphotyrosines on a growth factor receptor bind to specific molecules that mediate different signaling pathways. Cell, 69,413-423.

Feng, S., Chen, J.K., Yu, H., Simon, J.A. and Schreiber, S.L. (1994) Two binding orientations for peptides to the Src SH3 domain: development of a general model for SH3-ligand interactions. Science, 266, 1241- 1247.

Fensome, A., Cunningham, E., Prosser, S., Tan, S.K., Swigart, P., Thomas, G., Hsuan, J. and Cockcroft, S. (1996) ARF and PITP restore GTPyS-stimulated protein secretion from cytosol-depleted HL60 cells by promoting PIP2 synthesis. Curr. Biol, 6, 730-738.

Flanagan, C.A., Schnieders, E.A., Emerick, A.W., Kunisawa, R., Admon, A. and Thomer, J. (1993) Phosphatidylinositol 4-kinase: gene structure and requirement for yeast cell viability. Science, 262, 1444- 1448.

Flanagan, C.A. and Thomer, J. (1992) Purification and characterization of a soluble phosphatidylinositol 4- kinase from the yeast Saccharomyces cerevisiae. J. Biol. Chem., 267, 24117-24125.

200 Flick, J.S. and Thomer, J. (1993) Genetic and biochemical characterization of a phosphatidylinositol- specific phospholipase C in Saccharomyces cerevisiae. Mol. Cell. Biol, 13, 5861-5876.

Fogg, V.C., Azpiazu, I., Linder, M.E., Smrcka, A., Scarlata, S. and Gautam, N, (2001) Role of the y subunit prenyl moiety in G protein (3y complex interaction with phospholipase Cp. J. Biol Chem., 276, 41797- 41802.

Ford, M.G., Pearse, B.M., Higgins, M.K., Vallis, Y., Owen, D.J., Gibson, A., Hopkins, C.R., Evans, P R. and McMahon, H.T. (2001) Simultaneous binding of PtdIns(4,5)P2 and clathrin by API80 in the nucléation of clathrin lattices on membranes. Science, 291, 1051-1055.

Fruman, D.A., Meyers, RE. and Cantley, L.C. (1998) Phosphoinositide kinases. Annu. Rev. Biochem., 67, 481-507.

Fukami, K., Furuhashi, K., Inagaki, M., Endo, T., Hatano, S. and Takenawa, T. (1992) Requirement of phosphatidylinositol 4,5-bisphosphate for a-actinin function. Nature, 359, 150-152.

Fukami, K., Nakao, K., Inoue, T., Kataoka, Y., Kurokawa, M., Fissore, R.A., Nakamura, K., Katsuki, M., Mikoshiba, K., Yoshida, N. and Takenawa, T. (2001) Requirement of phospholipase CÔ4 for the zona pellucida-induced acrosome reaction. Science, 292, 920-923.

Fukami, K. and Takenawa, T. (1992) Phosphatidic acid that accumulates in platelet-derived growth factor- stimulated Balb/c 3T3 cells is a potential mitogenic signal.J. Biol Chem., 267, 10988-10993.

Fullwood, Y., dos Santos, M. and Hsuan, J.J. (1999) Cloning and characterisation of a novel human phosphatidylinositol transfer protein, rdgp. J. Biol Chem., 274, 31553-31558.

Funaki, M., Katagiri, H., Kanda, A., Anai, M., Nawano, M., Ogihara, T., Inukai, K., Fukushima, Y., Ono, H., Yazaki, Y., Kikuchi, M., Oka, Y. and Asano, T. (1999) p85/pl 10-type phosphatidylinositol kinase phosphorylates not only the D-3, but also the D-4 position of the inositol ring. J. Biol Chem., 274, 22019- 22024.

Gabev, E., Kasianowicz, J., Abbott, T. and McLaughlin, S. (1989) Binding of neomycin to phosphatidylinositol 4,5-bisphosphate. Biochim. Biophys. Acta, 979, 105-112.

Gaits, F., Fourcade, O., Le Balle, F., Gueguen, G., Gaige, B., Gassama-Diagne, A., Fauvel, J., Salles, J.P., Mauco, G., Simon, M.F. and Chap, H. (1997) Lysophosphatidic acid as a phospholipid mediator: pathways of synthesis. FEBS Lett., 410, 54-58.

Garcia, P., Gupta, R., Shah, S., Morris, A.J., Rudge, S.A., Scarlata, S., Petrova, V., McLaughlin, S. and Rebecchi, M.J. (1995) The pleckstrin homology domain of phospholipase C-Ô1 binds with high affinity to phosphatidylinositol 4,5-bisphosphate in bilayer membranes. Biochemistry, 34, 16228-16234.

Garcia-Bustos, J.F., Marini, F., Stevenson, I., Frei, C. and Hall, M.N. (1994) PIKl, an essential phosphatidylinositol 4-kinase associated with the yeast nucleus. EMBOJ., 13, 2352-2361.

Gary, J.D., Wurmser, A.E., Bonangelino, C.J., Weisman, L.S. and Emr, S.D. (1998) Fab Ip is essential for PtdIns(3)P 5-kinase activity and the maintenance of vacuolar size and membrane homeostasis. J. Cell Biol, 143, 65-79.

Gaullier, J.-M., Simonsen, A., D'Arrigo, A., Bremnes, B., Stenmark, H. and Aasland, R. (1998) FYVE fingers bind PtdIns(3)P. Nature, 394, 432-433.

Gehrmann, T., Vereb, G., Schmidt, M., Klix, D., Meyer, H.E., Varsânyi, M. and Heilmeyer, L.M.G.j. (1996) Identification of a 200K polypeptide as type 3 phosphatidylinositol 4-kinase from bovine brain by partial protein and cDNA sequencing. Biochim. Biophys. Acta, 1311, 53-63.

Geijtenbeek, T.B.H., de Groot, E., van Baal, J., Brunink, F., Westerman, J., Snoek, G.T. and Wirtz, K.W.A. (1994) Characterization of mouse phosphatidylinositol transfer protein expressed in Escherichia coli. Biochim. Biophys. Acta, 1213, 309-318.

Gillooly, D.J., Morrow, I.C., Lindsay, M., Gould, R, Bryant, N.J., Gaullier, J.M., Parton, R.G. and Stenmark, H. (2000) Localization of phosphatidylinositol 3-phosphate in yeast and mammalian cells. EMBOJ, 19, 4577-4588.

201 Gilmore, A.P. and Burridge, K. (1996) Regulation of vinculin binding to talin and actin by phosphatidyl- inositol-4-5-bisphosphate. Nature, 381, 531-535.

Godi, A., Fertile, P., Meyers, R., Marra, P., Di Tullio, G., lurisci, C., Luini, A., Corda, D. and De Matteis, M.A, (1999) ARF mediates recruitment of PtdIns-4-OH kinase p and stimulates synthesis of PtdIns(4,5)P2 on the Golgi complex. Nat. Cell Biol, 1, 280-287.

Godi, A., Santone, I., Fertile, P., Devarajan, P., Stabach, P.R., Morrow, J.S., Di Tullio, G., Polishchuk, R., Petrucci, T.C., Luini, A. and De Matteis, M.A. (1998) ADP ribosylation factor regulates spectrin binding to the Golgi complex. Proc. Natl Acad. Sol USA, 95, 8607-8612.

Goldschmidt-Clermont, P.J., Kim, J.W., Machesky, L.M., Rhee, S.G. and Pollard, T.D. (1991) Regulation of phospholipase C-yl by profilin and tyrosine phosphorylation. Science, 251, 1231-1233.

Goldschmidt-Clermont, P.J., Machesky, L.M., Baldassare, J.J. and Pollard, T.D. (1990) The actin-binding protein profilin binds to PIPj and inhibits its hydrolysis by phospholipase C. Science, 247, 1575-1578.

Gomperts, B.D. and Tatham, P.E.R. (1992) Regulated exocytotic secretion from permeabilized cells. Methods Enzymol, 219, 178-189.

Gout, I., Dhand, R., Hiles, I.D., Fry, M.J., Panayotou, G., Das, P., Truong, O., Totty, N.F., Hsuan, J., Booker, G.W., Campbell, I D. and Waterfield, M.D. (1993) The GTPase dynamin binds to and is activated by a subset of SH3 domains. Cell, 75, 25-36.

Granja, C., Lin, L.L., Yunis, E.J., Relias, V. and Dasgupta, J.D. (1991) PLCyl, a possible mediator of T cell receptor function. J Biol Chem., 266, 16277-16280.

Greenfield, C., Hiles, I., Waterfield, M.D., Federwisch, M., Wollmer, A., Blundell, T.L. and McDonald, N. (1989) Epidermal growth factor binding induces a conformational change in the external domain of its receptor. EMBO V., 8, 4115-4123.

Gu, M., Warshawsky, I. and Majerus, P.W. (1992) Cloning and expression of megakaryocyte protein- tyrosine-phosphatase with sequence homology to retinaldehyde binding protein and yeast SEC14p. Proc. Natl Acad. Scl USA, 88, 2980-2984.

Guo, J. and Yu, F.X. (1997) Cloning and characterization of human homologue ofDrosophila retinal degeneration B: a candidate gene for degenerative retinal diseases.Dev. Genet., 20, 235-245.

Guo, K., Nichol, R., Skehel, P., Dormann, D., Weijer, C.J., Williams, J.G. and Pears, C. (2001) A Dictyostelium nuclear phosphatidylinositol phosphate kinase required for developmental gene expression. EMBOJ., 20, 6017-6027.

Guo, S., Stolz, L.E., Lemrow, S.M. and York, J.D. (1999) SAC 1-like domains of SACl, INP52, and INP53 and of human synaptojanin encode polyphosphoinositide phosphatases. J. Biol Chem., 274, 12990-12995.

Halstead, J.R., Roefs, M., Ellson, C.D., D'Andrea, S., Chen, C.-S., D'Santos, C.S. and Divecha, N. (2001) A novel pathway of cellular phosphatidylinositol(3,4,5)-trisphosphate synthesis is regulated by oxidative stress. Curr. Biol, 11, 386-395.

Hama, H., Schnieders, E.A., Thomer, J., Takemoto, J.Y. and DeWald, D.B. (1999) Direct involvement of phosphatidylinositol 4-phosphate in secretion in the yeast Saccharomyces cerevisiae. J. Biol. Chem., 274, 34294-34300.

Hamilton, B.A., Smith, D.J., Mueller, K.L., Kerrebrock, A.W., Bronson, R.T., van Berkel, V., Daly, M.J., Kruglyak, L., Reeve, M.P., Nemhauser, J.L., Hawkins, T.L., Rubin, E.M. and Lander, E.S. (1997) The vibrator mutation causes neurodegeneration via reduced expression of PlTPa: positional complementation cloning and extragenic suppression. Neuron, 18, 711-722.

Hamm, H.E. and Gilchrist, A. (1996) Heterotrimeric G proteins. Curr. Opin. Cell Biol, 8, 189-196.

Hara, S., Swigart, P., Jones, D. and Cockcroft, S. (1997) The first 5 amino acids of the carboxy terminus of phosphatidylinositol transfer protein (PITP) a play a critical role in inositol lipid signaling. J. Biol Chem., 272, 14908-14913.

Harlan, J.E., Hajduk, P.J., Yoon, H.S. and Fesik, S.W. (1994) Pleckstrin homology domains bind to phosphatidylinositol-4,5-bisphosphate. Nature, 371, 168-170. 202 Hashimoto, A., Takeda, K., Inaba, M., Sekimata, M., Kaisho, T., Ikehara, S., Homma, Y., Akira, S. and Kurosaki, T. (2000) Cutting edge: essential role of phospholipase C-y2 in B cell development and function. J. Immunol, 165, 1738-1742.

Haugh, J.M., Schooler, K., Wells, A., Wiley, H.S. and Lauffenburger, D.A. (1999) Effect of epidermal growth facor receptor internalization on regulation of the phospholipase C-yl signaling pathway. J. Biol Chem., 274, 8958-8965.

Hay, J.C., Fisette, P.L., Jenkins, G.H., Fukami, K., Takenawa, T., Anderson, R.A. and Martin, T.F. (1995) ATP-dependent inositide phosphorylation required for Ca^+-activated secretion. Nature, 374, 173-177.

Hay, J.C. and Martin, T.F. (1993) Phosphatidylinositol transfer protein required for ATP-dependent priming of Ca^^-activated secretion. Nature, 366, 572-575.

Heath, J.K., White, S.J., Johnstone, C.N., Catimel, B., Simpson, R.J., Moritz, R.L., Tu, G.F., Ji, H., Whitehead, R.H., Groenen, L.C., Scott, A.M., Ritter, G., Cohen, L., Welt, S., Old, L.J., Nice, B.C. and Burgess, A.W. (1997) The human A3 3 antigen is a transmembrane glycoprotein and a novel member of the immunoglobulin superfamily. Proc. Natl Acad. Scl USA, 94,469-474.

Helin, K. and Beguinot, L. (1991) Internalization and down-regulation of the human epidermal growth factor receptor are regulated by the carboxyl-terminal tyrosines. J. Biol Chem., 266, 8363-8368.

Helmkamp, G.M. (1980a) Effects of phospholipid fatty acid composition and membrane fluidity on the activity of bovine brain phospholipid exchange protein. Biochemistry, 19, 2050-2056.

Helmkamp, G.M. (1980b) Interaction of bovine brain phospholipid exchange protein with liposomes of different lipid composition. Biochim. Biophys. Acta, 595, 222-234.

Hendricks, K.B., Wang, B.Q., Schnieders, E.A. and Thomer, J. (1999) Yeast homologue of neuronal ffequenin is a regulator of phosphatidylinositol-4-OH kinase. Nat. Cell Biol, 1, 234-241.

Hermosura, M.C., Takeuchi, H., Fleig, A., Riley, A.M., Potter, B.V., Hirata, M. and Penner, R. (2000) InsP^ facilitates store-operated calcium influx by inhibition of InsPg 5-phosphatase. Nature, 408, 735-740.

Hildebrandt, J.P., Plant, T.D. and Meves, H. (1997) The effects of bradykinin on currents in NG108-15 cells treated with U73122, a phospholipase C inhibitor, or neomycin. Br. J. Pharmacol, 120, 841-850.

Hinchliffe, K.A., Irvine, R.F. and Divecha, N. (1996) Aggregation-dependent, integrin-mediated increases in cytoskeletally associated PtdIns(4,5)P2 levels in human platelets are controlled by translocation of Ptdlns 4-P 5-kinase C to the cytoskeleton. EMBOJ., 15, 6516-6524.

Hinchliffe, K.A., Irvine, R.F. and Divecha, N. (1998) Regulation of PtdIns4P 5-kinase C by thrombin- stimulated changes in its phosphorylation state in human platelets. Biochem. J., 329, 115-119.

Hirai, H. and Varmus, H.E. (1990) Mutations in src homology regions 2 and 3 of activated chicken c-src that result in preferential transformation of mouse or chicken cells. Proc. Natl. Acad. Scl USA, 87, 8592- 8596.

Hiroaki, H., Ago, T., Ito, T., Sumimoto, H. and Kohda, D. (2001) Solution structure of the PX domain, a target of the SH3 domain. Nat. Struct. Biol, 8, 526-530.

Homma, K., Terui, S., Minemura, M., Qadota, H., Anraku, Y., Kanaho, Y. and Ohya, Y. (1998) Phosphatidylinositol-4-phosphate 5-kinase localized on the plasma membrane is essential for yeast cell morphogenesis. J. Biol Chem., 273, 15779-15786.

Homma, Y. and Emori, Y. (1995) A dual functional signal mediator showing RhoGAP and phospholipase C-Ô stimulating activities. EMBO. J., 14, 286-291.

Homma, Y. and Emori, Y. (1997) Purification and assay of PLC-Ô. In Shears, S. (ed.) Signalling by Inositides - A Practical Approach. IRL Press, pp. 99-116.

Honda, A., Nogami, M., Yokozeki, T., Yamazaki, M., Nakamura, H., Watanabe, H., Kawamoto, K., Nakayama, K., Morris, A.J., Frohman, M.A. and Kanaho, Y. (1999) Phosphatidylinositol 4-phosphate 5- kinase a is a downstream effector of the small G protein ARF6 in membrane ruffle formation. Cell, 99, 521-532.

203 Honegger, A., Dull, T.J., Bellot, F., van Obberghen, E., Szapary, D., Schmidt, A., Ullrich, A. and Schlessinger, J. (1988) Biological activities of EGF-receptor mutants with individually altered autophosphorylation sites. EMBOJ., 7, 3045-3052.

Honegger, A.M., Dull, T.J., Felder, S., van Obberghen, E., Bellot, F., Szapary, D., Schmidt, A., Ullrich, A. and Schlessinger, J. (1987a) Point mutation at the ATP binding site of EGF receptor abolishes protein tyrosine kinase activity and alters cellular routing. Cell, 51, 199-209.

Honegger, A.M., Szapary, D., Schmidt, A., Lyall, R., van Obberghen, E., Dull, T.J., Ullrich, A. and Schlessinger, J. (1987b) A mutant EGF receptor with defective protein tyrosine kinase activity is unable to stimulate proto-oncogene expression and DNA synthesis. Mol Cell Biol, 7, 4567-4571.

Hsuan, J. and Cockcroft, S. (2001) The PITP family of phosphatidylinositol transfer proteins. Genome BW., 2, 3011.1-3011.8.

Hsuan, J.J. (1993) Oncogene regulation by growth factors.Anticancer Res., 13, 2521-2532.

Hsuan, J.J., Minogue, S. and dos Santos, M. (1998) Phosphoinositide 4- and 5-kinases and the cellular roles of phosphatidylinositol 4,5-bisphosphate./I t/v. Cancer Res., 74, 167-216.

Hsuan, J.J. and Tan, S.K. (1997) Growth factor dependent phosphoinositide signalling.Int. J. Biochem. Cell B/o/., 29, 415-435.

Hsuan, J.J., Totty, N. and Waterfield, M.D. (1989) Identification of a novel autophosphorylation site (P4) on the epidermal growth factor receptor. Biochem. J., 262, 659-663.

Hsueh, Y.P., Kim, E. and Sheng, M. (1997) Disulfide-linked head-to-head multimerization in the mechanism of ion channel clustering by PSD-95. Neuron, 18, 803-814.

Hu, G.F. (1998) Neomycin inhibits angiogenin-induced angiogenesis. Proc. Natl Acad. Scl USA, 95, 9791- 9795.

Hubner, S., Couvillon, A.D., Kas, J.A., Bankaitis, V.A., Vegners, R., Carpenter, C.L. and Janmey, P.A. (1998) Enhancement of phosphoinositide 3-kinase (PI 3-kinase) activity by membrane curvature and inositol-phospholipid-binding peptides. Eur. J. Biochem., 258, 846-853.

Hung, A.Y. and Sheng, M. (2002) PDZ domains: structural modules for protein complex assembly. J. Biol Chem., 277, 5699-5702.

Hunter, T. (1995) When is a lipid kinase not a lipid kinase? When it is a protein kinase.Cell, 83, 1-4.

Hwang, S.C., Jhon, D.-Y., Bae, Y.S., Kim, J.H. and Rhee, S.G. (1996) Activation of phospholipase C-y by the concerted action of tau proteins and arachidonic acid.J. Biol. Chem., 271, 18342-18349.

Ikonomov, O.C., Sbrissa, D., Mlak, K., Kanzaki, M., Pessin, J. and Shisheva, A. (2002) Functional dissection of lipid and protein kinase signals of PlKfyve reveals the role of Ptdlns 3,5-P2 production for endomembrane integrity. J. Biol Chem., 277, 9206-9211.

Irvine, R. (2000) Nuclear lipid signaling. Science's STKE http:// www.stke.org/cgi/content/full/QC_sigtrans; 2000/48/re I.

Irvine, R. and Cullen, P. (1996) Inositol phosphates - whither bound. Curr. Biol, 6, 537-540.

Irvine, R.F. (2002) Nuclear lipid signaling. Science's STKE http://stke.sciencemag.org/cgi/content/full/ sigtrans;2002/150/rel 3.

Irvine, R.F. and Schell, M.J. (2001) Back in the water: the return o f the inositol phosphates.Nal Rev. Mol Cell Biol, 2, 327-338.

Ishihara, H., Shibasaki, Y., Kizuki, N., Katagiri, H., Yazaki, Y., Asano, T. and Oka, Y. (1996) Cloning of cDNAs encoding two isoforms of 68-kDa type 1 phosphatidylinositol 4-phosphate 5-kinase. J. Biol Chem., 271,23611-23614.

Ishihara, H., Shibasaki, Y., Kizuki, N., Wada, T., Yazaki, Y., Asano, T. and Oka, Y. (1998) Type 1 phosphatidylinositol-4-phosphate 5-kinases. Cloning of the third isoform and deletion/substitution analysis of members of this novel lipid kinase family. J. Biol Chem., 273, 8741-8748.

204 Itoh, T., Ijuin, T. and Takenawa, T. (1998) A novel phosphatidylinositoI-5-phosphate 4-kinase (phosphatidylinositol-phosphate kinase Ily) is phosphorylated in the endoplasmic reticulum in response to mitogenic signals. J. Biol. Chem., 273, 20292-20299.

Itoh, T., Ishihara, H., Shibasaki, Y., Oka, Y. and Tajenawa, T. (2000) Autophosphorylation of type I phosphatidylinositol phosphate kinase regulates its lipid kinase activity. J. Biol. Chem., 275, 19389-19394.

Itoh, T., Koshiba, S., Kigawa, T., Kikuchi, A., Yokoyama, S. and Takenawa, T. (2001) Role of the ENTH domain in phosphatidylinositol-4,5-bisphosphate binding and endocytosis. Science, 291, 1047-1051.

Jackson, P. and Baltimore, D. (1989) N-terminal mutations activate the leukemogenic potential of the myristoylated form of c-abl. EMBOJ., 8, 449-456.

Jackson, T.R., Stephens, L.R. and Hawkins, P.T. (1992) Receptor specificity of growth factor-stimulated synthesis of 3-phosphorylated inositol lipids in Swiss 3T3 cells. J. Biol. Chem., 267, 16627-16636.

Janmey, P.A. and Stossel, T.P. (1987) Modulation of gelsolin function by phosphatidylinositol 4,5- bisphosphate. Nature, 325, 362-364.

Jenkins, G.H., Fisette, P.L. and Anderson, R.A. (1994) Type I phosphatidylinositol 4-phosphate 5-kinase isoforms are specifically stimulated by phosphatidic acid. J. Biol. Chem., 269, 11547-11554.

Jenkins, G.H., Subrahmanyam, G. and Anderson, R.A. (1991) Purification and reconstitution of phosphatidylinositol 4-kinase from human erythrocytes. Biochim. Biophys. Acta, 1080, 11-18.

Ji, G.S., Ermini, S., Baulida, J., Sun, P.L. and Carpenter, G. (1998) Epidermal growth factor signaling and mitogensis in PLCyl null mouse embryonic fibroblasts. Mol. Biol. Cell, 9, 749-757.

Ji, Q.-S., Winnier, G.E., Niswender, K.D., Horstman, D., Wisdom, R., Magnuson, M.A. and Carpenter, G. (1997) Essential role of the tyrosine kinase substrate phospholipase C-yl in mammalian growth and development. Proc. Natl. Acad. Sci. USA, 94, 2999-3003.

Johnson, R.M., Connelly, P.A., Sisk, R.B., Pobiner, B.F., Hewlett, E.L. and Garrison, J.C. (1986) Pertussis toxin or phorbol 12-myristate 13-acetate can distinguish between epidermal growth factor- and angiotensin- stimulated signals in hepatocytes. Proc. Natl. Acad. Sci. USA, 83, 2032-2036.

Johnstone, C.N., Tebbutt, N.C., Abud, H.E., White, S.J., Stenvers, K.L., Hall, N.E., Cody, S.H., Whitehead, R.H., Catimel, B., Nice, E.C., Burgess, A.W. and Heath, J.K. (2000) Characterization of mouse A33 antigen, a definitive marker for basolateral surfaces of intestinal epithelial cells. Am. J. Physiol. Gastrointest. Liver Physiol, 279, G500-510.

Jones, A., T. and Wessling-Resnick, M. (1998) Inhibition of in vitro endosomal vesicle fusion activity by aminoglycoside antibiotics. J. Biol Chem., 273, 25301-25309.

Jones, D.H., Morris, J.B., Morgan, C.P., Kondo, H., Irvine, R.F. and Cockcroft, S. (2000a) Type I phosphatidylinositol 4-phosphate 5-kinase directly interacts with ADP-ribosylation factor 1 and is responsible for phosphatidylinositol 4,5-bisphosphate synthesis in the Golgi compartment. J. Biol Chem., 275, 13962-13966.

Jones, D.R., Sanjuan, M.A. and Merida, I. (2000b) Type la phosphatidylinositol 4-phosphate 5-kinase is a putative target for increased intracellular phosphatidic acid.FEBS Lett., 476, 160-165.

Jones, G.A. and Carpenter, G. (1993) The regulation of phospholipase C-yl by phosphatidic acid. J. Biol C/îew.,268, 20845-20850.

Jones, P.L., Crack, J. and Rabinovitch, M. (1997) Regulation of tenascin-C, a vascular smooth muscle cell survival factor that interacts with the a^P^ integrin to promote epidermal growth factor receptor phosphorylation and growth. J. Cell Biol, 139, 279-293.

Jones, S.M., Alb, J.G., Phillips, S.E., Bankaitis, V.A. and Howell, K.E. (1998) A phosphatidylinositol 3- kinase and phosphatidylinositol transfer protein act synergistically in formation of constitutive transport vesicles from the trans-Golgi network. J. Biol Chem., 273, 10349-10354.

Jouannic, N., Lepetit, M., Vergnolle, C., Cantrel, C., Gardies, A.-M., Kader, J.-C. and Arondel, V. (1998) Isolation of a cDNA from Arabidopsis thaliana that complements the sec 14 mutant o f yeast. Eur. J. Biochem., 258, 402-410. 205 Kamat, A. and Carpenter, G. (1997) Phospholipase C-yl: regulation of enzyme function and role in growth factor-dependent signal transduction. Cytokine Growth Factor Rev., 8, 109-117.

Kanaho, Y. and Suzuki, T. (2002) Phosphoinositide kinases as enzymes that produce versatile signaling lipids, phosphoinositides. J. Biochem., 131, 503-509.

Kashishian, A., Kazlauakas, A. and Cooper, J.A. (1992) Phosphorylation sites in the PDGF receptor with different specificities for binding GAP and PI3 kinase in vivo. EMBO J.,\\, 1373-1382.

Kauf&nann-Zeh, A., Klinger, R., Endemann, G., Waterfield, M.D., Wetzker, R. and Hsuan, J.J. (1994) Regulation of human type II phosphatidylinositol kinase activity by epidermal growth factor-dependent phosphorylation and receptor association. J. Biol. Chem., 269, 31243-31251.

Kauffmann-Zeh, A., Thomas, G.M., Ball, A., Prosser, S., Cunningham, E., Cockcroft, S. and Hsuan, J.J. (1995) Requirement for phosphatidylinositol transfer protein in epidermal growth factor signaling. Science, 268, 1188-1190.

Kazlauskas, A., Kashishian, A., Cooper, J.A. and Valius, M. (1992) GTPase-activating protein and phosphatidylinositol 3-kinase bind to distinct regions of the platelet-derived growth factor receptor P subunit. Mol. Cell. Biol., 12, 2534-2544.

Kearns, B.G., Alb, J.G. and Bankaitis, V.A. (1998a) Phosphatidylinositol transfer proteins: the long and winding road to physiological function. Trends Cell Biol., 8, 276-282.

Kearns, B.G., McGee, T.P., Mayinger, P., Gedvilaite, A., Phillips, S.E., Kagiwada, S. and Bankaitis, V.A. (1997) Essential role for diacylglycerol in protein transport from the yeast Golgi complex. Nature, 387, 101-105.

Kearns, M.A., Monks, D.E., Fang, M., Rivas, M.P., Courtney, P.O., Chen, J., Prestwich, G.D., Theibert, A.B., Dewey, R E. and Bankaitis, V.A. (1998b) Novel developmentally regulated phosphoinositide binding proteins from soybean whose expression bypasses the requirement for an essential phosphatidylinositol transfer protein in yeast. EMBOJ., 17, 4004-4017.

Kelley, G.G., Reks, S.E., Ondrako, J.M. and Smrcka, A.V. (2001) Phospholipase Ce: a novel Ras effector. EMBOJ, 20, 743-754.

Kim, D., Jun, K.S., Lee, S.B., Kang, N.-G., Min, O.S., Kim, Y.-H., Ryu, S.H., Suh, P.-G. and Shin, H.-S. (1997) Phospholipase C isozymes selectively couple to specific neurotransmitter receptors. Nature, 389, 290-293.

Kim, H.K., Kim, J.W., Zilberstein, A., Margolis, B., Kim, J.G., Schlessinger, J. and Rhee, S.G. (1991) PDGF stimulation of inositol phospholipid hydrolysis requires PLC-yl phosphorylation on tyrosine residues 783 and 1254. Cell, 65, 435-441.

Kim, J.W., Sim, S.S., Kim, U.H., Nishibe, S., Wahl, M.I., Carpenter, G. and Rhee, S.G. (1990) Tyrosine residues in bovine phospholipase C-y phosphorylated by the epidermal growth factor receptor in vitro. J. Biol Chem., 265, 3940-3943.

Kim, U.-H., Kim, J.W. and Rhee, S.G. (1989) Phosphorylation of phospholipase C-y by cAMP-dependent protein kinase. J. Biol. Chem., 264, 20167-20170.

King, L.A. and Possee, R.D. (1992) The Baculovirus Expression System: A Laboratory Guide. Chapman & Hall.

Klauck, T.M., Xu, X., Mousseau, B. and Jaken, S. (1996) Cloning and characterisation of a glucocorticoid- induced diacylglycerol kinase. J. Biol Chem., 271, 19781-19788.

Koelle, M.R. and Horvitz, H.R. (1996) EGL-10 regulates G protein signaling in the C. elegans nervous system and shares a conserved domain with many mammalian proteins. Cell, 84, 115-125.

Kubatzky, K.F., Ruan, W., Gurezka, R., Cohen, J., Ketteler, R., Watowich, S.S., Neumann, D., Langosch, D. and Klingmuller, U. (2001) Self assembly of the transmembrane domain promotes signal transduction through the erythropoietin receptor. Curr. Biol, 11, 110-115.

Kular, G., Loubtchenkov, M., Swigart, P., Whatmore, J., Ball, A., Cockcroft, S. and Wetzker, R. (1997) Co­ operation of phosphatidylinositol transfer protein with phosphoinositide 3-kinase y in the formylmethionyl-

206 leucylphenylalanine-dependent production of phosphatidylinositol 3,4,5-trisphosphate in human neutrophils. Biochem. J., 325, 299-301.

Kundra, V., Escobeto, J.A., Kazlauskas, A., Kim, H.K. and Rhee, S.G. (1994) Regulation of chemotaxis by the platelet-derived growth-factor receptor-p. Nature, 367, 474-476.

Kunz, J., Wilson, M.P., Kisseleva, M., Hurley, J.H., Majerus, P.W. and Anderson, R.A. (2000) The activation loop of phosphatidylinositol phosphate kinases determines signaling specificity. Mol Cell, 5, 1- 1 1 .

Kutateladze, T.G., Ogbum, K.D., Watson, W.T., de Beer, T., Emr, S.D., Burd, C.G. and Overduin, M. (1999) Phosphatidylinositol 3-phosphate recognition by the FYVE domain. Mol Cell, 3, 805-811.

Lassing, I. and Lindberg, U. (1985) Specific interaction between phosphatidylinositol 4,5-bisphosphate and profilactin. Nature, 314, 472-474.

Lassing, I. and Lindberg, U. (1988) Specificity of the interaction between phosphatidylinositol 4,5- bisphosphate and the profilin-actin complex. J. Cell Biochem., 37, 255-267.

Lauener, R., Shen, Y., Duronio, V. and Salari, H. (1995) Selective inhibition of phosphatidylinositol 3- kinase by phosphatidic cid and related lipids. Biochem. Biophys. Res. Com., 215, 8-14.

Lechleider, R.J., Sugimoto, S., Bennet, A.M., Kashishian, S.E., Walsh, G.T. and Neel, B.G. (1993) Activation of the SH2-containing phosphotyrosine phosphatase SH-PTP2 by its binding site, phosphotyrosine 1009, on the human platelet-derived growth factor receptor. J. Biol Chem., 268, 21478- 21481.

Lee, C., Fisher, S.K., Agranoff, B.W. and Hajra, A.K. (1991) Quantitative analysis of molecular species of diacylglycerol and phosphatidate formed upon muscarinic receptor activation of human SK-N-SH neuroblastoma cells. J. Biol Chem., 266, 22837-22846.

Lee, S B. and Rhee, S.G. (1995) Significance of PIP2 hydrolysis and regulation of phospholipase C isozymes. Curr. Opin. Cell Biol, 7, 183-189.

Lee, S.B., Shin, S.H., Hepler, J.R., Gilman, A.G. and Rhee, S.G. (1993) Activation of phospholipase C-(32 mutants by G protein and Py subunits. J. Biol Chem., 268, 25952-25957.

Lemmon, M.A., Ferguson, K.M., O'Brien, R., Sigler, P.B. and Schlessinger, J. (1995) Specific and high- affmity binding of inositol phosphates to an isolated pleckstrin homology domain. Proc. Natl Acad. Scl USA, 92, 10472-10476.

Lenferink, A.E., Pinkas-Kramarski, R., van de Poll, M.L., van Vugt, M.J., Klapper, L.N., Tzahar, E., Waterman, H., Sela, M., van Zoelen, E.J. and Yarden, Y. (1998) Differential endocytic routing of homo- and hetero-dimeric ErbB tyrosine kinases confers signaling superiority to receptor heterodimers. EMBO J., 17, 3385-3397.

Leslie, N.R., Biondi, R.M. and Alessi, D.R. (2001) Phosphoinositide-regulated kinases and phosphoinositide phosphatases. Chem. Rev., 101, 2365-2380.

Lev, S., Hernandez, J., Martinez, R., Chen, A., Plowman, G. and Schlessinger, J. (1999) Identification of a novel family of targets of PYK2 related to Drosophila retinal degeneration B (rdgB) protein. Mol Cell Biol, 19, 2278-2288.

Li, H., Tremblay, J.M., Yarbrough, L.R. and Helmkamp, G.M. (2002) Both isoforms of mammalian phosphatidylinositol transfer protein are capable of binding and transporting sphingomyelin. Biochim. Biophys. Acta, 1580, 67-76.

Li, S.C., Songyang, Z., Vincent, S.J., Zwahlen, C., Wiley, S., Cantley, L., Kay, L.E., Forman-Kay, J. and Pawson, T. (1997) High-affmity binding of the Drosophila Numb phosphotyrosine-binding domain to peptides containing a Gly-Pro-(p)Tyr motif. Proc. Natl Acad. Scl USA, 94, 7204-7209.

Li, X., Routt, S.M., Xie, Z., Cui, X., Fang, M., Kearns, M.A., Bard, M., Kirsch, D.R. and Bankaitis, V.A. (2000a) Identification of a novel family of nonclassic yeast phosphatidylinositol transfer proteins whose function modulates phospholipase D activity and Secl4p-independent cell growth. Mol Biol Cell, 11, 1989-2005.

207 Li, X., Xie, Z. and Bankaitis, V.A. (2000b) Phosphatidylinositol/phosphatidylcholine transfer proteins in yeast. Biochim. Biophys. Acta, 1486, 55-71.

Li, Z., Jiang, H., Xie, W., Zhang, Z., Smrcka, A.V. and Wu, D. (2000c) Roles of PLC-P2 and -p3 and PDKy in chemoattractant-mediated signal transduction. Science, 287, 1046-1049.

Liang, M. and Garrison, J.C. (1992) Epidermal growth factor activates phospholipase C in rat hepatocytes via a different mechanism from that in A 431 or ratlhER cells. Mol. Pharmacol, 42, 743-752.

Liao, H.-J., Kume, T., McKay, C., Xu, M.-J., Ihle, J.N. and Carpenter, G. (2002) Absence of erythrogenesis and vasculogenesis in f/eg 1-deficient mice. J. Biol Chem., 277, 9335-9341.

Lin, H., Choi, J.H. and Vancura, A. (1998) Phosphoinositide-specific phospholipase C interacts with phosphatidylinositol kinase homolog T0R2. Biochem. Biophys. Res. Commun., 252, 285-289.

Lin, S.-Y., Makino, K., Xia, W., Matin, A., Wen, Y., Kwong, K.Y., Bourguignon, L. and Hung, M.-C. (2001) Nuclear localization of EGF receptor and its potential new role as a transcription factor.Nat. Cell Biol, 3, 802-808.

Ling, L.E., Schulz, J.T. and Cantley, L.C. (1989) Characterization and purification of membrane-associated phosphatidylinositol-4-phosphate kinase from human red blood cells. J. Biol Chem., 264, 5080-5088.

Liscovitch, M. and Cantley, L.C. (1995) Signal transduction and membrane traffic: the PlTP/phosphoinositide connection. Cell, 81, 659-662.

Liscovitch, M., Chalifa, V., Pertile, P., Chen, C.S. and Cantley, L.C. (1994) Novel function of phosphatidylinositol 4,5-bisphosphate as a cofactor for brain membrane phospholipase D. J. Biol Chem., 269,21403-21406.

Litosch, I. (2000) Regulation of phospholipase C-(31 activity by phosphatidic acid. Biochemistry, 39, 7736- 7743.

Litvak, V., Shaul, Y.D., Shulewitz, M., Amarilio, V., Carmon, S. and Lev, S. (2002) Targeting of Nir2 to lipid droplets is regulated by a specific threonine residue within its Pl-transfer domain. Curr. Biol, 12, 1513-1518.

Livneh, E., Prywes, R., Kashles, O., Reiss, N., Sasson, I., Mory, Y., Ullrich, A. and Schlessinger, J. (1986) Reconstitution of human epidermal growth factor receptors in cultured hamster cells. J. Biol Chem., 261, 12490-12497.

Loijens, J.C. and Anderson, R.A. (1996) Type I phosphatidylinositol-4-phosphate 5-kinases are distinct members of this novel lipid kinase family. J. Biol Chem., 271, 32937-32943.

Lomasney, J.W., Cheng, H.-F., Wang, L. P., Kuan, Y.-S., Liu, S.-M., Fesik, S.W. and Kim, K. (1996) Phosphatidylinositol 4,5-bisphosphate binding to the pleckstrin homology domain of phospholipase C-Ô1 enhances enzyme activity. J. Biol Chem., 271, 25316-25326.

Lopez Garcia, F., Szyperski, T., Dyer, J.H., Choinowski, T., Seedorf, U., Hauser, H. and Wuthrich, K. (2000) NMR structure of the sterol carrier protein-2: implications for the biological role. J. Mol Biol, 295, 595-603.

Lopez, I., Mak, E.C., ding, J., Hamm, H.E. and Lomasney, J.W. (2001) A novel bifunctional phospholipase

C that is regulated by G a ,2 and stimulates the Ras/mitogen-activated protein kinase pathway. J. Biol Chem., 276, 2758-2765.

Lopez, M.C., Nicaud, J.M., Skinner, H.B., Vergnolle, C., Kader, J.C., Bankaitis, V.A. and Gaillardin, C. (1994) A phosphatidylinosito 1/phosphatidylcholine transfer protein is required for differentiation of the dimorphic yeast Yarrowia lipolytica from the yeast to the mycelial form. J. Cell Biol, 125, 113-127.

Lopez-Girona, A., Fumari, B., Mondesert, O. and Russell, P. (1999) Nuclear localization of Cdc25 is regulated by DNA damage and a 14-3-3 protein. Nature, 397, 172-175.

Lowenstein, E.J., Daly, R.J., Batzer, A.G., Li, W., Margolis, B., Lammers, R., Ullrich, A., Skolnik, E.Y., Bar, S.D. and Schlessinger, J. (1992) The SH2 and SH3 domain-containing protein GRB2 links receptor tyrosine kinases to ras signaling. Cell, 70, 431-442.

208 Loyet, K.M., Kowalchyk, J.A., Chaudhary, A., Chen, J., Prestwich, G.D. and Martin, T.F.J. (1998) Specific binding of phosphatidylinositol 4,5-bisphosphate to calcium-dependent activator protein for secretion (CAPS), a potential phosphoinositide effector protein for regulated exocytosis. J. Biol. Chem., 273, 8337- 8343.

Lu, C., Vihtelic, T.S., Hyde, D.R. and Li, T. (1999a) A neuronal-specific mammalian homolog of the Drosophila retinal degeneration B gene with expression restricted to the retina and dentate gyrus. J. Neurosci., 19, 7317-7325.

Lu, P.-J., Shieh, W.-R., Rhee, S.G., Yin, H.L. and Chen, C.-S. (1996) Lipid products of phosphoinositide 3- kinase bind human profilin with high affinity. Biochemistry, 35, 14027-14034.

Lu, P.-J., Zhou, X.Z., Shen, M. and Lu, K.P. (1999b) Function of WW domains as phosphoserine- or phosphothreonine-binding modules. Science, 283, 1325-1328.

Luttrell, L.M., Rocca, G.J.D., van Biesen, T., Luttrell, D.K. and Lefkowitz, R.J. (1997) GPy subunits mediate src-dependent phosphorylation of the epidermal growth factor receptor. J. Biol. Chem., 272, 4637- 4644.

Lykidis, A., Jackson, P.O., Rock, C.O. and Jackowski, S. (1997) The role of CDP-diacylglycerol synthetase and phosphatidylinositol synthase activity levels in the regulation of cellular phosphatidylinositol content. J. Biol. Chem., 272, 33402-33409.

Machesky, L.M., Goldschmidt-Clermont, P.J. and Pollard, T.D. (1990) The affinities of human platelet and Acanthamoeba profilin isoforms for polyphosphoinositides account for their relative abilities to inhibit phospholipase C. Cell Regul., 1, 937.

Macias, M.J., Hyvonen, M., Baraldi, E., Schultz, J., Sudol, M., Saraste, M. and Oschkinat, H. (1996) Structure of the WW domain of a kinase-associated protein complexed with a proline-rich peptide. Nature, 382, 646-649.

Mackay, D.J.G. and Hall, A. (1998) Rho GTPases.J. Biol. Chem., 273, 20658-20658.

Majerus, P.W., Kisseleva, M.V. and Norris, F.A. (1999) The role of phosphatases in inositol signaling reactions, y. Biol. Chem., 274, 10669-10672.

Maraldi, N.M., Zini, N., Santi, S., Riccio, M., Falconi, M., Capitani, S. and Manzoli, F.A. (2000) Nuclear domains involved in inositol lipid signal transduction. Adv. Enzyme Regul, 40, 219-253.

Marcus, S., Polverino, A., Barr, M. and Wigler, M. (1994) Complexes between STE5 and components of the pheromone-responsive mitogen-activated protein kinase module. Proc. Natl Acad. Scl USA, 91, 7762- 7766.

Margolis, B., Rhee, S.G., Felder, S., Mervic, M., Lyall, R., Levitzki, A., Ullrich, A., Zilberstein, A. and Schlessinger, J. (1989a) EGF induces tyrosine phosphorylation of phospholipase C-II: a potential mechanism for EGF receptor signaling. Cell, 57, 1101-1107.

Margolis, B.L., Lax, I., Kris, R., Dombalagian, M., Honegger, A.M., Howk, R., Givol, D,, Ullrich, A. and Schlessinger, J. (1989b) All autophosphorylation sites of epidermal growth factor (EGF) receptor and HER2/neu are located in their carboxyl-terminal tails. J. Biol Chem., 264, 10667-10671.

Marheineke, K., Grünewald, S., Christie, W. and Reilander, H. (1998) Lipid composition ofSpodoptera frugiperda (Sf9) and Trichoplusia ni (Tn) insect cells used for baculovirus infection. FEBS Lett., 441, 49- 52.

Martin, T.F.J. (1997) Greasing the Golgi budding machine.Nature, 387, 21-22.

Mazzotti, G., Zini, N., Rizzi, E., Rizzoli, R., Galanzi, A., Ognibene, A., Santi, S., Matteucci, A. and Maraldi, N.M. (1995) Immunocytochemical detection of phosphatidylinositol 4,5-bisphosphate localization sites within the nucleus. J. Histochem. Cytochem., 43, 181-191.

McEwen, R.K., Dove, S.K., Cooke, F.T., Painter, G.F., Holmes, A.B., Shisheva, A., Ohya, Y., Parker, P.J. and Michell, R.H. (1999) Complementation analysis in PtdlnsP kinase-deficient yeast mutants demonstrates that Schizosaccharomyces pombe and murine Fab Ip homologues are phosphatidylinositol 3-phosphate 5- kinases. J. Biol. Chem., 274, 33905-33912.

209 McGee, T.P., Skinner, H.B., Whitters, E.A., Henry, S.A. and Bankaitis, V.A, (1994) A phosphatidylinositol transfer protein controls the phosphatidylcholine content of yeast Golgi membranes. J. Cell Biol., 124, 273- 87.

Meijer, H.J.G., Berrie, C.P., lurisci, C., Divecha, N., Musgrave, A. and Munnik, T. (2001) Identification of a new polyphosphoinositide in plants, phosphatidylinositol 5-monophosphate (PtdlnsSP), and its accumulation upon osmotic stress. Biochem. J., 360, 491-498.

Meisenhelder, J., Suh, P.G., Rhee, S.G. and Hunter, T. (1989) Phospholipase C-y is a substrate for the PDGF and EGF receptor protein-tyrosine kinases in vivo and in vitro. Cell, 57, 1109-1122.

Mendrola, J.M., Berger, M.B., King, M.G. and Lemmon, M.A. (2002) The single transmembrane domains of ErbB receptors self-associate in cell membranes. J. Biol. Chem., 277,4704-4712.

Meyers, R. and Cantley, L.C. (1997) Cloning and characterization of a wortmannin-sensitive human phosphatidylinositol 4-kinase. J. Biol. Chem., 272, 4384-4390.

Michell, R.H. (1975) Inositol phospholipids and cell surface receptor function. Biochim. Biophys. Acta, 415, 81-147.

Michell, R.H., Kirk, C.J., Maccallum, S.H. and Hunt, P.A. (1988) Inositol lipids: receptor-stimulated hydrolysis and cellular lipid pools. Phil. Trans. R. Soc. Lond, B 320, 239-246.

Milligan, S.C., Alb, J.G., Jr., Elagina, R.B., Bankaitis, V.A. and Hyde, D.R. (1997) The phosphatidyl inositol transfer protein domain of Drosophila retinal degeneration B protein is essential for photoreceptor cell survival and recovery from light stimulation. J. Cell Biol, 139, 351-363.

Minogue, S., Anderson, J.S., Waugh, M.G., dos Santos, M., Corless, S., Cramer, R. and Hsuan, J.J. (2001) Cloning of a human type 11 phosphatidylinositol 4-kinase reveals a novel lipid kinase family. J. Biol Chem.,216, 16635-16640.

Misra, S. and Hurley, J.H. (1999) Crystal structure of a phosphatidylinositol 3-phosphate-specific membrane-targeting motif, the FYVE domain of Vps27p. Cell, 97, 657-666.

Miyamoto, S., Teramoto, H., Gutkind, J.S. and Yamada, K.M. (1996) Integrins can collaborate with growth factors for phosphorylation of receptor tyrosine kinases and MAP kinase activation: roles of integrin aggregation and occupancy of receptors. J. Cell Biol, 135, 1633-1642.

Mohammadi, M., Honegger, A.M., Rotin, D., Fischer, R., Bellot, F., Li, W., Dionne, C.A., Jaye, M., Rubinstein, M. and Schlessinger, J. (1991) A tyrosine-phosphorylated carboxy-terminal peptide of the fibroblast growth factor receptor (Fig) is a binding site for the SH2 domain of phospholipase C-yl. Mol Cell Biol, 11, 5068-5078.

Monaco, M.E., Alexander, R.J., Snoek, G.T., Moldover, N.H., Wirtz, K.W.A. and Walden, P.D. (1998) Evidence that mammalian phosphatidylinositol transfer protein regulates phosphatidylcholine metabolism. Biochem. J., 335, 175-179.

Monaco, M.E. and Gershengom, M.C. (1992) Subcellular organization of receptor-mediated phosphoinositide turnover. Endocr. Rev., 13, 707-718.

Moolenaar, W.H. (1994) LPA: a novel lipid mediator with diverse biological actions. Trends Cell Biol, 4, 213-219.

Moolenaar, W.H., Bierman, A.J., Tilly, B.C., Verlaan, I., Defize, L.K.H., Honegger, A.M., Ullrich, A. and Schlessinger, J. (1988) A point mutation at the ATP binding site of EGF receptor abolishes signal transduction. EMBOJ., 7, 707-710.

Moran, M.F., Polakis, P., McCormick, F., Pawson, T. and Ellis, C. (1991) Protein-tyrosine kinases regulate the phosphorylation, protein interactions, subcellular distribution, and activity of p21"^ GTPase-activating protein. Mol Cell. Biol, 11, 1804-1812.

Morris, J.B., Hinchliffe, K.A., Ciruela, A., Letcher, A.J. and Irvine, R.F. (2000) Thrombin stimulation of platelets causes an increase in phosphatidylinositol 5-phosphate revealed by mass assay. FEBS Lett., 475, 57-60.

210 Nakagawa, T., Goto, K. and Kondo, H. (1996a) Cloning and characterization of a 92 kDa soluble phosphatidylinositol 4-kinase, Biochem. J., 320, 643-649.

Nakagawa, T., Goto, K. and Kondo, H. (1996b) Cloning, expression, and localization of 230-kDa phosphatidylinositol 4-kinase. y. Biol. Chem.,H\, 12088-12094.

Nakanishi, S., Catt, K.J. and Balia, T. (1995) A wortmannin-sensitive phosphatidylinositol 4-kinase that regulates hormone-sensitive pools of inositolphospholipids. Proc. Natl. Acad. Sci. USA, 92, 5317-5321.

Nakaoka, H., Perez, D.M., Baek, K.J., Das, T., Husain, A., Misono, K., Im, M.-J. and Graham, R.M. (1994) G|^: a GTP-binding protein with transglutaminase activity and receptor signaling function. Science, 264, 1593-1596.

Neri, L.M., Borgatti, P., Capitani, S. and Martelli, A.M. (1998) Nuclear diacylglycerol produced by phosphoinositide-specific phospholipase C is responsible for nuclear translocation of protein kinase C-a. J. Biol. Chem., 273, 29738-29744.

Ni, C.-Y., Murphy, M.P., Golde, T.E. and Carpenter, G. (2001) y-Secretase cleavage and nuclear localization of ErbB-4 receptor tyrosine kinase. Science, 294, 2179-2181.

Nishibe, S., Wahl, M.I., Hernandez, S.S., Tonks, N.K., Rhee, S.G. and Carpenter, G. (1990) Increase of the catalytic activity of phospholipase C-yl by tyrosine phosphorylation. Science, 250, 1253-1256.

Nishimura, R., Li, W., Kashishian, A., Mondino, A., Zhou, M., Cooper, J. and Schlessinger, J. (1993) Two signaling molecules share a phosphotyrosine-containing binding site in the platelet-derived growth factor receptor. Mol. Cell. Biol., 13, 6889-6896.

Nishizuka, V. (1988) The molecular heterogeneity of protein kinase C and its implications for cellular regulation. Nature, 334, 661-665.

Noh, D.-Y., Shin, S.H. and Rhee, S.G. (1995) Phosphoinositide-specific phospholipase C and mitogenic signaling. Biochim. Biophys. Acta, 1242, 99-114.

Obermeier, A., Halfter, H., Wiesmuller, K.H., Jung, G., Schlessinger, J. and Ullrich, A. (1993a) Tyrosine 785 is a major determinant of Trk-substrate interaction. EMBOJ., 12, 933-941.

Obermeier, A., Lammers, R., Wiesmuller, K.-H., Jung, G., Schlessinger, J. and Ullrich, A. (1993b) Identification of Trk binding sites for SHC and phosphatidylinositol 3'-kinase and formation of a multimeric signaling complex. J. Biol. Chem., 268, 22963-22966.

Odom, A.R., Stahlberg, A., Wente, S.R. and York, J.D. (2000) A role for nuclear inositol 1,4,5- trisphosphate kinase in transcriptional control. Science, 287,2026-2029.

Odorizzi, G., Babst, M. and Emr, S.D. (1998) Fab Ip PtdIns(3)P 5-kinase function essential for protein sorting in the multivesicular body. Cell, 95, 847-858.

Odorizzi, G., Babst, M. and Emr, S.D. (2000) Phosphoinositide signaling and the regulation of membrane trafficking in yeast. Trends Biochem. Sci., 25, 229-235.

Ohashi, M., de Vries, K.J., Frank, R., Snoek, G., Bankaitis, V., Wirtz, K. and Huttner, W.B. (1995) A role for phosphatidylinositol transfer protein in secretory vesicle formation. Nature, 377, 544-547.

Olashaw, N.E., Rhee, S.G. and Pledger, W.J. (1990) Cyclic AMP agonists induce the phosphorylation of phospholipase C-tau and of a 76 kDa protein co-precipitated by anti-(phospholipase C-x) monoclonal antibodies in BALB/c-3T3 cells. Relationship to inositol phosphate formation. Biochem. J., 272, 297-303.

Olayioye, M.A., Graus-Porta, D., Beerli, R.R., Rohrer, J., Gay, B. and Hynes, N.E. (1998) ErbB-1 and ErbB-2 acquire distinct signaling properties dependent upon their dimérisation partner. Mol. Cell. Biol, 18, 5042-5051.

Olayioye, M.A., Neve, R.M., Lane, H.A. and Hynes, N.E. (2000) The ErbB signaling network: receptor heterodimerization in development and cancer. EMBOJ., 19, 3159-3167.

O'Sullivan, A.J., Brown, A.M., Freeman, H.N.M. and Gomperts, B.D. (1996) Purification and identification of Foad-Il, a cytosolic protein that regulates secretion in streptolysin-0 permeabilised mast cells, as a rac/rhoGDl complex. Mol. Biol. Cell, 7, 397-408.

211 Oude Weemink, P.A., Schulte, P., Guo, Y., Wetzel, J., Amano, M., Kaibuchi, K., Haverland, S., Vofi, M., Schmidt, M., Mayr, G.W. and Jakobs, K.H. (2000) Stimulation of phosphatidylinositol-4-phosphate 5- kinase by Rho-kinase. J. Biol. Chem., 275, 10168-10174.

Paetkau, D.W., Elagin, V.A., Sendi, L.M. and Hyde, D.R. (1999) Isolation and characterisation of Drosophila retinal degeneration B suppressors. Genetics, 151, 713-724.

Panaretou, C., Domin, J., Cockcroft, S. and Waterfield, M.D. (1997) Characterization of pi 50, an adaptor protein for the human phosphatidylinositol (Ptdlns) 3-kinase. Substrate presentation by phosphatidylinositol transfer protein to the plSO.Ptdins 3-kinase complex. J. Biol. Chem., 272, 2477-2485.

Paris, S., Beraud-Dufour, S., Robineau, S., Bigay, J., Antonny, B., Chabre, M. and Chardin, P. (1997) Role of protein-phospholipid interactions in the activation of ARFl by the guanine nucleotide exchange factor Amo. J. Biol. Chem., 272, 22221-22226.

Park, D., Jhon, D.Y., Lee, C.W., Lee, K.H. and Rhee, S.G. (1993) Activation of phospholipase C isozymes by G protein Py subunits. J. Biol. Chem., 268, 4573-4576.

Park, D.J., Min, H.K. and Rhee, S.G. (1992) Inhibition of CD3-linked phospholipase C by phorbol ester and by cAMP is associated with decreased phosphotyrosine and increased phosphoserine contents of PLC-yl. J. Biol. Chem., 267, 1496-1501.

Park, S.J., Itoh, T. and Takenawa, T. (2001) Phosphatidylinositol 4-phosphate 5-kinase type I is regulated through phosphorylation response by extracellular stimuli. J. Biol. Chem., 276, 4781-4787.

Parker, P.J., Hemmings, B.A. and Gierschik, P. (1994) PH domains and phospholipases-a meaningful relationship? Trends Biochem. Sci., 19, 54-55.

Parton, R.G. (1996) Caveolae and caveolins. Curr. Opin. Cell Biol, 8, 542-548.

Paterson, A., Filtz, T.M. and Harden, T.K. (1997) Baculovims-promoted expression, purification, and functional assay of G-protein regulated PLC-P isoenzymes. In Shears, S. (ed.) Signalling by Inositides - A Practical Approach. IRL Press, pp. 85-98.

Paterson, H.F., Savopoulos, J.W., Perisic, O., Cheung, R., Ellis, M.V., Williams, R.L. and Katan, M. (1995) Phospholipase C 61 requires a pleckstrin homology domain for interaction with the plasma membrane. Biochem. J., 312, 661-666.

Patki, V., Lawe, D.C., Corvera, S., Virbasius, J.V. and Chawla, A. (1998) A functional PtdIns(3)P-binding motif. Nature, 394, 433-434.

Paul, K.S., Bogan, A.A. and Waters, M.G. (1998) Phosphatidylinositol transfer protein (PITPa) stimulates in vitro intra-Golgi transport. FEBS Letters, 431, 91-96.

Pawson, T. (1995) Protein modules and signalling networks. Nature, 373, 573-580.

Pawson, T. and Scott, J.D. (1997) Signaling through scaffold, anchoring, and adaptor proteins.Science, 278, 2075-2080.

Payne, W.E. and Fitzgerald-Hayes, M. (1993) A mutation in PLCl, a candidate phosphoinositide-specific phospholipase C gene from Saccharomyces cerevisiae, causes aberrant mitotic chromosome segregation. Mol Cell Biol, 13, 4351-4364.

Payrastre, B., Nie vers, M., Boonstra, J., Breton, M., Verkleij, A.J. and van Bergen en Henegouwen, P.M. (1992) A differential location of phosphoinositide kinases, diacylglycerol kinase, and phospholipase C in the nuclear matrix. J. Biol Chem., 267, 5078-5084.

Payrastre, B., Plantavid, M., Breton, M., Chambaz, E. and Chap, H. (1990) Relationship between phosphoinositide kinase activities and protein tyrosine phosphorylation in plasma membranes from A431 cells. Biochem. J., 272, 665-670.

Payrastre, B., van Bergen en Henegouwen, P.M., Breton, M., den Hartigh, J.C., Plantavid, M., Verkleij, A.J. and Boonstra, J. (1991) Phosphoinositide kinase, diacylglycerol kinase, and phospholipase C activities associated to the cytoskeleton: effect of epidermal growth factor. J. Cell Biol, 115, 121-128.

212 Pertile, P., Liscovitch, M., Chalifa, V, and Cantley, L.C. (1995) Phosphatidylinositol 4,5-bisphosphate is required for activation of phospholipase D in U937 cells. J. Biol. Chem., 270, 5130-5135.

Phillips, S.E., Sha, B., Topalof, L., Xie, Z., Alb, J.G., Klenchin, V.A., Swigart, P., Cocokcroft, S., Martin, T.F.J., Luo, M. and Bankaitis, V.A. (1999) Yeast Secl4p deficient in phosphatidylinositol transfer activity is functional in vivo. Mol. Cell, 4, 187-197.

Pierce, K.L., Premont, R.T. and Lefkowitz, R.J. (2002) Seven-transmembrane receptors. Nat. Rev. Mol. Cell Biol, 3, 639-650.

Pike, L.J. (1992) Phosphatidylinositol 4-kinases and the role of polyphosphoinositides in cellular regulation. Endocr. Rev., 13, 692-706.

Pike, L.J. and Bakes, A.T. (1987) Epidermal growth factor stimulates the production of phosphatidylinositol monophosphate and the breakdown of polyphosphoinositides in A431 cells. J. Biol Chem., 262, 1644- 1651.

Pongs, O., Lindemeier, J., Zhu, X.R., Theil, T., Engelkamp, D., Krah-Jentgens, I., Lambrecht, H.G., Koch, K.W., Schwemer, J. and Rivosecchi, R. (1993) Frequenin-a novel calcium-binding protein that modulates synaptic efficacy in the Drosophila nervous system. Neuron, 11, 15-28.

Pouting, C.P. and Aravind, L. (1999) START: a lipid-binding domain in StAR, HD-ZIP and signaling proteins. Trends Biochem. Scl, 24, 130-132.

Pouting, C.P. and Benjamin, D.R. (1996) A novel family of Ras-binding domains. Trends Biochem. Scl, 21, 422-425.

Pouting, C.P. and Bork, P. (1996) Pleckstrin's repeat performance: a novel domain in G-protein signaling? Trends Biochem. Scl, 21, 245-246.

Pouting, C.P. and Parker, P.J. (1996) Extending the C2 domain family: C2s in PKCs delta, episilon, eta, theta, phospholipases, GAPs and perforin. Protein Scl, 5, 162-166.

Pouting, C.P., Phillips, C., Davies, K.E. and Blake, D.J. (1997) PDZ domains: targeting signalling molecules to sub-membranous sites. BioEssays, 19,469-479.

Prasad, K.V., Kapeller, R., Janssen, O., Repke, H., Duke-Cohan, J.S., Cantley, L.C. and Rudd, C.E. (1993) Phosphatidylinositol (PI) 3-kinase and PI 4-kinase binding to the CD4-p56'‘^'^ complex: the p56*‘^*‘ SH3 domain binds to PI 3-kinase but not PI 4-kinase. Mol Cell Biol, 13, 7708-7717.

Prenzel, N., Zwick, E., Daub, H., Leserer, M., Abraham, R., Wallasch, C. and Ullrich, A. (1999) EGF receptor transactivation by G-protein-coupled receptors requires metalloproteinase cleavage of proHB-EGF. Nature, 402, 884-888.

Price, L.S., Norman, J.C., Ridley, A.J. and Koffer, A. (1995) The small GTPases Rac and Rho as regulators of secretion in mast cells. Curr. Biol, 5, 68-73.

Pronk, G.J., McGlade, J., Pelicci, G., Pawson, T. and Bos, J.L. (1993) Insulin-induced phosphorylation of the 46- and 52-kDa She proteins. J. Biol Chem., 268, 5748-5753.

Prosser, S., Sarra, R., Swigart, P., Ball, A. and Cockcroft, S. (1997) Deletion of 24 amino acids from the C- terminus of phosphatidylinositol transfer protein causes loss of phospholipase C-mediated inositol lipid signalling. Biochem. J., 324, 19-23.

Prywes, R., Livneh, E., Ullrich, A. and Schlessinger, J. (1986) Mutations in the cytoplasmic domain of EGFR affect EGF binding and receptor internalisation. EMBOJ., 5, 2179-2190.

Raghuram, V., Mak, D.D. and Foskett, J.K. (2001) Regulation of cystic fibrosis transmembrane conductance regulator single-channel gating by bivalent PDZ-domain-mediated interaction. Proc. Natl Acad. Scl USA, 98, 1300-1305.

Rameh, L.E., Arvidsson, A., Carraway, K.L., Couvillon, A.D., Rathbun, G., Crompton, A., VanRenterghem, B., P., C.M., Ravichandran, K.S., Burakoff, S.J., Wang, D.S., Chen, C.S. and Cantley, L.C. (1997a) A comparative analysis of the phosphoinositide binding specificity of pleckstrin homology domains. J. Biol. Chem., 272, 22059-22066.

213 Rameh, L.E., Chen, C.S. and Cantley, L.C. (1995) Phosphatidylinositol (3,4,5)Pg interacts with SH2 domains and modulates PI 3-kinase association with tyrosine-phosphorylated proteins. Cell, 83, 821-830.

Rameh, L.E., Rhee, S.G., Spokes, K., Kazlauskas, A., Cantley, L.C. and Cantley, L.G. (1998) Phosphoinositide 3-kinase regulates phospholipase Cy-mediated calcium signaling. J. Biol. Chem., 273, 23750-23757.

Rameh, L.E., Tolias, K.F., Duckworth, B.C. and Cantley, L.C. (1997b) A new pathway for synthesis of phosphatidylinositol-4,5-bisphosphate. Nature, 390, 192-196.

Randazzo, P.A. and Kahn, R.A. (1994) GTP hydrolysis by ADP-ribosylation factor is dependent on both an ADP-ribosylation factor GTPase-activating protein and acid phospholipids. J. Biol. Chem., 269, 10758- 10763.

Rando, O.J., Zhao, K., Janmey, P. and Crabtree, G.R. (2002) Phosphatidylinositol-dependent actin filament binding by the SWl/SNF-like RAF chromatin remodeling complex. Proc. Natl. Acad. Sci. USA, 99, 2824- 2829.

Ranganathan, R., Lu, K.P., Hunter, T. and Noel, J.P. (1997) Structural and functional analysis o f the mitotic rotamase Pinl suggests substrate recognition is phosphorylation dependent. Cell, 89, 875-886.

Rao, V.D., Misra, S., Boronenkov, I.V., Anderson, R.A. and Hurley, J.H. (1998) Structure of type lip phosphatidylinositol phosphate kinase: A protein kinase fold flattened for interfacial phosphorylation. Cell, 94, 829-839.

Ren, X.-D., Bokoch, G.M., Traynor-Kaplan, A., Jenkins, G.H., Anderson, R.A. and Schwartz, M.A. (1996) Physical association of the small GTPase rho with a 68-kDa phosphatidylinositol 4-phosphate 5-kinase in Swiss 3T3 cells. Mol. Biol. Cell, 7, 435-442.

Renkonen, O., Gahmberg, C.G., Simons, K. and Kaariainen, L. (1972) The lipids of the plasma membranes and endoplasmic reticulum from cultured baby hamster kidney cells (BHK21). Biochim. Biophys. Acta, 255, 66-78.

Rhee, S.G. and Bae, Y.S. (1997) Regulation of phosphoinositide-specific phospholipase C isozymes. J. Biol. Chem.,112, 15045-15048.

Rhee, S.G., Suh, P.-G., Ryu, S.H. and Lee, S.Y. (1989) Studies of inositol phospholipid-specific phospholipase C. Science, 244, 546-550.

Ridley, A.J. and Hall, A. (1992) The small GTP-binding protein rho regulates the assembly of focal adhesions and actin stress fibers in response to growth factors. Cell, 70, 389-399.

Riese, D.J. and Stem, D.F. (1998) Specificity within the EGF family/ErbB receptor family signaling network. BioEssays, 20, 41-48.

Rizo, J. and Sudhof, T.C. (1998) C2-domains, structure and function of a universal Ca^^-binding domain.J. Biol. Chem., 273, 15879-15882.

Ronnstrand, L., Mori, S., Arridsson, A.-K., Eriksson, A., Wemstedt, C., Heilman, U., Claesson-Welsh, L. and Heldin, C.-H. (1992) Identification of two C-terminal autophosphorylation sites in the PDGF P-recepor: involvement in the interaction with phospholipase C-y. EMBOJ., 11, 3911-3919.

Rosen, L.B. and Greenberg, M.E. (1996) Stimulation of growth factor receptor signal transduction by activation of voltage-sensitive calcium channels. Proc. Natl. Acad. Sci. USA, 93, 1113-1118.

Ross, C.A., MacCumber, M.W., Glatt, C.E. and Snyder, S.H. (1989) Brain phospholipase C isozymes: Differential mRNA localizations by in situ hybridization. Proc. Natl. Acad. Sci. USA, 86, 2923-2927.

Rothbacher, U., Laurent, M.N., Deardorff, M.A., Klein, P.S., Cho, K.W. and Fraser, S.E. (2000) Dishevelled phosphorylation, subcellular localization and multimerization regulate its role in early embryogenesis. EA/BO J., 19, 1010-1022.

Rozenvayn, N. and Flaumenhaft, R. (2001) Phosphatidylinositol 4,5-bisphosphate mediates Ca^+-induced platelet a-granule secretion: evidence for type 11 phosphatidylinositol 5-phosphate 4-kinase function. J. Biol. Chem., 276, 22410-22419.

214 Rudd, C.E. (1999) Adaptors and molecular scaffolds in immune cell signaling. Cell, 96, 5-8.

Ryu, S.H., Cho, K.S., Lee, K.Y., Suh, P.G. and Rhee, S.G. (1987) Purification and characterization of two immunologically distinct phosphoinositide-specific phospholipases C from bovine brain. J. Biol. Chem., 262, 12511-12518.

Sahly, I., Bar-Nachum, S., Suss-Toby, E., Rom, A, and Peretz, A. (1992) Calcium channel blockers inhibit retinal degeneration in the retinal degeneration-B mutant of Drosophila. Proc. Natl. Acad. Sci. USA, 89, 435-439.

Sakane, F., Yamada, K., Imai, S. and Kanoh, H. (1991) Porcine 80-kDa diacylglycerol kinase is a calcium- binding and calcium/phospholipid-dependent enzyme and undergoes calcium-dependent translocation. J. Biol. Chem., 266, 7096-7100.

Salama, S.R., Cleves, A.E., Malehom, D.E., Whitters, E.A. and Bankaitis, V.A. (1990) Cloning and characterization of Kluyveromyces lactis SEC14, a gene whose product stimulates Golgi secretory function in Saccharomyces cerevisiae. J. Bacterial., 172, 4510-4521.

Salardi, A., Caffrey, J.J., Snyder, S.H. and Shears, S.B. (2000) Inositol polyphosphate multikinase (ArgRIII) determines nuclear mRNA export in Saccharomyces cerevisiae. FEBS Lett., 468, 28-32.

Sambrook, J., Fritsch, E.F. and Maniatis, T. (1989)Molecular Cloning: A Laboratory Manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor.

Sato, Y., Arai, H., Miyata, A., Tokita, S., Yamamoto, K., Tanabe, T. and Inoue, K. (1993) Primary structure of a-tocopherol transfer protein from rat liver. Homology with cellular retinaldehyde protein. J. Biol. Chem., 26^, 17705-17710.

Sbrissa, D., Ikonomov, O.C. and Shisheva, A. (1999) PlKfyve, a mammalian ortholog of yeast Fab Ip lipid kinase, synthesizes 5-phosphoinositides. J. Biol. Chem., 274, 21589-21597.

Sbrissa, D., Ikonomov, O.C. and Shisheva, A. (2000) PlKfyve lipid kinase is a protein kinase: downregulation of 5'-phosphoinositide product formation by autophosphorylation. Biochemistry, 39, 15980- 15989.

Sbrissa, D., Ikonomov, O.C. and Shisheva, A. (2002) Phosphatidylinositol 3-phosphate-interacting domains in PlKfyve. J. Biol. Chem., 277, 6073-6079.

Schaap, D., van, d.W.J., van, B.W., van, d.B.R. and Ploegh, H.L. (1993) Diacylglycerol kinase is phosphorylated in vivo upon stimulation of the epidermal growth factor receptor and serine/threonine kinases, including protein kinase C-e. Biochem. J., 289, 875-881.

Schein, C.H. (1989) Production of soluble recombinant proteins in bacteria. BioTechnology, 7, 1141-1147.

Schein, C.H. and Notebom, M.H.M. (1988) Formation of soluble recombinant proteins in E. coli is favored by lower growth temperature. BioTechnology, 6,291-294.

Schermoly, M.J. and Helmkamp, G.M. (1983) The inactivity of brain phospholipid transfer protein toward phosphatidylinositol 4-phosphate. Brain Res., 268, 197-200.

Schmidt, M., Evellin, S., Weemink, P.A.O., vom Dorp, F., Rehmann, H., Lomasney, J.W. and Jakobs, K.H. (2001) A new phospholipase-C-calcium signalling pathway mediated by cyclic AMP and a Rap GTPase. Nat. Cell Biol, Z, 1020-1024.

Schouten, A., Agianian, B., Westerman, J., Kroon, J., Wirtz, K.W.A. and Gros, P. (2002) Structure o f apo- phosphatidylinositol transfer protein a provides insight into membrane association. EMBO J., 21, 2117- 2121 .

Schu, P.V., Takegawa, K., Fry, M.J., Stack, J.H., Waterfield, M.D. and Emr, S.D. (1993) Phosphatidylinositol 3-kinase encoded by yeast VPS34 gene essential for protein sorting. Science, 260, 88- 91.

Scott, G.K., Dodson, J.M., Montgomery, P.A., Johnson, R.M., Sarup, J.C., Wong, W.L., Ullrich, A., Shepard, H.M. and Benz, C.C. (1991) p i85™^^ signal transduction in breast cancer cells.J. Biol Chem., 266, 14300-14305.

215 Scott, K. and Zuker, C.S. (1998) Assembly of the Drosophila phototransduction cascade into a signalling complex shapes elementary responses. Nature, 395, 805-808.

Segui, B., Allen-Baume, V. and Cockcroft, S. (2002) Phosphatidylinositol transfer protein p displays minimal sphingomyelin transfer activity and is not required for biosynthesis and trafficking of sphingomyelin. Biochem. J., in the press.

Seidel-Dugan, C., Meyer, B.E., Thomas, S.M. and Brugge, J.S. (1992) Effects of SH2 and SH3 deletions on the functional activities of wild-type and transforming variants of c-Src. Mol. Cell Biol, 12, 1835-1845.

Serth, J., Weber, W., Freeh, M., Wittinghofer, A. and Pingoud, A. (1992) Binding of the H-ras p21 GTPase activating protein by the activated epidermal growth factor receptor leads to inhibition of the p21 GTPase activity in vitro. Biochemistry, 31, 6361-6365.

Sha, B. and Luo, M. (1999) PI transfer protein: the specific recognition of phospholipids and its functions. Biochim. Biophys. Acta, 1441, 268-277.

Sha, B., Phillips, S.E., Bankaitis, V.A. and Luo, M. (1998) Crystal structure of the Saccharomyces cerevisiae phosphatidylinositol transfer protein. Nature, 391, 506-510.

Shibasaki, Y., Ishihara, H., Kizuki, N., Asano, T., Oka, Y. and Yazaki, Y. (1997) Massive actin polymerization induced by phosphatidylinositol-4-phosphate 5-kinase in vivo. J. Biol Chem., 272, 7578- 7581.

Shibatohge, M., Kariya, K., Liao, Y., Hu, C.-D., Watari, Y., Goshima, M., Shima, F. and Kataoka, T. (1998) Identification of PLC210, a Caenorhabditis elegans phospholipase C, as a putative effector of Ras. J. Biol Chem., 273, 6218-6222.

Shieh, B.H. and Zhu, M.Y. (1996) Regulation of the TRP Ca^'*’ channel by INAD in Drosophila photoreceptors. Neuron, 16, 991-998.

Shisheva, A., Rusin, B., Ikonomov, O.C., DeMarco, C. and Sbrissa, D. (2001) Localization and insulin- regulated relocation of phosphoinositide 5-kinase PlKfyve in 3T3-L1 adipocytes. J. Biol Chem., 276, 11859-11869.

Shisheva, A., Sbrissa, D. and Ikonomov, O. (1999) Cloning, characterization, and expression of a novel Zn^^-binding FYVE finger-containing phosphoinositide kinase in insulin-sensitive cells. Mol Cell Biol, 19, 623-634.

Simon, J.-P., Morimoto, T., Bankaitis, V.A., Gottlieb, T.A., Ivanov, I.E., Adesnik, M. and Sabatini, D.D. (1998) An essential role for the phosphatidylinositol transfer protein in the scission of coatomer-coated vesicles fi*om the trans-Golgi network. Proc. Natl Acad. Scl USA, 95, 11181-11186.

Simonsen, A., Wurmser, A.E., Emr, S.D. and Stenmark, H. (2001) The role of phosphoinositides in membrane transport. Curr. Opin. Cell Biol, 13, 485-492.

Skinner, H.B., Alb, J.G., Whitters, E.A., Helmkamp, G.M. and Bankaitis, V.A. (1993) Phospholipid transfer activity is relevant to but not sufficient for the essential function of the yeast SEC 14 gene product. EMBOJ., 12, 4775-4784.

Skinner, H.B., McGee, T.P., McMaster, C.R., Fry, M.R., Bell, R.M. and Bankaitis, V.A. (1995) The Saccharomyces cerevisiae phosphatidylinositol-transfer protein effects a ligand-dependent inhibition of choline-phosphate cytidylyltransferase activity. Proc. Natl Acad. Scl USA, 92, 112-116.

Skippen, A., Jones, D.H., Morgan, C.P., Li, M. and Cockcroft, S. (2002) Mechanism of ADP ribosylation factor-stimulated phosphatidylinositol 4,5-bisphosphate synthesis in HL60 cells. J. Biol. Chem., 277, 5823- 5831.

Smith, C.D. and Chang, K.J. (1989) Regulation of brain phosphatidylinositol-4-phosphate kinase by GTP analogues. A potential role for guanine nucleotide regulatory proteins. J. Biol Chem., 264, 3206-3210.

Smith, M.R., Liu, Y.-L., Kim, S.R., Bae, Y.S., Kim, C.G., Kwon, K.-S., Rhee, S.G. and Kung, H.-F. (1996) PLCyl Src homology domain induces mitogenesis in quiescent NIH 3T3 fibroblasts. Biochem. Biophys. Res. Commun., 222, 186-193.

216 Smith, M.R., Liu, Y.-L., Matthews, N.T., Rhee, S.G., Sung, W.K. and Kung, H.-F. (1994) Phospholipase C- yl can induce DNA synthesis by a mechanism independent of its lipase activity. Proc. Natl. Acad. Sci. USA, 91, 6554-6558.

Snoek, G.T., Berrie, C.P., Geijtenbeek, T.B.H., van der Helm, H.A., Cadee, J.A., Lurisci, C., Corda, D. and Wirtz, K.W.A. (1999) Overexpression of phosphatidylinoditol transfer protein a in NIH3T3 cells activates a phospholipase A. J. Biol. Chem., 274, 35393-35399.

Snoek, G.T., Westerman, J., Wouters, F.S. and Wirtz, K.W. (1993) Phosphorylation and redistribution of the phosphatidylinositol-transfer protein in phorbol 12-myristate 13-acetate- and bombesin-stimulated Swiss mouse 3T3 fibroblasts. Biochem. J., 291, 649-656.

Sohn, R.H. and Goldschmidt-Clermont, P.J. (1994) Profilin: at the crossroads of signal transduction and the actin cytoskeleton. Bioessays, 16,465-472.

Soler, C., Beguinot, L. and Carpenter, G. (1994) Individual epidermal growth factor receptor autophosphorylation sites do not stringently define association motifs for several SH2-containing proteins. J. Biol. Chem., 269, 12320-12324.

Somerharju, P., Van Paridon, P. and Wirtz, K.W. (1983) Phosphatidylinositol transfer protein from bovine brain. Substrate specificity and membrane binding properties. Biochim. Biophys. Acta, 731, 186-195.

Song, C., Hu, C.-D., Masago, M., Kariya, K., Yamawaki-Kataoka, Y., Shibatohge, M., Wu, D., Satoh, T. and Kataoka, T. (2001a) Regulation of a novel human phospholipase C, PLCs, through membrane targeting by Ras. J. Biol. Chem., 276, 2752-2757.

Song, X., Xu, W., Zhang, A., Huang, G., Liang, X., Virbasius, J.V., Czech, M.P. and Zhou, G.W. (2001b) Phox homology domains specifically bind phosphatidylinositol phosphates. Biochemistry, 40, 8940-8944.

Songyang, Z., Shoelson, S.E., Chaudhuri, M., Gish, G., Pawson, T., Haser, W.G., King, F., Roberts, T., Ratnofsky, S., Lechleider, R.J., Neel, B.G., Birge, R.B., Fajardo, I.E., Chou, M.M., Hanafrisa, H., Schaffhausen, B. and Cantley, L.C. (1993) SH2 domains recognise specific phosphopeptide sequences. Cell, 72, 767-778.

Sorkin, A., Waters, C., Overholser, K.A. and Carpenter, G. (1991) Multiple autophosphorylation site mutations of the epidermal growth factor receptor. Analysis of kinase activity and endocytosis. J. Biol. Chem., 266, 8355-8362.

Sreenivas, A., Patton-Vogt, J.L., Bruno, V., Griac, P. and Henry, S.A. (1998) A role for phospholipase D (Pldlp) in growth, secretion, and regulation of membrane lipid synthesis in yeast. J. Biol. Chem., 273, 16635-16638.

Stark, W.S., Chen, D.-M., Johnson, M.A. and Frayer, K.L. (1983) The rdgB gene o f Drosophila: retinal degeneration in different alleles and inhibition by norpA. J. Insect Physiol, 29, 123-131.

Stenmark, H., Aasland, R., Toh, B.H. and D'Arrigo, A. (1996) Endosomal localization of the autoantigen EEAl is mediated by a zinc-binding FYVE finger. J. Biol Chem., 271, 24048-24054.

Stephens, L., Jackson, T.R. and Hawkins, P.T. (1993) Activation of phosphatidylinositol 4,5-bisphosphate supply by agonists and non-hydrolysable GTP analogues. Biochem. J., 296, 481-488.

Stephens, L.R., Hughes, K.T. and Irvine, R.F. (1991) Pathway of phosphatidylinositol(3,4,5)-trisphosphate synthesis in activated neutrophils. Nature, 351, 33-39.

Stock, S.D., Hama, H., DeWald, D.B. and Takemoto, J.Y. (1999) S'ÆC/4-dependent secretion in Saccharomyces cerevisiae. Nondependence on sphingolipid synthesis-coupled diacylglycerol production. J. Biol Chem., 274, 12979-12983.

Stover, D R., Becker, M., Liebetanz, J. and Lydon, N.B. (1995) Src phosphorylation of the epidermal growth factor receptor at novel sites mediates receptor interaction with Src and p85a.J. Biol Chem., 270, 15591-15597.

Stryjek-Kaminska, D., Piiper, A. and Zeuzem, S. (1996) Epidermal growth factor regulates adenylate cyclase activity via Gj and Gjl-2 proteins in pancreatic acinar membranes. Biochem. J., 316, 87-91.

217 Sugimoto, S., Wandless, T.J., Shoelson, S.E., Neel, B.G. and Walsh, C.T. (1994) Activation of the SH2- containing protein tyrosine phosphatase, SH-PTP2, by phosphotyrosine-containing peptides derived from insulin receptor substrate-1. J. Biol. Chem., 269, 13614-13622.

Sun, H., Seyer, J.M. and Patel, T.B. (1995) A region in the cytosolic domain of the epidermal growth factor receptor antithetically regulates the stimulatory and inhibitory guanine nucleotide-binding regulatory proteins of adenylyl cyclase. Proc. Natl. Acad. Sci. USA, 92, 2229-2233.

Sutton, R.B., Davletov, B.A., Berghuis, A.M., Sudhof, T.C. and Sprang, S.R. (1995) Structure of the first Cj domain of synaptotagmin I: A novel Ca^Vphospholipid-binding fold. Cell, 80, 929-938.

Suzuki, E. and Hirosawa, K. (1994) Immunolocalization of a Drosophila phosphatidylinositol transfer protein (rdgB) in normal and rdgA mutant photoreceptor cells with special reference to the subrhabdomeric cistemae. J. Electron Microsc., 43, 183-189.

Takahashi, S., Katagiri, T., Hirayama, T., Yamaguchi-Shinozaki, K. and Shinozaki, K. (2001) Hyperosmotic stress induces a rapid and transient increase in inositol 1,4,5-trisphosphate independent of abscisic acid in Arabidopsis cell culture. Plant Cell Physiol, 42, 214-222.

Takemoto, Y., Furuta, M., Sato, M., Kubo, M. and Hashimoto, Y. (1999) Isolation and characterization of a novel HSl SH3 domain binding protein, HS1BP3. Int. Immunol, 11, 1957-1964.

Tanaka, O. and Kondo, H. (1994) Localization of mRNA for three novel members ((33, (34 and y2) of phospholipase C family in mature rat brain. Neurosci Lett., 182, 17-20.

Tanaka, S. and Hosaka, K. (1994) Cloning of a cDNA encoding a second phosphatidylinositol transfer protein of rat brain by complementation of the yeast secI4 mutation. J. Biochem., 115, 981-984.

Tanaka, S., Yamashita, S. and Hosaka, K. (1995) Cloning and expression of human cDNA encoding phosphatidylinositol transfer protein p. Biochim. Biophys. Acta, 1259, 199-202.

Tang, W., Bunting, M., Zimmerman, G.A., McIntyre, T.M. and Prescott, S.M. (1996) Molecular cloning of a novel human diacylglycerol kinase highly selective for arachidonate-containing substrates. J. Biol Chem., 271, 10237-10241.

Tanner, K.G. and Kyte, J. (1999) Dimerization of the extracellular domain of the receptor for epidermal growth factor containing the membrane-spanning segment in response to treatment with epidermal growth factor. J. Biol. Chem., 274, 35985-35990.

Teitelbaum, I., Strasheim, A. and Berl, T. (1990) Epidermal growth factor-stimulated phosphoinositide hydrolysis in cultured rat inner medullary collecting tubule cells. Regulation by G protein, calcium, and protein kinase C.J Clin Invest, 85, 1044-1050.

Terui, T., Kahn, R.A. and Randazzo, F.A. (1994) Effects of acid phospholipids on nucleotide exchange properties of ADP-ribosylation factor 1. Evidence for specific interaction with phosphatidylinositol 4,5- bisphosphate. J. Biol. Chem., 269, 28130-28135.

Thomas, G.M., Cunningham, E., Fensome, A., Ball, A., Totty, N.F., Truong, O., Hsuan, J.J. and Cockcroft, S. (1993) An essential role for phosphatidylinositol transfer protein in phospholipase C-mediated inositol lipid signaling. Cell, 74, 919-928.

Thomas, G.M.H., Cunningham, E. and Cockcroft, S. (1994) Purification of phosphatidylinositol transfer protein from brain cytosol for reconstituting G-protein-regulated phosphoinositide-specific phospholipase C-(3 isozymes. Methods Enzymol, 238, 168-181.

Thomas, G.M.H., Geny, B. and Cockcroft, S. (1991) Identification of a novel cytosolic poly­ phosphoinositide-specific phospholipase C (PLC-86) as the major G-protein-regulated enzyme. EMBO J., 10,2507-2512.

Thomas, S.M. and Brugge, J.S. (1997) Cellular functions regulated by Src family kinases. Annu. Rev. Cell Dev. Biol, 13, 513-609.

Thompson, D.M., Cochet, C., Chambaz, E.M. and Gill, G.N. (1985) Separation and characterization of a phosphatidylinositol kinase activity that co-purifies with the epidermal growth factor receptor. J. Biol Chem., 260, 8824-8830.

218 Toker, A. (2002) Phosphoinositides and signal transduction. Cell. Mol. Life Set., 59, 761-779.

Toker, A. and Cantley, L.C. (1997) Signalling through the lipid products of phosphoinositide-3-OH kinase. Nature, 387, 673-676.

Tolias, K.F., Cantley, L.C. and Carpenter, C.L. (1995) Rho family GTPases bind to phosphoinositide kinases. J. Biol. Chem., 270, 17656-17659.

Tolias, K.F., Couvillon, A.D., Cantley, L.C. and Carpenter, C.L. (1998a) Characterization of a Rac- and RhoGDI-associated lipid kinase signalling complex. Mol. Cell. Biol, 18, 762-770.

Tolias, K.F., Hartwig, J.H., Ishihara, H., Shibasaki, Y., Cantley, L.C. and Carpenter, C.L. (2000) Type la phosphatidylinositol-4-phosphate 5-kinase mediates Rac-dependent actin assembly. Curr. Biol, 10, 153- 156.

Tolias, K.F., Rameh, L.E., Ishihara, H., Shibasaki, Y., Chen, J., Prestwich, G.D., Cantley, L.C. and Carpenter, C.L. (1998b) Type I phosphatidylinositol-4-phosphate 5-kinases synthesize the novel lipids phosphatidylinositol 3,5-bisphosphate and phosphatidylinositol 5-phosphate. J. Biol Chem., 273, 18040- 18046.

Touhara, K., Inglese, J., Pitcher, J.A., Shaw, G. and Lefkowitz, R.J. (1994) Binding of G protein Py- subunits to pleckstrin homology domains. J. Biol Chem., 269, 10217-10220.

Tremblay, J.M., Helmkamp, G.M. and Yarbrough, L.R. (1996) Limited proteolysis of rat phosphatidylinositol transfer protein by trypsin cleaves the C terminus, enhances binding to lipid vesicles, and reduces phospholipid transfer activity. J. Biol Chem., 271, 21075-21080.

Tremblay, J.M., Voziyan, P.A., Helmkamp, G.M. and Yarbrough, L.R. (1998) The C-terminus of PITP modulates membrane interactions and transfer activity but not phospholipid binding. Biochim. Biophys. Acta, 1389,91-100.

Trotter, P.J., Wu, W.-I., Pedretti, J., Yates, R. and Voelker, D.R. (1998) A genetic screen for aminophospholipid transport mutants identifies the phosphatidylinositol 4-kinase, Stt4p, as an essential component in phosphatidylserine metabolism. J. Biol Chem., 273, 13189-13196.

Trub, T., Choi, W.E., Wolf, G., Ottinger, E., Chen, Y., Weiss, M. and Shoelson, S.E. (1995) Specificity of the PTB domain of She for P turn-forming pentapeptide motifs amino-terminal to phosphotyrosine. J. Biol Chem.,Ills, 18205-18208.

Tsai, W., Morielli, A.D. and Peralta, E.G. (1997) The ml muscarinic acetylcholine receptor transactivates the EGF receptor to modulate ion channel activity. EMBOJ., 16, 4597-4605.

Tsujishita, Y. and Hurley, J.H. (2000) Structure and lipid transport mechanism of a StAR-related domain. Nat. Struct. Biol, 7, 404-414.

Tsunoda, S., Sierralta, J., Sun, Y., Bodner, R., Suzuki, E., Becker, A., Socolich, M. and Zuker, C.S. (1997) A multivalent PDZ-domain protein assembles signalling complexes in a G-protein-coupled cascade. Nature, 388, 243-249.

Tzahar, E., Pinkas-Kramarski, R., Moyer, J.D., Klapper, L.N., Alroy, I., Levkowitz, G., Shelly, M., Henis, S., Eisenstein, M., Ratzkin, B.J., Sela, M., Andrews, G.C. and Yarden, Y. (1997) Bivalence of EGF-like ligands drives the ErbB signalling network. EMBOJ., 16, 4938-4950.

Uchida, T., Matozaki, T., Noguchi, T., Yamao, T., Horita, K., Suzuki, T., Fujioka, Y., Sakamoto, C. and Kasuga, M. (1994) Insulin stimulates the phosphorylation of Tyr^^* and the catalytic activity of FTP 1C, a protein tyrosine phosphatase with Src homology-2 domains. J. Biol Chem., 269, 12220-12228.

Ullrich, A. and Schlessinger, J. (1990) Signal transduction by receptors with tyrosine kinase activity. Cell, 61,203-212.

Urumow, T. and Wieland, O H. (1988) Evidence for a cholera-toxin-sensitive G-protein involved in the regulation of phosphatidylinositol 4-phosphate kinase of rat liver membranes. Biochim. Biophys. Acta, 972, 232-238.

219 Urumow, T, and Wieland, O.H. (1990) Purification and partial characterization of phosphatidylinositol-4- phosphate kinase from rat liver plasma membranes. Further evidence for a stimulatory G-protein. Biochim. Biophys. Acta, 1052, 152-158.

Valius, M., Bazenet, C. and Kazlauskas, A. (1993) Tyrosines 1021 and 1009 are phosphorylation sites in the carboxy terminus of the platelet-derived growth factor receptor p subunit and are required for binding of phospholipase C y and a 64-kD protein, respectively. Mol. Cell. Biol, 13, 133-143. van der Bend, R.L., de Widt, J., Hilkmann, H. and van Blitterswijk, W.J. (1994) Diacylglycerol kinase in receptor-stimulated cells converts its substrate in a topologically restricted manner. J. Biol Chem., 269, 4098-4102. van der Geer, P., Wiley, S., Gish, G.D., Lai, V.K., Stephens, R., White, M.F., Kaplan, D. and Pawson, T. (1996) Identification of residues that control specific binding of the She phosphotyrosine-binding domain to phosphotyrosine sites. Proc. Natl Acad. Scl USA, 93, 963-968. van Horck, F.P., Lavazais, E., Eickholt, B.J., Moolenaar, W.H. and Divecha, N. (2002) Essential role of type la phosphatidylinositol 4-phosphate 5-kinase in neurite remodeling. Curr. Biol, 12, 241-245. van Paridon, P.A., Gadella, T.W.J., Somerharju, P.J. and Wirtz, K.W.A. (1987a) On the relationship between the dual specificity of the bovine brain phosphatidylinositol transfer protein and membrane phosphatidylinositol levels. Biochim. Biophys. Acta, 903, 68-77. van Paridon, P.A., Gadella, T.W.J., Somerharju, P.J. and Wirtz, K.W.A. (1988a) Properties of the binding sites for the sn-1 and sn-2 acyl chains on the phosphatidylinositol transfer protein from bovine brain. Biochemistry, 27, 6208-6214. van Paridon, P.A., Gadella, T.W.J. and Wirtz, K.W.A. (1988b) The effect of polyphosphoinositides and phosphatidic acid on the phosphatidylinositol transfer protein from bovine brain: a kinetic study. Biochim. Biophys. Acta, 943, 76-86. van Paridon, P.A., Visser, A.J.W.G. and Wirtz, K.W.A. (1987b) Binding of phospholipids to the phosphatidylinositol transfer protein from bovine brain as studied by steady-state and time-resolved fluorescence spectroscopy. Biochim. Biophys. Acta, 898, 172-180. van Tiel, C.M., Luberto, C., Snoek, G.T., Hannun, Y.A. and Wirtz, K.W.A. (2000a) Rapid replenishment of sphingomyelin in the plasma membrane upon degradation by sphingomyelinase in NIH3T3 cells overexpressing the phosphatidylinositol transfer protein |3. Biochem. J., 346, 537-543. van Tiel, C.M., Westerman, J., Paasman, M., Wirtz, K.W.A. and Snoek, G.T. (2000b) The protein kinase C- dependent phosphorylation of serine 166 is controlled by the phospholipid species bound to the phosphatidylinositol transfer protein a. J. Biol Chem., 275, 21532-21538. van Tiel, C.M., Westerman, J., Paasman, M.A., Hoebens, M.M., Wirtz, K.W.A. and Snoek, G.T. (2002) The Golgi localization of phosphatidylinositol transfer protein (3 requires the protein kinase C-dependent phosphorylation of serine-262 and is essential for maintaining plasma membrane sphingomyelin levels. J. Biol Chem., 277, 22447-22452.

Vanhaesebroeck, B., Leevers, S.J., Panayotou, G. and Waterfield, M.D. (1997) Phosphoinositide 3-kinases: a conserved family of signal transducers. Trends Biochem. Scl, 22, 267-272.

Vann, L.R., Wooding, F.B.P., Irvine, R.F. and Divecha, N. (1997) Metabolism and possible compartmentalization of inositol lipids in isolated rat-liver nuclei. Biochem. J., 327, 569-576.

Vamai, P. and Balia, T. (1998) Visualization of phosphoinositides that bind pleckstrin homology domains: calcium- and agonist-induced dynamic changes and relationship to myo-[^H] inositol-labeled phosphoinositide pools. J. Cell Biol, 143, 501-510.

Vega, Q.C., Cochet, C., Filhol, O., Chang, C.P., Rhee, S.G. and Gill, G.N. (1992) A site of tyrosine phosphorylation in the C terminus of the epidermal growth factor receptor is required to activate phospholipase C. Mol Cell Biol, 12, 128-135.

Verveer, P.J., Wouters, F.S., Reynolds, A.R. and Bastiaens, P.I.H. (2000) Quantitative imaging of lateral ErbBl receptor signal propagation in the plasma membrane.Science, 290, 1567-1570.

220 Vihtelic, T.S., Goebl, M., Milligan, S., O'Tousa, J.E. and Hyde, D.R. (1993) Localization ofDrosophila retinal degeneration B, a membrane-associated phosphatidylinositol transfer protein. J. Cell Biol, 122, 1013-1022.

Vihtelic, T.S., Hyde, D.R. and O'Tousa, J.E. (1991) Isolation and characterization of the Drosophilaretinal degeneration B (rdgB) gene. Genetics, 127, 761-768.

Vogel, S. and Hoppe, J. (1986) Polyamines stimulate the phosphorylation of phosphatidylinositol in membranes from A431 cells. Eur. J. Biochem., 154, 253-257.

Voziyan, P.A., Tremblay, J.M., Yarbrough, L.R. and Helmkamp, G.M., Jr. (1996) Truncations of the C- terminus have different effects on the conformation and activity of phosphatidylinositol transfer protein. Biochemistry, 35, 12526-12531.

Wahl, M.I., Jones, G.A., Nishibe, S., Rhee, S.G. and Carpenter, G. (1992) Growth factor stimulation of phospholipase C-yl activity. Comparative properties of control and activated enzymes. J. Biol Chem., 267, 10447-10456.

Wahl, M.I., Nishibe, S., Kim, J.W., Kim, H., Rhee, S.G. and Carpenter, G. (1990) Identification of two epidermal growth factor-sensitive tyrosine phosphorylation sites of phospholipase C-y in intact HSC-1 cells. J. Biol Chem., 265, 3944-3948.

Walch-Solimena, C. and Novick, P. (1999) The yeast phosphatidylinositol-4-OH kinase Pikl regulates secretion at the Golgi. Nat. Cell Biol, 1, 523-525.

Walker, D.H., Dougherty, N. and Pike, L.J. (1988) Purification and characterization of a phosphatidyl inositol kinase from A431 cells. Biochemistry, 27, 6504-6511.

Walker, D.H. and Pike, L.J. (1987) Phosphatidylinositol kinase is activated in membranes derived from cells treated with epidermal growth factor. Proc. Natl Acad. Scl USA, 84, 7513-7517.

Walton, G.M., Chen, W.S., Rosenfeld, M.G. and Gill, G.N. (1990) Analysis of deletions of the carboxyl terminus of the epidermal growth factor receptor reveals self phosphorylation at tyrosine 992 and enhanced in vivo tyrosine phosphorylation of cell substrates. J. Biol Chem., 265, 1750-1754.

Walz, B., Ukhanov, K. and Zimmermann, B. (2000) Actions of neomycin on electrical light responses, Ca^^ release, and intracellular Ca^^ changes in photoreceptors of the honeybee drone. J. Comp. Physiol, 186, 1019-1029.

Wang, S., Gebre-Medhin, S., Betsholtz, C., Stalberg, P., Zhou, Y., Larsson, C., Weber, G., Feinstein, R., Oberg, K., Gobi, A. and Skogseid, B. (1998) Targeted disruption of the mouse phospholipase C |33 gene results in early embryonic lethality. FEBS Lett., 441, 261-265.

Wang, S., Yue, H., Derin, R.B., Guggino, W.B. and Li, M. (2000) Accessory protein facilitated CFTR- CFTR interaction, a molecular mechanism to potentiate the chloride channel activity. Cell, 103, 169-179.

Weisz, O.A., Gibson, G.A., Leung, S.-M., Roder, J. and Jeromin, A. (2000) Overexpression of frequenin, a modulator of phosphatidylinositol 4-kinase, inhibits biosynthetic delivery of an apical protein in polarized Madin-Darby canine kidney cells. J. Biol. Chem., 275, 24341-24347.

Wenk, M.R., Pellegrini, L., Klenchin, V.A., Di Paolo, G., Chang, S., Daniell, L., Arioka, M., Martin, T.F. and De Camilli, P. (2001) PIP kinase ly is the major PI(4,5)P2 synthesizing enzyme at the synapse. Neuron, 32, 79-88.

Wennstrom, S., Siegbahn, A., Yokote, K., Arvidsson, A.K., Heldin, C.H., Mori, S. and Claesson-Welsh, L. (1994) Membrane ruffling and chemotaxis transduced by the PDGF beta-receptor require the binding-site for phosphatudylinositol 3-kinase. Oncogene, 9, 651-660.

Westergren, T., Dove, S.K., Sommarin, M. and Pical, C. (2001) AtPIP5Kl, an Arabidopsis thaliana phosphatidylinositol phosphate kinase, synthesizes Ptdlns(3,4)?2 and PtdIns(4,5)P2 In vitro and is inhibited by phosphorylation. Biochem. J., 359, 583-589.

Westerman, J., de, V.K., Somerharju, P., Timmermans, H.J., Snoek, G.T. and Wirtz, K.W. (1995) A sphingomyelin-transferring protein from chicken liver. Use of pyrene-labeled phospholipid. J. Biol Chem., 270, 14263-14266.

221 Wetzker, R., Klinger, R., Hsuan, J., Fry, M.J., Kauffinann-Zeh, A., Muller, E., Frunder, H. and Waterfield, M, (1991) Purification and characterization of phosphatidylinositol 4-kinase from human erythrocyte membranes. Eur. J. Biochem., 200, 179-185.

Whatmore, J., Wiedemann, C., Somerharju, P., Swigart, P. and Cockcroft, S. (1999) Resynthesis of phosphatidylinositol in permeabilized neutrophils following phospholipase Cbeta activation: transport of the intermediate, phosphatidic acid, from the plasma membrane to the endoplasmic reticulum for phosphatidylinositol resynthesis is not dependent on soluble lipid carriers or vesicular transport. Biochem. J., 341, 435-444.

Whiteford, C.C., Brearley, C.A. and Ulug, E.T. (1997) Phosphatidylinositol 3,5-bisphosphate defines a novel PI 3-kinase pathway in resting mouse fibroblasts. Biochem. J., 323, 597-601.

Wiedemann, C., Schâfer, T. and Burger, M.M. (1996) Chromaffin granule-associated phosphatidylinositol 4-kinase activity is required for stimulated secretion. EMBOJ., 15, 2094-2101.

Wirtz, K.W. (1991) Phospholipid transfer proteins. Annu. Rev. Biochem., 60, 73-99.

Wirtz, K.W. (1997) Phospholipid transfer proteins revisited. Biochem. J., 324, 353-360.

Wong, K. and Cantley, L.C. (1994) Cloning and characterization of a human phosphatidylinositol 4-kinase. J. Biol. Chem., 269, 28878-28884.

Wong, K., Meyers, R. and Cantley, L.C. (1997) Subcellular locations of phosphatidylinositol 4-kinase isoforms. J. Biol. Chem., 272, 13236-13241.

Woscholski, R. and Parker, P.J. (1997) Inositol lipid 5-phosphatases - traffic signals and signal traffic. Trends Biochem. Sci., 22, 427-431.

Wu, W.-L, Routt, S., Bankaitis, V.A. and Voelker, D.R. (2000) A new gene involved in the transport- dependent metabolism of phosphatidylserine, PSTB2/PDR17, shares sequence similarity with the gene encoding the phosphatidylinositol/phosphatidylcholine transfer protein, SEC 14. J. Biol. Chem., 275, 14446- 14456.

Xie, W., Samoriski, G.M., McLaughlin, J.P., Romoser, V.A., Smrcka, A., Hinkle, P.M., Bidlack, J.M., Gross, R.A., Jiang, H. and Wu, D. (1999) Genetic alteration of phospholipase C (33 expression modulates behavioral and cellular responses to p opioids. Proc. Natl. Acad. Sci. USA, 96, 10385-10390.

Xie, Z., Fang, M. and Bankaitis, V.A. (2001) Evidence for an intrinsic toxicity of phosphatidylcholine to Secl4p-dependent protein transport from the yeast Golgi complex. Mol. Biol. Cell, 12, 1117-1129.

Xie, Z., Fang, M., Rivas, M.P., Faulkner, A.J., Stemweis, P.C., Engebrecht, J. and Bankaitis, V.A. (1998) Phospholipase D activity is required for suppression of yeast phosphatidylinositol transfer protein defects. Proc. Natl. Acad. Sci. USA, 95, 12346-12351.

Xu, Y., Hortsman, H., Sect, L.F., Wong, S.H. and Hong, W. (2001a) SNX3 regulates endosomal function through its PX-domain-mediated interaction with PtdIns(3)P. Nat. Cell Biol., 3, 658-666.

Xu, Y., Seet, L.-F., Hanson, B. and Hong, W. (2001b) The Phox homology (PX) domain, a new player in phosphoinositide signalling. Biochem. J., 360, 513-530.

Yagisawa, H., Sakuma, K., Paterson, H.F., Cheung, R., Allen, V., Hirata, H., Watanabe, Y., Hirata, M., Williams, R.L. and Katan, M. (1998) Replacements of single basic amino acids in the pleckstrin homology domain of phospholipase C-Ô1 alter the ligand binding, phospholipase activity, and interaction with the plasma membrane. J. Biol. Chem., 273, 417-424.

Yamamoto, A., DeWald, D.B., Boronenkov, I.V., Anderson, R.A., Emr, S.D. and Koshland, D. (1995) Novel PI(4)P 5-kinase homologue. Fab Ip, essential for normal vacuole function and morphology in yeast. Mol. Biol. Cell, 6, 525-539.

Yamamoto, K., Graziani, A., Carpenter, C., Cantley, L.C. and Lapetina, E.G. (1990) A novel pathway for the formation of phosphatidylinositol 3,4-bisphosphate. Phosphorylation of phosphatidylinositol 3- monophosphate by phosphatidylinositol 3-monophosphate 4-kinase. J. Biol. Chem., 265, 22086-22089.

Yamamoto, K. and Lapetina, E.G. (1990) Protein kinase C-mediated formation of phosphatidylinositol 3,4- bisphosphate in human platelets. Biochem. Biophys. Res. Commun., 168, 466-472.

222 Yamauchi, T., Ueki, K., Tobe, K., Tamemoto, H., Sekine, N., Wada, M., Honjo, M., Takahashi, M., Takahashi, T., Hirai, H., Tushima, T., Akanuma, Y., Fujita, T., Komuro, I., Yazaki, Y. and Kadowaki, T. (1997) Tyrosine phosphorylation of the EGF receptor by the kinase Jak2 is induced by growth hormone. Nature, 390, 91-96.

Yamazaki, M., Miyazaki, H., Watanabe, H., Sasaki, T., Maehama, T., Frohman, M.A. and Kanaho, Y. (2002) Phosphatidylinositol 4-phosphate 5-kinase is essential for ROCK-mediated neurite remodeling. J. Biol. Chem., Ill, 17226-17230.

Yang, L., Camoratto, A.M., Baffy, 0., Raj, S., Manning, D.R. and Williamson, J R. (1993) Epidermal growth factor-mediated signaling of G(i)-protein to activation of phospholipases in rat-cultured hepatocytes. J. Biol. Chem., 268, 3739-3746.

Yang, L.J., Baffy, G., Rhee, S.G., Manning, D., Hansen, C.A. and Williamson, J.R. (1991) Pertussis toxin- sensitive Gi protein involvement in epidermal growth factor-induced activation of phospholipase C-y in rat hepatocytes. J. Biol. Chem., 266, 22451-22458.

Yang, L.J., Rhee, S.G. and Williamson, J.R. (1994) Epidermal growth factor-induced activation and translocation of phospholipase C-yl to the cytoskeleton in rat hepatocytes. J. Biol. Chem., 269, 7156-7162.

Yarden, Y. and Sliwkowski, M.X. (2001) Untangling the ErbB signalling network. Nat. Rev. Mol. Cell Biol, 2, 127-137.

Ye, K., Aghdasi, B., Luo, H.R., Moriarity, J.L., Wu, F.Y., Hong, J.J., Hurt, K.J., Bae, S.S., Suh, P.G. and Snyder, S.H. (2002) Phospholipase C-yl is a physiological guanine nucleotide exchange factor for the nuclear GTPase PIKE. Nature, 415, 541-544.

Yoder, M.D., Thomas, L.M., Tremblay, J.M., Oliver, R.L., Yarbrough, L.R. and Helmkamp, G.M. (2001) Structure of a multifunctional protein: mammalian phosphatidylinositol transfer protein complexed with phosphatidylcholine. J. Biol. Chem., 276, 9246-9252.

York, J.D., Odom, A.R., Murphy, R., Ives, E.B. and Wente, S.R. (1999) A phospholipase C-dependent inositol polyphosphate kinase pathway required for efficient messenger RNA export. Science, 295, 96-100.

Yoshida, S., Ohya, Y., Goebl, M., Nakano, A. and Anraku, Y. (1994a) A novel gene, STT4, encodes a phosphatidylinositol 4-kinase in the PKCl protein kinase pathway of Saccharomyces cerevisiae. J. Biol Chem., 269, 1166-1172.

Yoshida, S., Ohya, Y., Nakano, A. and Anraku, Y. (1994b) Genetic interactions among genes involved in the STT4-PKC1 pathway oïSaccharomyces cerevisiae. Mol Gen. Genet., 242, 631-640.

Yu, H., Fukami, K., Watanabe, Y., Ozaki, C. and Takenawa, T. (1998) Phosphatidylinositol 4,5- bisphosphate reverses the inhibition of RNA transcription caused by histone HI. Eur. J. Biochem., 251, 281-287.

Zhang, X., Loijens, J.C., Boronenkov, I.V., Parker, G.J., Norris, F.A., Chen, J., Thum, O., Prestwich, G.D., Majerus, P.W. and Anderson, R.A. (1997) Phosphatidylinositol-4-phosphate 5-kinase isozymes catalyze the synthesis of 3-phosphate-containing phosphatidylinositol signaling molecules. J. Biol Chem., 272, 17756- 17761.

Zhao, K., Wang, W., Rando, O.J., Xue, Y., Swiderek, K., Kuo, A. and Crabtree, G.R. (1998) Rapid and phospholipid-dependent binding of the SWI/SNF-like BAF complex to chromatin after T lymphocyte receptor signalling. Cell, 95, 625-636.

Zhao, X., Vamai, P., Tuymetova, G., Balia, A., Toth, Z.E., Oker-Blom, C., Roder, J., Jeromin, A. and Balia, T. (2001) Interaction of neuronal calcium sensor-1 (NCS-1) with phosphatidylinositol 4-kinase (3 stimulates lipid kinase activity and affects membrane trafficking in COS-7 cells. J. Biol Chem., 276, 40183-40189.

Zhao, X.-H., Bondeva, T. and Balia, T. (2000) Characterization of recombinant phosphatidylinositol 4- kinase P reveals auto- and heterophosphorylation of the enzyme. J. Biol. Chem., 275, 14642-14648.

Zheng, Y., Glaven, J.A., Wu, W.J. and Cerione, R.A. (1996) Phsophatidylinositol 4,5-bisphosphate provides an alternative to gaunine nucleotide exchnage factors by stimulating the dissociation of GDP fi’om Cdc42Hs. J. Biol Chem., 271, 23815-23819.

223 Zhou, M.M., Ravichandran, K.S., Olejniczak, E.F., Petros, A.M., Meadows, R.P., Sattler, M., Harlan, J.E., Wade, W.S., Burakoff, S.J. and Fesik, S.W. (1995) Structure and ligand recognition of the phosphotyrosine binding domain of She. Nature, 378, 584-592.

Zimmermann, P., Meerschaert, K., Reekmans, G., Leenaerts, I., Small, J.V., Vandekerckhove, J., David, 0. and Gettemans, J. (2002) PIPj-PDZ domain binding controls the association of syntenin with the plasma membrane. Mol. Cell, 9, 1215-1225.

Zini, N., Martelli, A.M., Cocco, L., Manzoli, F.A. and Maraldi, N.M. (1993) Phosphoinositidase C isoforms are specifically localized in the nuclear matrix and cytoskeleton of Swiss 3T3 cells. Exptl. Cell Res., 208, 257-269.

Zini, N., Sabatelli, P., Faenza, I., Ognibene, A. and Maraldi, N.M. (1996) Interleukin la induces variations of the intranuclear amount of phosphatidylinositol 4,5-bisphosphate and phospholipase C pi in human osteosarcoma Saos-2 cells. Histochem. J., 28, 495-504.

Zuker, C.S. (1996) The biology of vision of Drosophila. Proc. Natl. Acad. Sci. USA, 93, 571-576.

Zwick, E., Daub, H., Aoki, N., Yamaguchi-Aoki, Y., Tinhofer, I., Maly, K. and Ullrich, A. (1997) Critical role of calcium-dependent epidermal growth factor receptor transactivation in PC 12 cell membrane depolarisation and bradykinin signaling. J. Biol. Chem., 272, 24767-24770.

224