MOLECULAR BIOLOGY INTELLIGENCE UNIT
MOLECULAR BIOLOGY INTELLIGENCE UNIT
Landes Bioscience, a bioscience publisher, is ...
183 downloads
1359 Views
3MB Size
Report
This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form
MOLECULAR BIOLOGY INTELLIGENCE UNIT
MOLECULAR BIOLOGY INTELLIGENCE UNIT
Landes Bioscience, a bioscience publisher, is making a transition to the internet as Eurekah.com.
Biotechnology Intelligence Unit Medical Intelligence Unit Molecular Biology Intelligence Unit Neuroscience Intelligence Unit Tissue Engineering Intelligence Unit
DEKKER
INTELLIGENCE UNITS
Lodewijk V. Dekker
Second Edition MBIU
The chapters in this book, as well as the chapters of all of the five Intelligence Unit series, are available at our website.
Second Edition
9 790306 478634
Protein Kinase C
I SBN 0- 306- 47863- 3
Protein Kinase C
MOLECULAR BIOLOGY INTELLIGENCE UNIT
Protein Kinase C Second Edition Lodewijk V. Dekker, M.D. Ionix Pharmaceuticals, Ltd. Cambridge, U.K. and Departments of Medicine and Biology University College London London, U.K.
LANDES BIOSCIENCE / EUREKAH.COM GEORGETOWN, TEXAS U.S.A.
KLUWER ACADEMIC / PLENUM PUBLISHERS NEW YORK, NEW YORK U.S.A.
PROTEIN KINASE C SECOND EDITION
Molecular Biology Intelligence Unit Landes Bioscience / Eurekah.com Kluwer Academic / Plenum Publishers Copyright ©2004 Eurekah.com and Kluwer Academic / Plenum Publishers All rights reserved. No part of this book may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopy, recording, or any information storage and retrieval system, without permission in writing from the publisher, with the exception of any material supplied specifically for the purpose of being entered and executed on a computer system; for exclusive use by the Purchaser of the work. Printed in the U.S.A. Kluwer Academic / Plenum Publishers, 233 Spring Street, New York, New York, U.S.A. 10013 www.wkap.nl/ Please address all inquiries to the Publishers: Landes Bioscience / Eurekah.com, 810 South Church Street, Georgetown, Texas, U.S.A. 78626 Phone: 512.863.7762; Fax: 512.863.0081 www.eurekah.com www.landesbioscience.com Protein Kinase C, 2nd Edition, edited by Lodewijk V. Dekker, Landes / Kluwer dual imprint / Landes series: Molecular Biology Intelligence Unit ISBN: 0-306-47863-3 While the authors, editors and publisher believe that drug selection and dosage and the specifications and usage of equipment and devices, as set forth in this book, are in accord with current recommendations and practice at the time of publication, they make no warranty, expressed or implied, with respect to material described in this book. In view of the ongoing research, equipment development, changes in governmental regulations and the rapid accumulation of information relating to the biomedical sciences, the reader is urged to carefully review and evaluate the information provided herein.
Library of Congress Cataloging-in-Publication Data CIP applied for but not received at time of publication.
CONTENTS Preface .................................................................................................. ix 1. Introduction ........................................................................................... 1 Lodewijk V. Dekker The Protein Phosphorylation Reaction .................................................. 1 Protein Kinases ...................................................................................... 1 Protein Kinase C ................................................................................... 2 Protein Kinase C Isotypes and Their Domain Structure ........................ 2 Defining the PKC Family ...................................................................... 4 This Book ............................................................................................. 5 2. Structure, Function and Membrane Interactions of C1 Domains ........... 8 Matthew Pearson and James H. Hurley C1 Domain Structure and Stereospecific Ligand Recognition ............... 9 Membrane Binding and Subcellular Localization of C1 Domains ....... 11 Typical C1 Domains in Proteins Other than PKC .............................. 12 Atypical C1 Domains .......................................................................... 12 C1 Domains in the Context of Multidomain Signaling Proteins ......... 13 3. Structural and Functional Specialization of C2 Domains in Protein Kinase C .............................................................................. 16 Eric A. Nalefski C2 Domain Structure .......................................................................... 17 C2 Domain Interactions with Ligands ................................................. 20 Mechanism of Membrane Targeting of PKC by C2 Domains ............. 25 4. Regulation of Protein Kinase C by Membrane Interactions ................. 36 Alexandra C. Newton Protein Kinase C: At the Heart of Lipid Signalling .............................. 36 The Three Cofactors: Phosphatidylserine, Diacylglycerol, and Ca2+ ..... 37 Translocation of Protein Kinase C ....................................................... 42 Phosphorylation: Regulator of Protein Kinase C’s Subcellular Location ........................................................................ 44 5. Protein Kinase C Regulation by Protein Interactions ........................... 49 Susan Jaken Examples of PKC Binding Proteins ..................................................... 51 Discussion ........................................................................................... 59 6. Regulation of Protein Kinase C by Phosphorylation ............................ 63 Alex Toker Phosphorylation at the Activation Loop .............................................. 64 Phosphorylation at the Turn Motif ..................................................... 67 Phosphorylation at the Hydrophobic Site ............................................ 68 The Role of the Phosphoinositide 3-Kinase Pathway ........................... 69 Regulation by Dephosphorylation ....................................................... 70 Implications for Cellular Function ...................................................... 71 Regulation of Protein Kinase C by Tyrosine Phosphorylation ............. 71
7. Down-Regulation of Protein Kinase C ................................................. 76 Nigel T. Goode and Nicola Smart The Proteolytic Mechanism ................................................................. 78 Role of Down-Regulation.................................................................... 82 8. Protein Kinase C in Yeast: The Cell Wall Integrity Pathway................ 87 Pilar Perez, Beatriz Santos and Pedro M. Coll Yeasts PKC Homologues ..................................................................... 87 Rho GTPases and PKC Activation ...................................................... 88 S. cerevisiae Pkc1p and the Cell Integrity Pathway ............................... 90 Cell Integrity Pathway and Cell Cycle ................................................. 91 Interaction with Other Pathways ......................................................... 93 Function of Pck1p and Pck2p in S. pombe Cell Integrity and Cell Wall Biosynthesis .............................................................. 93 9. Specificity in Atypical Protein Kinase C Signaling: κB Paradigm ........................................................................ 100 The NF-κ Jorge Moscat and María T. Diaz-Meco The NF-κB Pathway ......................................................................... 100 Adapters for the aPKCs to NF-κB Activation .................................... 103 Functional Relevance of the aPKCs in the Ras-NF-κB Crossroad ..... 106 Specificity during Cell Signaling ........................................................ 106 Blockade of NF-κB by the Endogenous aPKC Inhibitor Par-4 .......... 109 10. Protein Kinase C: A Molecular Information Storage Device in Neurons .............................................................................. 114 Laura A. Schrader, Coleen M. Atkins, Michael Leitges, J. David Sweatt and Edwin J. Weber Activity-Dependent Synaptic Plasticity .............................................. 114 Inhibitors of PKC Block LTP ............................................................ 116 Activating PKC Mimics LTP ............................................................ 117 Biochemical Studies........................................................................... 118 Oxidative Modification of PKC ........................................................ 118 Protein Kinase Substrates Phosphorylated during LTP ...................... 119 Cerebellar Long-Term Depression ..................................................... 121 PKC in Learning and Memory .......................................................... 121 Changes in Activity, Cellular Distribution and Phosphorylation of Substrates .................................................................................. 122 Pharmacological Inhibition of PKC ................................................... 123 Isotype Specific PKC Knockout ........................................................ 123
11. Involvement of PKC in the Sensation of Pain .................................... 134 Vittorio Vellani and Peter A. McNaughton Role of PKC in Peripheral Nociception ............................................. 136 Sensitization of the Heat-Gated Current by PKC .............................. 136 Desensitization of VR1 by Calcineurin .............................................. 138 Phosphorylation of VR1 .................................................................... 138 PKC Isotypes Involved in VR1 Phosphorylation ............................... 139 PKC-ε and Inflammatory Pain .......................................................... 139 Actions of PKC on Voltage-Sensitive Na Currents ............................ 140 Specific Involvement of PKC-ε in Neuropathic Pain ......................... 141 Conclusion: Peripheral Nociception .................................................. 141 Involvement of PKC in Central Pain Processing ................................ 141 Protein Kinase C-γ and Persistent Pain .............................................. 142 PKC Activity, Opioid Tolerance and Pathological Pain States Are Related .................................................................................... 142 12. The Protein Kinase C Gene Module: Cellular Functions in T Lymphocytes .............................................................................. 147 Gottfried Baier The PKC Kinases: A Gene Family of Nine Isotypes .......................... 147 Regulation of PKC Activity in T Cells ............................................... 149 Signaling Specificity of PKC in T Cells—A Complex Affair .............. 150 Role of PKC in T Lymphocyte Physiology ........................................ 153 PKC and T Cell Activation ............................................................... 153 T Cell Antigen Receptor .................................................................... 153 Integrin Receptors ............................................................................. 155 Cytokine Receptors ........................................................................... 156 Are There any Other Critical Pathways Modulated by PKC .............. 157 Role of PKC in Lymphocyte Development ....................................... 157 Role of PKC in Lymphocyte Homeostasis ......................................... 158 PKCs in Lymphocyte Disorders ........................................................ 159 Synopsis and Future Perspectives ....................................................... 159 13. Protein Kinase C Isotype Function in Neutrophils ............................. 165 Lodewijk V. Dekker Basic Cell Biological Properties of Neutrophils .................................. 165 PKC Isotypes in Neutrophils ............................................................. 168 Studying PKC Isotype Function in Neutrophils ................................ 171 Molecular Targets of PKC in Neutrophils ......................................... 173 Index .................................................................................................. 191
EDITOR Lodewijk V. Dekker, Ph.D. Ionix Pharmaceuticals, Ltd. Cambridge, U.K. and Departments of Medicine and Biology University College London London, U.K. Chapters 1, 13
CONTRIBUTORS Colleen M. Atkins Vollum Institute Oregon Health Sciences University Portland, Oregon, U.S.A.
Nigel T. Goode Veterinary Basic Sciences The Royal Veterinary College London, U.K.
Chapter 10
Chapter 7
Gottfried Baier Institute for Medical Biology and Human Genetics University of Innsbruck Innsbruck, Austria
James H. Hurley Laboratory of Molecular Biology National Institute of Diabetes and Digestive and Kidney Diseases National Institutes of Health U.S. Department of Health and Human Services Bethesda, Maryland, U.S.A.
Chapter 12
Pedro M. Coll Edificio Departamental Consejo Superior de Investigaciones Cientificas Instituto de Microbiologia Bioquímica Universidad de Salamanca Salamanca, Spain
Chapter 2
Susan Jaken Lilly Research Laboratories Lilly Corporate Center Indianapolis, Indiana, U.S.A.
Chapter 8
Chapter 5
María T. Diaz-Meco Centro de Biologia Molecular Severo Ochoa Consejo Superior de Investigaciones Cientificas Universidad Autónoma Madrid, Spain
Michael Leitges Max Planck Institute for Experimental Endocrinology Hannover, Germany
Chapter 9
Chapter 10
Peter A. McNaughton Department of Pharmacology University of Cambridge Cambridge, U.K. Chapter 11
Jorge Moscat Centro de Biologia Molecular Severo Ochoa Consejo Superior de Investigaciones Cientificas Universidad Autónoma Madrid, Spain Chapter 9
Eric A. Nalefski U.S. Genomics Woburn, Massachusetts, U.S.A
Laura A. Schrader Division of Neuroscience Baylor College of Medicine Houston, Texas, U.S.A. Chapter 10
Nicola Smart Molecular Medicine Unit Institute of Child Health University College London London, U.K. Chapter 7
Chapter 3
Alexandra C. Newton Department of Pharmacology University of California at San Diego La Jolla, California, U.S.A.
J. David Sweatt Division of Neuroscience Baylor College of Medicine Houston, Texas, U.S.A. Chapter 10
Chapter 4
Matthew Pearson Laboratory of Molecular Biology National Institute of Diabetes and Digestive and Kidney Diseases National Institutes of Health U.S. Department of Health and Human Services Bethesda, Maryland, U.S.A. Chapter 2
Pilar Perez Edificio Departamental Consejo Superior de Investigaciones Cientificas Instituto de Microbiologia Bioquímica Universidad de Salamanca Salamanca, Spain Chapter 8
Beatriz Santos Edificio Departamental Consejo Superior de Investigaciones Cientificas Instituto de Microbiologia Bioquímica Universidad de Salamanca Salamanca, Spain Chapter 8
Alex Toker Department of Pathology Beth Israel Deaconess Medical Center Harvard Medical School Boston, Massachusetts, U.S.A. Chapter 6
Vittorio Vellani Department of Pharmacology University of Cambridge Cambridge, U.K. Chapter 11
Edwin J. Weber Division of Neuroscience Baylor College of Medicine Houston, Texas, U.S.A. Chapter 10
PREFACE
A
t the time of publication of the first edition in 1997, Protein Kinase C was a well-established kinase and a significant body of knowledge had been generated on its structure, regulation and function. Since then the research has continued apace and this second edition builds on new developments in each of these areas. The continued analysis of the Protein Kinase C building blocks has provided novel insights into their structure and function and their relationship to structures in other proteins. This has greatly increased our understanding of the classic control mechanism of Protein Kinase C, its interaction with lipid membranes and cofactors. More recently, the regulation of Protein Kinase C by phosphorylation and protein interactions has received much attention, to the extent that these are now established regulatory principles. Genetic targeting of Protein Kinase C has contributed and will continue to contribute to our understanding of the role of this kinase in the living organism. Deciding on suitable timing for a follow-up edition was demanding and always has an arbitrary element. However the tremendous progress in the field makes this a good moment to capture some of the more recent research and to place it in the context of what has gone before. At the same time, the fast pace of research makes it impossible to cover all new data and for these, the reader is referred to the primary literature. It is my hope that this publication will be used as a reference work in order to place the current research in context and as a starting point to learn more about this fascinating kinase. I am grateful to all contributors without whom this project would not have been possible. Lodewijk V. Dekker, Ph.D.
CHAPTER 1
Introduction Lodewijk V. Dekker
T
he discovery of protein phosphorylation in the 1950s was a seminal event which started off a thriving area of research. These days it would be difficult, if not impossible to name any physiological function in which phosphorylation does not play a regulatory role. Recent evidence has revealed the importance of the phosphorylation process in the accurate control of the cell cycle, circadian rhythm, neurotransmission, immune response and countless other biological functions. The fact that a number of Nobel prizes have been awarded over the past years to researchers directly active in this field is a testament to the importance of the phosphorylation process. The application of our basic knowledge of protein phosphorylation events to medicine has resulted in the development of new, powerful anticancer drugs, and promises as much for the future of human health as for the future health of this field of research.
The Protein Phosphorylation Reaction The phosphorylation reaction comprises the transfer of a γ-phosphate from the phosphate donor (ATP) to the acceptor protein. In mammalian proteins, serine, threonine and tyrosine residues are the phosphate acceptor residues.1 Histidine phosphorylation occurs in prokaryotes and has also been identified in yeast, molds and plants.2 The phosphorylation event introduces charge into the acceptor protein resulting in its functional modification. In molecular terms, phosphorylation may result in conformational changes in the protein or may change its surface appearance. As such it may become active or inactive, act to recruit protein components into a functional complex or may disassociate from protein complexes. Protein kinases carry out the phosphotransfer reaction. Their catalytic domain contains a binding site for the phosphate donor ATP as well as a groove to accommodate the phosphate acceptor region of the protein substrate. These interactions allow correct juxtaposition of the γ-phosphate of ATP to the phosphoacceptor residue so that the phosphotransfer reaction can take place. Apart from their core catalytic domain, many kinases contain additional domains and motifs which determine the context in which their catalytic function is executed. These include auto-inhibitory elements, which keep the kinase in an inactive conformation, and functional modules which allow the interaction with lipid factors, second messengers or proteins. Such elements provide a significant control over the kinase and restrict its function to the environment where the controlling substance is present.
Protein Kinases Hundreds of different protein kinases have been identified and with the increase in our knowledge of the genome—both in humans and other species—the extent of the protein kinase ‘superfamily’ becomes clear.3 Protein kinases, as defined by the presence of a conserved protein kinase catalytic domain, comprise the second largest family of protein domains in the nematode worm Caenorhabditis elegans and a similarly significant family in mammalian Protein Kinase C, 2nd Edition, edited by Lodewijk Dekker. ©2004 Eurekah.com and Kluwer Academic / Plenum Publishers.
2
Protein Kinase C
species.4 The sheer number of kinases makes some form of classification desirable. Using the primary sequence conservation within the catalytic core region as a criterion, kinases have been divided in various groups.1,3 To a degree, this division matches certain functional properties of the kinases involved. For example, the large group of tyrosine kinases shares sequence elements distinct from those in groups of serine/threonine kinases. The serine/threonine kinases can be divided in several groups which each show a degree of primary sequence conservation within their kinase domains. Protein Kinase C (PKC) falls within the group of AGC kinases which also embraces the cAMP-dependent protein kinase (PKA) and the cGMP-dependent protein kinase (PKG; hence the term AGC) in addition to Protein Kinase B (PKB or Akt), G-protein-regulated kinases (GRKs), S6 Kinase and several others. Outside their catalytic domain, the AGC kinases are quite different reflecting the distinct control mechanisms impinging on them. PKA is activated by binding of the second messenger cAMP, followed by dissociation of the catalytic subunit. PKC is under the control of Ca2+ and diacylglycerol and PKB is regulated acutely by phosphoinositides. Furthermore, kinases of the AGC group are subject to control by phosphorylation, providing additional inputs from those messenger systems that regulate the relevant kinases or phosphatases.
Protein Kinase C The action of extracellular hormones is mediated inside the cell by intracellular molecules, “second messengers”, which activate the cellular response machinery. The second messenger concept became established in the 1960s, with the cAMP system as the major paradigm.5 With the identification of the cAMP-dependent protein kinase (PKA) as the target for cAMP, an intracellular signal transduction chain emerged.6,7 Work by Nishizuka and coworkers in the mid and late 1970s suggested the existence of a cyclic nucleotide-independent kinase (Protein Kinase C) which could be activated by a combination of Ca2+, diacylglycerol and phospholipids.8-10 This discovery together with earlier observations on agonist induced inositol lipid turnover, which suggested the existence of lipid derived second messengers like diacylglycerol, provided the cornerstone to a second signal transduction pathway which ultimately proved to be involved in a plethora of biological functions (Fig. 1). The basic signal transduction pathway is dependent upon the ability of particular cell surface receptors to stimulate the activity of phospholipase C, which in turn acts upon phosphatidylinositol 4,5-bisphosphate to yield diacylglycerol and inositol 1,4,5-trisphosphate, finally resulting in PKC activation. It has become clear recently that inputs from other lipid-generating systems impinge on this basic control pathway. In particular, phosphatidylinositol 3,4,5 trisphosphate (PIP3) activates phosphoinositide dependent kinase (PDK) which phosphorylates numerous kinases including PKC, a modification which is essential to their activity.11 Since PIP3 itself can be seen as a second messenger, it thus appears that PKC activity is determined by the intracellular availability of a number of such messengers. The function of PKC is checked on this basis, not just acutely after receptor activation, when a burst of second messenger is generated, but also in time—by the continued presence of low amounts of second messenger inside the cell—and in space—by restricting the availability of second messenger inside the cell.
Protein Kinase C Isotypes and Their Domain Structure The above suggests that PKC is a single molecular entity however in reality this is not the case and in fact a family of ‘PKC isotypes’ can be defined (Fig. 2).12 PKC-α, β and γ were the first to be identified, as cDNAs obtained from brain libraries.13,14 Low stringency screening led to the identification of PKC-δ, ε and ζ whilst screening of non-brain libraries identified PKC-η, θ and ι.15 Further efforts have not led to the identification of additional PKC genes, although variants have been picked up of some of the isotypes. 12 Alignment of the different isotypes reveals several areas of high homology (Fig. 2). All PKCs contain a C-terminal catalytic domain and an N-terminal regulatory domain. As mentioned
Introduction
3
Figure 1. The PKC signal transduction pathway. The basic pathway involves allosteric activation of PKC by the intracellular messengers, diacylglycerol and Ca2+, which are generated by phospholipase C which in turn is activated by receptor input. PKC activity is affected by a number of additional inputs, including several kinases, phosphatases and intracellular binding proteins.
before, the catalytic domains show a high degree of homology, and also share homology with equivalent domains from other kinases. The regulatory domains are more diverse and they form the basis for subdivision of the isotypes into three subgroups: classical PKCs (PKC-α, β and γ), novel PKCs (PKC-δ, ε, η and θ) and atypical PKCs (PKC-ζ and ι; note that the λ isotype represents the mouse orthologue of human PKC-ι). The regulatory domain of the classical PKCs contains several conserved functional modules: C1a, C1b and C2. The C1a and b modules in these isotypes are located N-terminal of the C2 module. Functionally C1 represents the phorbol ester/diacylglycerol binding site whilst the C2 module comprises the Ca2+ binding site. Although these modules were first identified in PKC, neither one is unique to PKC—over the years they have been found in numerous other proteins where they execute similar functions. C1 and C2 modules are also present in novel PKCs, however in these PKC isotypes the C2 region is located towards the N-terminus. Although the C2 module of novel PKC isotypes has the overall structural features of a C2 barrel, it is significantly different from that in classical PKCs and does not coordinate Ca2+. Two classes of C2 modules can be found in novel PKC isotypes, one present in PKC-δ and -θ and the other present in PKC-ε and -η. Atypical PKCs differ fundamentally from classical and novel isotypes in that they are not responsive to Ca2+, diacylglycerol or phorbol ester. This is because they contain no recognizable C2 module and a C1 module which is structurally different from the typical C1 module as defined in classical and novel PKCs. All PKC isotypes contain an autoinhibitory motif in their regulatory domain. In all cases this element is localized N-terminal of the C1 module. The autoinhibitory motif resembles a PKC consensus substrate phosphorylation site sequence with a central alanine residue representing a phosphate accepting serine/threonine residue. For this reason the autoinhibitory motif is generally referred to as pseudosubstrate site. The pseudosubstrate site slots into the substrate binding groove in the catalytic domain so as to keep the catalytic domain in an inactive conformation. Activation of PKC essentially involves the dissociation of the pseudosubstrate site from this catalytic groove, which now becomes accessible to a ‘real’ substrate. The dissociation of the pseudosubstrate site is provoked in an allosteric fashion by the binding of cofactors like diacylglycerol and Ca2+ to their respective modules on the regulatory domain.
4
Protein Kinase C
Figure 2. The PKC family and the domain structure of PKC. Ten PKC isotypes have been identified which fall into three classes: classical, novel and atypical PKCs. Each class has a characteristic domain structure based on the presence and/or topological organisation of C1 domains (typical or atypical-tC1, aC1), C2 domains, pseudosubstrate sites (indicated by small ellipsoid) and catalytic domains.
Defining the PKC Family Additional kinases may or may not be included in the “PKC family”. PKC-µ, PKC-ν and PKD2 are closely related kinases containing C1 modules as present in the PKC isotypes mentioned above. Catalytic domain conservation criteria suggest that these kinases may not fall within the PKC family and these kinases are generally considered to form a separate family, the PKD family. Other kinases that are closely related to PKC are PKN and PKC-related kinases, which possess a catalytic domain highly homologous to that of PKC and a C2 module in their regulatory domain. The existence of such kinases justifies the question as to what constitutes the PKC family. It is of interest to consider the presence of related isotypes in the lower eukaryotic nematode worm Caenorhabditis elegans. In this organism, several PKC isotypes have been identified and they show a subdivision as indicated above for the mammalian proteins namely orthologues of PKC-α (classical), PKC-δ, PKC-ε (the two classes within the novel PKCs), and PKC-ι (atypical). Representatives of the PKB family, the PKD family and PKC-related kinases have also been defined in C. elegans. Thus these families may have evolved separately, becoming increasingly complex in the higher organisms. To a large extent, the PKC family can be distinguished within the AGC group on the basis of catalytic domain conservation. As mentioned above, outside their catalytic domain, the AGC kinases are quite different and distinctions between families can also be made on this basis. All PKC isotypes contain a C1 module in their regulatory domain, albeit that the structure (and functionality) varies between isotypes. The combination of catalytic domain primary conservation and the presence of a C1 signature module may thus be used to define the PKC family. In functional terms this definition is quite forced—the C1 module in atypical PKC isotypes does not bind diacylglycerol and one may argue that the atypical PKCs fall outside the family since they lack the characteristic diacylglycerol dependence of other PKC isotypes. Altogether, the AGC kinases appear to form a continuum, in which a conserved catalytic domain is juxtaposed to various different regulatory modules. Within this continuum patterns can be found and some kinases appear to be more closely related than others, falling into distinct ‘families’. PKC isotypes are an example of this. It is important to realise that kinase families have
Introduction
5
Table 1. Processes in which PKC isotypes have been implicated based upon analysis of PKC-deficient mice Isotype
Function
PKC-α PKC-β
Insulin receptor feedback16 B-cell receptor signalling, feedback regulation17 Mast cell activation18 Neutrophil NADPH oxidase activation19 Glucose transport20 Transcription factor activation in hypoxia21,22 Fear conditioning23 Hippocampal LTP24,25 Learning, addiction, anxiety26,27 Neuropathic pain28-30 Smooth muscle cell homeostasis/apoptosis31 B-cell tolerance32,33 Acute pain34 Alcohol addiction35 Macrophage activation, host defense36
PKC-γ
PKC-δ PKC-ε
PKC-η PKC-θ PKC-ζ
T-cell activation37 NFkB activation38 B-cell receptor signalling39
PKC-ι
been defined historically, this is certainly the case for the PKC family. Furthermore, even if strict categorization of each kinase is difficult, the classification in kinase families is a useful means to come to terms with the heterogeneous appearance of this functional domain. Ultimately, the aim should be to define the heterogeneity within the kinase superfamily in functional terms and each kinase, whether member of a family or classified on its own, will have to be treated individually. For the members of the PKC family, a start has been made with this definition by the phenotypic characterization of PKC isotype-deficient mice, revealing the cellular processes in which the isotypes may play a role (Table 1).
This Book The literature on PKC is vast and it is impossible to capture all the data in a single volume. This book has been divided in two main sections. The first section contains a description of the structure of the kinase and its basic biochemical properties. It highlights the enormous progress that has been made on the structure and function of the individual regulatory modules, the phosphorylation of PKC and the interaction with protein partners, all of which have proved invaluable in furthering our understanding of the place of the kinase in the cell. The second section focuses on cell biological approaches to PKC, loosely based on deriving functional information by manipulating the kinase. This section highlights some of the work in lower eukaryotic model systems, as well as studies of PKC isotype function in individual cell types. Altogether, it is hoped that this book will contribute to the continued evaluation of the PKC family and its individual members.
Protein Kinase C
6
References 1. Hanks SK, Quinn AM, Hunter T. The protein kinase family: conserved features and deduced phylogeny of the catalytic domains. Science 1988; 241:42-52. 2. Klumpp S, Krieglstein J. Phosphorylation and dephosphorylation of histidine residues in proteins. Eur J Biochem 2002; 269:1067-1071. 3. Smith CM, Shindyalov IN, Veretnik S et al. The protein kinase resource. Trends Biochem Sci 1997; 22:444-446. 4. Plowman GD, Sudarsanam S, Bingham J et al. The protein kinases of Caenorhabditis elegans: a model for signal transduction in multicellular organisms. Proc Natl Acad Sci USA 1999; 96:13603-13610. 5. Sutherland EW. On the biological role of cyclic AMP. JAMA 1970; 214:1281-1288. 6. Walsh DA, Perkins JP, Krebs EG. An adenosine 3',5'-monophosphate-dependant protein kinase from rabbit skeletal muscle. J Biol Chem 1968; 243:3763-3765. 7. Miyamoto E, Kuo JF, Greengard P. Adenosine 3',5'-monophosphate-dependent protein kinase from brain. Science 1968; 165:63-65. 8. Inoue M, Kishimoto A, Takai Y et al. Studies on a cyclic nucleotide-independent protein kinase and its proenzyme in mammalian tissues. II. Proenzyme and its activation by calcium-dependent protease from rat brain. J Biol Chem 1977; 252:7610-7616. 9. Kishimoto A, Takai Y, Nishizuka Y. Activation of glycogen phosphorylase kinase by a calciumactivated, cyclic nucleotide-independent protein kinase system. J Biol Chem 1977; 252:7449-7452. 10. Takai Y, Yamamoto M, Inoue M et al. A proenzyme of cyclic nucleotide-independent protein kinase and its activation by calcium-dependent neutral protease from rat liver. Biochem Biophys Res Commun 1977; 77:542-550. 11. Parker PJ, Parkinson SJ. AGC protein kinase phosphorylation and protein kinase C. Biochem Soc Trans 2001; 29:860-863. 12. Kofler K, Erdel M, Utermann G et al. Molecular genetics and structural genomics of the human protein kinase C gene module. Genome Biol 2002; 3:research0014.1-0014.10. 13. Parker PJ, Coussens L, Totty N et al. The complete primary structure of protein kinase C-the major phorbol ester receptor. Science 1986; 233:853-859. 14. Coussens L, Parker PJ, Rhee L et al. Multiple, distinct forms of bovine and human protein kinase C suggest diversity in cellular signaling pathways. Science 1986; 233:859-866. 15. Dekker LV and Parker PJ. Protein kinase C-a question of specificity. Trends Biochem Sci 1994; 19:73-77. 16. Leitges M, Plomann M, Standaert ML et al. Knockout of PKC alpha enhances insulin signaling through PI3K. Mol Endocrinol 2002; 16:847-858. 17. Leitges M, Schmedt C, Guinamard R et al. Immunodeficiency in protein kinase cbeta-deficient mice. Science 1996; 273:788-791. 18. Nechushtan H, Leitges M, Cohen C et al. Inhibition of degranulation and interleukin-6 production in mast cells derived from mice deficient in protein kinase Cbeta. Blood 2000; 95:1752-1757. 19. Dekker LV, Leitges M, Altschuler G et al. Protein kinase C-beta contributes to NADPH oxidase activation in neutrophils. Biochem J 2000; 347(Pt 1):285-289. 20. Standaert ML, Bandyopadhyay G, Galloway L et al. Effects of knockout of the protein kinase C beta gene on glucose transport and glucose homeostasis. Endocrinology 1999; 140:4470-4477. 21. Yan SF, Lu J, Zou YS et al. Protein kinase C-beta and oxygen deprivation. A novel Egr-1-dependent pathway for fibrin deposition in hypoxemic vasculature. J Biol Chem 2000; 275:11921-11928. 22. Harris RA, McQuilkin SJ, Paylor R et al. Mutant mice lacking the gamma isoform of protein kinase C show decreased behavioral actions of ethanol and altered function of gamma-aminobutyrate type A receptors [see comments]. Proc Natl Acad Sci USA 1995; 92:3658-3662. 23. Weeber EJ, Atkins CM, Selcher JC et al. A role for the beta isoform of protein kinase C in fear conditioning. J Neurosci 2000; 20:5906-5914. 24. Abeliovich A, Chen C, Goda Y et al. Modified hippocampal long-term potentiation in PKC gamma-mutant mice. Cell 1993; 75:1253-1262. 25. Kano M, Hashimoto K, Chen C et al. Impaired synapse elimination during cerebellar development in PKC gamma mutant mice. Cell 1995; 83:1223-1231. 26. Abeliovich A, Paylor R, Chen C et al. PKC gamma mutant mice exhibit mild deficits in spatial and contextual learning. Cell 1993; 75:1263-1271. 27. Bowers BJ, Owen EH, Collins AC et al. Decreased ethanol sensitivity and tolerance development in gamma- protein kinase C null mutant mice is dependent on genetic background. Alcohol Clin Exp Res 1999; 23:387-397.
Introduction
7
28. Martin WJ, Malmberg AB, Basbaum AI. PKC-gamma Contributes to a Subset of the NMDA-Dependent Spinal Circuits That Underlie Injury-Induced Persistent Pain. J Neurosci 2001; 21:5321-5327. 29. Ohsawa M, Narita M, Mizoguchi H et al. Reduced hyperalgesia induced by nerve injury, but not by inflammation in mice lacking protein kinase C gamma isoform. Eur J Pharmacol 2001; 429:157-160. 30. Malmberg AB, Chen C, Tonegawa S et al. Preserved acute pain and reduced neuropathic pain in mice lacking PKCgamma. Science 1997; 278:279-283. 31. Leitges M, Mayr M, Braun U et al. Exacerbated vein graft arteriosclerosis in protein kinase Cdelta-null mice. J Clin Invest 2001; 108:1505-1512. 32. Mecklenbrauker I, Saijo K, Zheng NY et al. Protein kinase Cdelta controls self-antigen-induced B-cell tolerance. Nature 2002; 416:860-865. 33. Miyamoto A, Nakayama K, Imaki H et al. Increased proliferation of B cells and auto-immunity in mice lacking protein kinase Cdelta. Nature 2002; 416:865-869. 34. Khasar SG, Lin YH, Martin A et al. A novel nociceptor signaling pathway revealed in protein kinase C epsilon mutant mice. Neuron 1999; 24:253-260. 35. Olive MF, Mehmert KK, Messing RO et al. Reduced operant ethanol self-administration and in vivo mesolimbic dopamine responses to ethanol in PKC epsilon-deficient mice. Eur J Neurosci 2000; 12:4131-4140. 36. Castrillo A, Pennington DJ, Otto F et al. Protein kinase Cepsilon is required for macrophage activation and defense against bacterial infection. J Exp Med 2001; 194:1231-1242. 37. Sun Z, Arendt CW, Ellmeier W et al. PKC-theta is required for TCR-induced NF-kappaB activation in mature but not immature T lymphocytes. Nature 2000; 404:402-407. 38. Leitges M, Sanz L, Martin P et al. Targeted disruption of the zetaPKC gene results in the impairment of the NF-kappaB pathway. Mol Cell 2001; 8:771-780. 39. Martin P, Duran A, Minguet S et al. Role of zetaPKC in B-cell signaling and function. EMBO J 2002; 21:4049-4057.
8
Protein Kinase C
CHAPTER 2
Structure, Function and Membrane Interactions of C1 Domains Matthew Pearson and James H. Hurley
S
ubcellular localization of proteins is a central organizing theme in signal transduction. The redistribution of signaling proteins to specific cellular membranes in response to specific events is a required step in many signaling pathways. Such redistribution is critical not only for colocalization of interacting proteins but it is frequently linked to their allosteric activation. Protein kinase C (PKC), the subject of this book, is an archetype for the study of the regulated subcellular redistribution and activation of an enzyme in response to the generation of a lipid signal.1-4 Lessons learned from the intensive study of the lipid-binding domains in PKC have general implications for a vast array of signaling mechanisms. Two of the best-known modular domains involved in recognizing specific lipids in membranes are named for their occurrence in, and their discovery in, PKC. These are the PKC homology-1 (C1) and -2 (C2) domains, the former being the subject of this chapter. The C1 domain has now been found in almost 400 proteins in the nonredundant protein database including PKC isozymes (Fig. 1).5-7 The activation of conventional and novel PKC isozymes by the second messenger diacylglycerol (DAG) and the tumor promoting phorbol esters is mediated by their C1 domains. Like PKCs, C1 domains can be divided into two classes based on their ligand binding properties: typical and atypical.5 Typical C1 domains are found in conventional and novel (i.e., α,β,γ,δ,ε,θ,η, and µ, but not λ/ι or ζ) PKCs, chimaerins, and Unc-13.5,6 More recently, their presence was discovered in a novel ras guanine nucleotide exchange factor, RasGRP,6,8,9 where they act as receptors for DAG and phorbol esters. The atypical C1 domains are so named for their presence in the atypical PKC-λ/ι and ζ, and they are also found in DAG kinases,10 a number of GTPase interacting proteins, and many putative signaling proteins of unknown function.7 The atypical domains do not bind DAG or phorbol esters, and instead are thought to act as receptors for proteins or other lipids. The past several years has seen an explosion of sequence information that has greatly expanded the class of recognized C1-containing proteins. In addition, the determination of structures of C1 domains, coupled to site-directed mutagenesis experiments, has led to detailed models for the ligand binding functions of the domains. The typical C1 domains of PKCs and the atypical C1 domain of Raf have been intensively studied. The use of fluorescently-tagged C1 domain constructs in cells has allowed the measurement in vivo of the kinetics of membrane localization in response to DAG or phorbol ester signals. This chapter provides an update on several of these experiments, which are quickly providing details of C1 domain function, specifically about ligand interaction, membrane localization, and the involvement of C1 domains in the allosteric activation of PKCs.
Protein Kinase C, 2nd Edition, edited by Lodewijk Dekker. ©2004 Eurekah.com and Kluwer Academic / Plenum Publishers.
Structure, Function and Membrane Interactions of C1 Domains
9
A
Figure 1. A) Sequence alignment of representative C1 domains, with known and predicted typical domains grouped above, and known and predicted atypical domains below. The secondary structural elements of the domain are shown above the alignment. The three residues critical for organizing the phorbol ester/DAG binding site are boxed. The C1 domain of GEK is predicted to be typical based on the conservation of these three residues, but it should be noted that a positively charged Lys residue replaces a highly conserved hydrophobic residue at the end of strand β2 and may affect the ligand/membrane-binding properties of this domain. The three boxed residues are also conserved in the atypical C1A domain of DAG kinase α. A deletion in the loop between strands β3 and β4 likely affects the placement of two of these residues, which may in turn affect ligand binding. Abbr: Mm: Mus musculus, Hs: Homo sapiens, Ce: Caenorhabditis elegans, Dm: Drosophila melanogaster. Figure is continued on the following page.
C1 Domain Structure and Stereospecific Ligand Recognition
Several structures of C1 domains have been determined using X-ray crystallography11 and NMR.12-15 C1 domains contain ~50 amino acids which are organized into two β-sheets and a short C-terminal α-helix (Fig. 2). The secondary structural elements are arranged around two Zn2+ ion binding motifs, each containing three cysteines and one histidine. These buried Zn2+ clusters are integral to the domain structure. At the tip of the domain, two β-strands have been “unzipped” or pulled apart, leaving a pocket where main chain hydrogen bonding groups are exposed to solvent. The crystal structure of the second C1 domain from PKC-δ bound to a phorbol acetate revealed the “unzipped” pocket to be the binding site for phorbol esters.11 Key Pro and Gln residues help maintain the unzipped conformation. Conserved C1 residues make hydrogen bonds to oxygens on the multi-ringed phorbol group which are believed to be analogous in part to the interactions made with DAG. In particular, the main chain polar groups of residues Thr242, Leu251, and Gly253 provide binding sites for the 3-, 4-, and 20-oxygens of the phorbol and for the 3-hydroxyl and one of the acyl oxygens of a modeled DAG. The conserved Pro,
10
Protein Kinase C
B
Figure 1, continued. B) Domain organization of several C1 domain-containing proteins with the C1 domains highlighted in gray. Abbr: C1: PKC homology-1, C2: PKC homology-2, PH: pleckstrin homology, RasGEFN: Ras guanine nucleotide exchange factor N-terminal motif, RasGEF: Ras guanine nucleotide exchange factor, SH2: Src homology-2, RhoGAP: GTPase activator protein for Rho, CNH: citron/NIK1, PBD: P21-Rho-binding domain, RA: Ras association, MYSc: myosin catalytic, IQ: calmodulin-binding motif, RBD: Ras binding domain, CH: calponin homology, RhoGEF: Rho guanine nucleotide exchange factor, SH3: Src homology-3, EF: EF-hand, DAGKc: diacylglycerol kinase catalytic, DAGKa: diacylglycerol kinase accessory, PDZ: PSD-95/Dlg/ZO-1/2 domain, RGS: regulator of G protein signaling, GRAM: glucosyltransferase/Rab/myotubularin.
Gly, and Gln residues together provide a fingerprint for DAG binding functionality that allows typical and atypical C1 domains to be classified and their DAG binding properties to be predicted based on their sequence.5 Site-directed mutagenesis suggests that hydrophobic residues surrounding the binding pocket also influence the binding of DAG or phorbol esters, but the effect is more pronounced in the presence of phospholipids.16 This suggests that the major role of these residues is interaction with the membrane as opposed to specific ligand interactions (see below). The typical PKC C1 domains appear to do more than just bind DAG, phorbol esters, and nonspecifically interact with other membrane constituents. Recently, one of the C1 domains of PKC-βII has been found to specifically recognize L-phosphatidylserine (PS).17 The only deep pocket on the C1 domain structure is the phorbol ester/DAG binding site. On the other hand there is precedent for binding of a charged phospholipid headgroup to a shallow pocket on the “side” of a small zinc finger domain. The FYVE domain, which has distant structural similarity to the C1 domain, binds phosphatidylinositol 3-phosphate in just such a manner.18,19
Structure, Function and Membrane Interactions of C1 Domains
11
Figure 2. Structure of the C1B domain of PKC-δ complexed with phorbol ester and docked onto a membrane surface (modifed from ref. 4). The myristate tail has been modeled onto the phorbol ester. The molecular surface is colored green for hydrophobic residues and blue for basic residues. Selected phorbol-binding and membrane-interacting residues are highlighted. The Zn2+ ions are colored cyan. The color version of this figure can be viewed at www.Eurekah.com.
L-PS is an essential coactivator of conventional and novel PKCs. Since PS binds to conventional PKC (α,β,γ) C2 domains in a calcium-dependent manner,20 it might have been expected that the stereospecific binding site for PS would have been on this domain. PKC has multiple PS binding sites, some stereospecific and some not.21 The nonstereospecific sites can be occupied by acidic lipids other than L-PS with similar effects on membrane binding. The binding of PS to the conventional PKC C2 domain may be a consequence of the preference of this C2 domain for acidic phospholipids, thus representing one of the nonstereospecific PS binding sites. However, Cho and colleagues have derived a conflicting interpretation from their mutagenesis experiments on PKC-α that suggests that the C2, and not the C1, domain specifically binds PS.21 Newton and coworkers found that the C2 domain of PKC-β does not bind stereospecifically to L-PS.17 The systems under study are sufficiently similar that isoform differences seem unlikely to explain the disparities in the findings. Thus the location of the stereospecific L-PS binding site of PKC remains somewhat controversial (see also Chapters 3 and 4).
Membrane Binding and Subcellular Localization of C1 Domains C1 domain binding to DAG or phorbol esters propels the insertion of the domain into phospholipid membranes.1-4 Results from structural analyses suggested a mechanism of membrane insertion,11 which has subsequently been confirmed by mutagenesis16 and biochemical and biophysical analysis.14,22 The “unzipped” β-strands in the domain form a polar pocket in
12
Protein Kinase C
what is otherwise a large hydrophobic surface at the end of the domain. The binding of phorbol ester (or DAG) puts a hydrophobic cap on this pocket, with the polar groups of the protein hydrogen bonding with the polar groups of the ligand, and the hydrophobic part of the ligand completing a contiguous hydrophobic surface on the domain. This surface can then insert into the membrane, burying the hydrophobic surface as well as allowing a basic ring of residues surrounding the hydrophobic patch to interact with the negatively charged surface of the plasma membrane. This model has been verified by NMR experiments that indicate that residues on the hydrophobic tip are inserted into the membrane core during micelle binding.14 In addition, site-directed mutagenesis experiments have revealed that conserved hydrophobic residues surrounding the ligand-binding pocket influence the interaction with the membrane,16,22 and that charged residues in the basic ring take part in nonspecific electrostatic interactions at the membrane surface.21,22 In summary, the mechanism of membrane insertion involves the formation of a ternary complex of the C1 domain with the ligand (phorbol ester or DAG) and the phospholipid membrane. Therefore, the affinity of C1 domains for a membrane depends not only on specific hydrogen bonding interactions with the lipid second messengers, but also on nonspecific hydrophobic and charged interactions with the bulk membrane. C1 domains from PKC coupled to green fluorescent protein (GFP) have been used in vivo to visualize membrane localization in response to signals.2-4,23 After activation of cell-surface receptors or the addition of phorbol ester or DAG, the C1 domains of PKC-γ localize to the plasma membrane of cells in a matter of seconds. Interestingly, the C1 domain constructs were also shown to localize to the nuclear membrane in response to more polar phorbol esters, suggesting that different signals may lead to different subcellular locations of PKC activation.23 Photobleaching and recovery experiments indicate the C1 domains bound to membranes in the presence of phorbol ester are relatively immobile, with a lateral diffusion that is slower than a C2 domain localized to the membrane by calcium ions.2 In addition, the C1-membrane interaction holds activated PKC at the membrane for ~ 8 seconds after the removal of the calcium coactivator.2 These dynamics indicate that the C1 domain interaction accounts for a substantial portion of the membrane affinity of activated PKCs.
Typical C1 Domains in Proteins Other than PKC
PKC is not the only receptor for DAG or phorbol ester signals.5,6,25 The class of nonPKC DAG and phorbol ester receptors each contain one or more C1 domains and include the chimaerins, Unc-13, and RasGRP. The C1 domains of chimaerins have similar affinities as PKC C1 domains for DAG and phorbol esters.25 The role that the C1 domains play in the activation of the chimaerin GTPase activating protein (GAP) domains (which are specific for Rac, a Rho family GTPase) remains unclear, but the domains are certainly involved in chimaerin localization to membranes which may be crucial for colocalization with Rac.25-27 The function of the Unc-13 protein found in Caenorhabditis elegans and the Munc-13 proteins, which are its mammalian orthologs, remains unknown. The C1 domains of Unc13 bind phorbol esters and DAG. The Munc-13 proteins are localized to synaptic junctions in neurons, suggesting that the localization of these proteins to specific membranes may play a role in their function. The most interesting recent addition to the class of typical C1 domain proteins is the Ras guanine nucleotide exchange factor, RasGRP. The C1 domains of RasGRP bind phorbol esters.8,9 As for the chimaerins, it remains to be seen whether the Ras-specific guanine nucleotide exchange activity of RasGRP is regulated by membrane localization. The presence of typical C1 domains in RasGRP presents a link between DAG signaling and Ras activation. Finally, the Drosophila kinase GEK represents a potential new DAG-binding C1 domain protein (Fig. 1), although functional data on its phorbol ester and DAG responsiveness have yet to be reported.
Atypical C1 Domains Atypical C1 domains are named for their presence in the atypical PKCs ζ and λ/ι. These C1 domains do not bind DAG or phorbol esters, and as a consequence, the atypical PKCs are not
Structure, Function and Membrane Interactions of C1 Domains
13
activated by these lipids. The functions of the atypical PKCs remain poorly understood. Ceramide is a known lipid second messenger whose structure resembles that of DAG. The hypothesis that ceramide activates PKC-ζ via its C1 domain has circulated for some time, but has been difficult to substantiate because of conflicting results in different laboratories. A recent report suggesting that ceramide directly activates PKC-ζ in vitro and in HEK-293 cells28 has revived this intuitively appealing notion. Strong evidence is still lacking for an activation mechanism that parallels the C1 domain-mediated DAG activation of other PKCs. Atypical C1 domains are also present in the diacylglycerol kinases and a group of proteins that interact with small GTPases. The function of C1 domains in DAG kinases remains unknown.10 Of course, these enzymes use DAG as a substrate, but the C1 domain is not required for enzyme activity and so is not involved in substrate binding.29 In addition, none of the DAG kinases have been shown to bind phorbol esters.10,29 The best-studied member of the C1-containing GTPase effectors is the Raf kinase. The C1 domain appears to be involved in the allosteric regulation of this enzyme, which involves binding to the Ras GTPase.30,21 The region of the C1 domain that is suggested to be involved in maintaining the inactive conformation of Raf partially overlaps the binding site for lipids on the typical C1s.30,31 This intramolecular interaction in Raf is disrupted by the binding to Ras, which may involve interactions between Ras and the C1 domain, although the Ras binding domain (RBD) of Raf is considered the main player in this interaction. A number of other atypical C1 domain-containing proteins have been identified7 but little is known about the role of their C1 domains in their function. The Drosophila SET domain binding factor (sbf ) contains two recently described putative membrane targeting modules, the DENN and GRAM domains, in addition to its atypical C1 domains. Several otherwise unclassified Drosophila proteins have atypical C1 domains, including RhoGEF proteins like shar pei, and several PDZ containing proteins of unknown function. Several myosin IX heavy chains contain atypical C1 domains. Lastly, RalGDS is an atypical C1 and RA (Ras association) domain-containing ras effector.
C1 Domains in the Context of Multidomain Signaling Proteins Conventional and novel PKCs have two C1 domains (C1a and C1b) either N- or C-terminal to a C2 domain, all preceding the kinase catalytic domain.1-4 The arrangement of these domains in three dimensions is unknown, and the structure determination of an intact PKC will be critical for understanding enzyme activation. The conformation of the domains with respect to one another clearly changes during activation. Features of the inactive and active conformations have been deduced by indirect means (see Chapter 4). Analysis of the kinetics of PKC activation and membrane localization has led to considerable insights into the positioning of the C1 domains in the active and inactive conformations.2 Based on the ~100 second delay of the membrane localization of intact PKCs relative to isolated C1 domains after stimulation by DAG, the C1 domains are almost certainly in a conformation where the DAG binding site is inaccessible in the enzymatically inactive state.2 The two C1 domains are not functionally identical in the context of the intact PKC32-34, and their relative roles differ between different PKC isoforms. In PKC-α, the two C1 domains are equally important in membrane targeting.33 However, site-directed mutagenesis experiments on PKC-α suggest that C1a inserts into membranes, while C1b does not.23 Intact PKC-βII and its isolated C1b domain bind phorbol esters with equal affinity, suggesting C1b is primarily responsible for targeting this isoform.17 In the case of PKC-γ, the isolated C1a is targeted to membranes more efficiently than C1b.24 Mutagenesis of PKC-δ has shown that C1b is the primary site for DAG-dependent membrane binding.34 Clearly, the interactions of the C1 domains with the rest of the enzyme have an important influence on the membrane-binding properties, and isoform-specific nuances of the interactions between C1 domains and the rest of the enzyme profoundly affect the outcome.
14
Protein Kinase C
The potential interactions between the C1 and C2 domains have generated a good deal of experimentation and debate. Mutagenesis experiments on PKC-α have been interpreted in terms of a model for PS specificity that postulates an intramolecular interaction of the C1a domain and C2 domains in the inactive kinase.22 Mutation of Asp55 to Ala in C1a leads to enhanced binding to membranes and lower specificity for PS. This was interpreted to suggest that this negatively-charged residue interacts with a positively charged residue on some other part of PKC before activation. Another possible explanation is that Asp55 is part of the PS binding site on the surface of the C1 domain. The two models are not mutually exclusive. In the former model, PS binding is required during activation to disrupt this charge interaction, presumably freeing the C1 domain to insert into a membrane containing DAG or phorbol esters. Confirmation of these and other models concerning the activation of PKC awaits the determination of the structure of an intact PKC.
Future Directions Over the past decade, the structure, function, and membrane targeting mechanism of typical C1 domains have been elucidated. The function of C1 domains in the context of intact PKCs has been explored, but so far the structure of an intact C1 domain-containing protein is lacking. Such a structure would be invaluable for understanding allosteric regulation through C1 domains. Understanding the function of atypical C1 domains has lagged behind the typicals. Aside from some information on the Raf-C1, much remains to be learned about this class, which comprises at least half of all C1 domains.
References 1. Newton AC, Johnson JE. Protein kinase C: a paradigm for regulation of protein function by two membrane-targeting modules. Biochim Biophys Acta Rev Biomembr 1998; 1376:155-172. 2. Oancea E, Meyer T. Protein kinase C as a molecular machine for decoding calcium and diacylglycerol signals. Cell 1998; 95:307-318. 3. Hurley JH, Meyer T. Subcellular targeting by membrane lipids. Curr Opin Cell Biol 2001; 13:146-152. 4. Hurley JH, Misra S. Signaling and subcellular targeting by membrane-binding domains. Annu Rev Biophys Biomol Struct 2000; 29:49-79. 5. Hurley JH, Newton AC, Parker PJ et al. Taxonomy and function of C1 protein kinase C homology domains. Protein Sci 1997; 6:477-480. 6. Ron D, Kazanietz MG. New insights into the regulation of protein kinase C and novel phorbol ester receptors. FASEB J 1999; 13:1658-1676. 7. Schultz J, Milpetz F, Bork P et al. SMART, a simple modular architecture research tool: identification of signaling domains. Proc Natl Acad Sci USA 1998; 95:5857–5864. 8. Ebinu JO, Bottorff DA, Chen EYW et al. RasGRP, a Ras guanylyl nucleotide-releasing protein with calcium- and diacylglycerol-binding motifs. Science 1998; 280:1082-1086. 9. Tognon CE, Kirk HE, Passmore LA et al. Regulation of RasGRP via a phorbol ester-responsive C1 domain. Mol Cell Biol 1998; 18:6995-7008. 10. Topham MK, Prescott SM. Mammalian diacylglycerol kinases, a family of lipid kinases with signaling functions. J Biol Chem 1999; 274:11447-11450. 11. Zhang G, Kazanietz MG, Blumberg PM et al. Crystal structure of the Cys2 activator-binding domain of protein kinase Cδ in complex with phorbol ester. Cell 1995; 81:917-924. 12. Hommel U, Zurini M, Luyten M. Solution structure of a cysteine-rich domain of rat protein kinase C. Nat Struct Biol 1994; 1:383-387. 13. Ichikawa S, Hatanaka H, Takeuchi Y et al. Solution structure of cysteine-rich domain of protein kinase C. J Biochem 1995; 117:566-574. 14. Xu RX, Pawelczyk T, Xia T-H et al. NMR structure of a protein kinase C-phorbol-binding domain and study of protein-lipid micelle interactions. Biochemistry 1997; 36:10709-10717. 15. Mott HR, Carpenter JW, Zhong S et al. The solution structure of the Raf-1 cysteine-rich domain: A novel Ras and phospholipid binding site. Proc Natl Acad Sci USA 1996; 93:8312-8317. 16. Wang QJ, Fang T-W, Nacro K et al. Role of hydrophobic residues in the C1b domain of protein kinase Cδ on ligand and phospholipid interactions. J Biol Chem 2001; 276:19580-19587.
Structure, Function and Membrane Interactions of C1 Domains
15
17. Johnson JE, Giorgione J, Newton AC. The C1 and C2 domains of protein kinase C are independent membrane targeting modules, with specificity for phosphatidylserine conferred by the C1 domain. Biochemistry 2000; 39:11360-11369. 18. Misra S, Hurley JH. Crystal structure of a phosphatidylinositol 3-phosphate-specific membranetargeting motif, the FYVE domain of Vps27p. Cell 1999; 97:657-666. 19. Kutateladze T, Overduin M. Mechanism of endosome docking by the FYVE domain. Science 2001; 291:1793-1796. 20. Verdaguer N, Corbalan-Garcia S, Ochoa WF et al. Ca-2+ bridges the C2 membrane-binding domain of protein kinase Cα directly to phosphatidylserine. EMBO J 1999; 18:6329-6338. 21. Mosior M, Newton AC. Mechanism of the apparent cooperativity in the interaction of protein kinase C with phosphatidylserine. Biochemistry 1997; 37:17271-17279. 22. Bittova L, Stahelin RV, Cho W. Roles of ionic residues of the C1 domain in protein kinase C-α activation and the origin of phosphatidylserine specificity. J Biol Chem 2001; 276:4218-4226. 23. Medkova M, Cho W. Interplay of C1 and C2 domains of protein kinase C-α in its membrane binding and activation. J Biol Chem 1999; 274:19852-19861. 24. Oancea E, Teruel MN, Quest AFG et al. Green fluorescent protein (GFP)-tagged cysteine-rich domains from protein kinase C as fluorescent indicators for diacylglycerol signaling in living cells. J Cell Biol 1998; 140:485-498. 25. Kazanietz MG. Eyes wide shut: protein kinase C isozymes are not the only receptors for the phorbol ester tumor promoters. Mol Carcinog 2000; 28:5-11. 26. Caloca MJ, Garcia-Bermejo ML, Blumberg PM et al. β2-chimaerin is a novel target for diacylglycerol: Binding properties and changes in subcellular localization mediated by ligand binding to its C1 domain. Proc Natl Acad Sci USA 1999; 96:11854-11859. 27. Caloca MJ, Wang HB, Delemos A et al. Phorbol esters and related analogs regulate the subcellular localization of β2-chimaerin, a nonprotein kinase C phorbol ester receptor. J Biol Chem 2001; 276:18303-18312. 28. Bourbon NA, Yun J, Kester M. Ceramide directly activates protein kinase Cζ to regulate a stress-activated protein kinase signaling complex. J Biol Chem 2000; 275:35617-35623. 29. Sakane F, Kai M, Wada I et al. The C-terminal part of diacylglycerol kinase α lacking zinc fingers serves as a catalytic domain. Biochem J 1996; 318:583-590. 30. Cutler Jr RE, Stephens RM, Saracino MR et al. Autoregulation of the Raf-1 serine/threonine kinase. Proc Natl Acad Sci USA 1998; 95:9214-9219. 31. Daub M, Jockel J, Quack T et al. The RafC1 cysteine-rich domain contains multiple distinct regulatory epitopes which control Ras-dependent Raf activation. Mol Cell Biol 1998; 18:6698-6710. 32. Slater SJ, Ho C, Kelly MB et al. Protein kinase Cα contains two activator binding sites that bind phorbol esters and diacylglycerols with opposite affinities. J Biol Chem 1996; 271:4627-4631. 33. Bogi K, Lorenzo PS, Acs P et al. Comparison of the roles of the C1a and C1b domains of protein kinase Cα in ligand induced translocation in NIH 3T3 cells. FEBS Lett 1999; 456:27-30. 34. Szallasi Z, Bogi K, Gohari S et al. Nonequivalent roles for the first and second zinc fingers of protein kinase Cδ - Effect of their mutation on phorbol ester-induced translocation in NIH 3T3 cells. J Biol Chem 1996; 271:18299-18301.
CHAPTER 3
Structural and Functional Specialization of C2 Domains in Protein Kinase C Eric A. Nalefski
Abstract
F
ull activation of protein kinase C (PKC) isotypes requires multiple contacts with the membrane surface via their C1 and C2 domains. PKC C2 domains bind various components of cellular membranes, including anionic phospholipids and membraneassociated proteins. Ca2+-dependent conventional isotypes of PKC (cPKC) are functionally distinguishable from novel Ca2+-independent (nPKC) isotypes by the mechanism of membrane binding displayed by their C2 domains: C2 domains from cPKC isotypes dock to anionic phospholipids when they bind Ca2+, whereas those from nPKC isotypes bind anionic phospholipids constitutively. Although C2 domains from both classes of isotypes are composed of similar anti-parallel β-strand sandwiches, the domains are distinguishable by their order in the primary structure, by arrangement of the β-strands within the domains, and notably, by the conformation and presence of Ca2+-coordinating residues within the interstrand loops. In cPKC C2 domains, Ca2+ binding triggers a large electrostatic change in the domain that promotes nonspecific interactions with the anionic membrane surface, increasing the apparent membrane association rate constant and decreasing the apparent dissociation rate constant. Shared coordination of Ca2+ by cPKC C2 domains and membrane phosphoryl groups helps retain the domain at the membrane surface. The lack of this mechanism in nPKC C2 domains indicates that the domain is highly adapted in cPKC isotypes for participation in rapid cell signaling events associated with Ca2+ mobilization. These properties illustrate the specialization of a common protein motif within a family of closely related enzymes that catalyze the same chemical reaction, specifically linking this enzymatic function with distinct cellular processes.
Introduction One of the hallmarks of PKC is stimulation of its kinase activity by anionic phospholipid membranes (see Chapters 2 and 4). Full activation of PKC requires phospholipid binding mediated by multiple contacts with the membrane surface. Two such membrane-binding regulatory domains in the Ca2+-dependent cPKC and Ca2+-independent nPKC isotypes are the C1 and C2 domains. C1 and C2 domains were initially identified as the first and second of four conserved domains (C1 - C4) within the first cPKC isotypes sequenced.1-5 Tandem C1 domains constitute the zinc finger diacylglycerol (DAG)-binding domains in PKC that bind the pharmacological agent phorbol ester (see Chapter 2). As is the case for the C2 domain, C1 domains are present in many other signaling proteins.6 The C3 and C4 domains encompass the two lobes of the Ser/Thr-kinase catalytic domain that bears similarity to other protein kinases.1-5 Protein Kinase C, 2nd Edition, edited by Lodewijk Dekker. ©2004 Eurekah.com and Kluwer Academic / Plenum Publishers.
Structural and Functional Specialization of C2 Domains in Protein Kinase C
17
Early reports correctly proposed that the C2 domain was responsible for the Ca2+-dependent regulation of cPKC isotypes.7 In support of this hypothesis, a domain homologous to the PKC C2 domain was identified in the primary structure of the α-isoform of cytosolic phospholipase A2 (cPLA2),8 a vertebrate enzyme that catalyzes the cleavage of arachidonic acid from membrane glycerophospholipids in a Ca2+-dependent manner.9 This domain was shown to constitute the Ca2+-dependent lipid-binding (CaLB) domain of cPLA2-α, serving to deliver a Ca2+-independent phospholipase domain to its membrane-sequestered substrate.8,10 The first of two tandem C2 domain homologues also identified in the synaptic vesicle protein synaptotagmin11 was shown to perform a similar function.12,13 That the C2 domain endowed CaLB function agreed with initial characterizations of the Ca2+-independent nPKC isotypes, which appeared to lack the C2 domain.14 However, C2-like domains were later identified in nPKC isotypes, albeit at a different position in the primary structure.15 Thus it became clear that the presence of this domain, per se, was not responsible for Ca2+-regulated activity of the cPKC isotypes. For the purposes of this review, these C2-like domains in nPKC isotypes will be referred to as C2 domains. Such classification acknowledges their likely shared ancestry with the “classical” C2 domains of cPKC isotypes, as revealed by the overall similarity in their three-dimensional structures, as well as their common function as membrane-targeting domains.
C2 Domain Structure In the cPKC-α, -β and -γ isotypes, functional domains are ordered in the primary structure C1-C2-C3-C4, whereas in the nPKC-δ, -θ, -ε, and -η isotypes they are ordered C2-C1-C3-C4 (see Chapter 1). To date, structures of isolated C2 domains from four different PKC isotypes have been determined by X-ray crystallography, including PKC-α,16 -β,17 -δ,18 and -ε.19 As in other signaling proteins,20-22 these C2 domains are composed primarily of eight anti-parallel β-strands, packed tightly together into a two-sheet sandwich (Fig. 1). In addition, short α-helices are interspersed within the interstrand loops. The order of β-strands within the C2 domain primary structure differs in cPKC and nPKC isotypes due to a circular permutation of the strands (Fig. 2), segregating the domains into two distinct topological classes recognized in other signaling proteins.21 C2 domains of nPKC are classified as type I (or P type, for phospholipase) domains, a form found in a vast collection of eukaryotic proteins expressed in single cell and multi-cellular eukaryotes, which includes cPLA2-α and phosphoinositide-specific phospholipase C (PLC). In contrast, C2 domains of cPKC are classified as type II (or S type, for synaptotagmin) domains, a variant form present in other metazoan signaling proteins, including the multimember family of synaptotagmins.23 In these domains, the N-terminal β1-strand in type I domains is permuted to the C-terminus in the type II variant form, and strand connections are altered slightly so that this strand occupies the same topological position in both three-dimensional structures (Fig. 1). The N- and C-termini of the type I and II C2 domains are located at opposite edges of the β-sandwich (Fig. 1), suggesting that C2 domains are connected to the remaining domains in the full-length enzyme differently within these two classes of isotypes. This difference in connectivity may have arisen in response to the shuffling of regulatory C1 and C2 domains and functional constraints to maintain a fixed orientation of these two domains to the catalytic domain in the full-length enzyme. Despite the permutation of β-strands, the overall structures of both types of domains are similar, reflecting the presence of the highly conserved β-sandwich structural core (Fig. 1). For instance, superposition of the PKC-ε (type II) and PKC-α (type I) C2 domains reveals a root-mean-squared deviation (rsmd) of only 1.02 Å over 85 Cα atoms.19 Within type I cPKC C2 domains, high sequence identity generates considerably greater structural similarity: superposition of the PKC-α and -β C2 domains, which are ~80% identical in sequence, yields a rsmd of only 0.43 Å over 130 equivalent residues.16 The high degree of sequence similarity between the cPKC-γ and the -α and -β isotypes suggests that the structures of the cPKC C2 domains are most similar (Fig. 3). In contrast, sequence differences between type II nPKC C2
18
Protein Kinase C
Figure 1. PKC C2 domain structures. Ribbon representation of three-dimensional structures of PKC-α, -δ and -ε C2 domains determined by x-ray crystallography. Strands of the facing β-sheet are darkened. Within this sheet, the N-terminal β-strand of type I C2 domains of cPKC isotypes and the corresponding C-terminal β-strand of type II domains of nPKC isotypes are blackened to highlight their similar topological positions in the two domains. Ca2+ ions bound to cPKC C2 domains are indicated by spheres. Note the striking overall similarity in the domains, which results from a highly conserved structural core formed by the interface of the two β-pleated sheets, as well as considerable structural diversity in the conformations of the interstrand loops. These three domains are proposed to represent the three structural subclasses of PKC C2 domains (see (Fig. 3)). Structures were rendered by Ras Mac (v 2.6) using coordinates obtained from the Brookhaven Protein Data Bank (PDB): identification codes 1DSY (α),16 1BDY (δ),18 and 1GMI (ε).19
Figure 2. Topological arrangement of β-strands in PKC C2 domains. Numbering and topological arrangement of β-strands in the two-sheet sandwiches of type I and type II PKC C2 domains. β-strands are shaded as in (Fig. 1). Interstrand loops proposed to form the ligand-binding surfaces are designated L1-L4.
domains yields structures that are as dissimilar to each other as they are to all other C2 domains, including the cPKC C2 domains: superposition of the PKC-δ and -ε C2 domains, which share only 19% sequence identity, generates a larger rmsd of 1.36 Å over 66 equivalent
Structural and Functional Specialization of C2 Domains in Protein Kinase C
19
Figure 3. Sequence alignment of PKC C2 domains. Manual sequence alignment of PKC C2 domains guided by the positions of conserved β-strands that form the C2 domain structural core. Secondary structures are indicated above the sequences using the same shading scheme as in (Figs. 1 and 2). Interstrand loops proposed to form the ligand-binding surfaces are designated L1-L4. Interstrand α-helices that define the δ/θ and ε/η subclasses of nPKC C2 domains are indicated in brackets. The five conserved Ca2+-coordinating aspartates present in the cPKC, but not the nPKC C2 domains are highlighted.
residues.19 Consequently, the C2 domains of the PKC-δ and -ε isotypes have been proposed to represent two distinct subclasses of the type II PKC C2 structural family.19 The C2 domain of PKC-η has been proposed to resemble that of PKC-ε because of their 47% sequence identity and the presence of certain characteristic structural elements, notably an α-helix in the β1-β2 loop (Fig. 1).19 Likewise, sequence similarity between the PKC-δ and -θ isotypes (Fig. 3) and the presence of residues that form an α-helix in the β6-β7 loop present in the PKC-δ but not the -ε isotype (Fig. 1) together suggest that the PKC-δ and -θ C2 domains form a separate structural subclass. Segregation of the PKC-δ/θ and -ε/η C2 domains into two structural subclasses distinct from the PKC-α/β/γ domains parallels the divergence of the PKC-δ/θ and -ε/η isotypes into two evolutionary branches distinct from that leading to the cPKC isotypes.24 Generally in C2 domains, the major determinants of the ligand-docking surface are formed by variable interstrand loops on one edge of the domain sandwich.25-27 Another consequence of the circular permutation of β-strands in PKC C2 domains is that different numbers of these interstrand loops are projected outward from this edge of the sandwich: three loops, designated L1-L3, are present in type I and four loops, L1-L4, in type II C2 domains (Fig. 3). Major structural differences exist between nPKC and cPKC C2 domains that are localized to these loops,18 especially in the conformations of the loops themselves (see Fig. 1), which is likely to reflect mechanistic differences in their interactions with ligands. Even finer differences have been recognized between the PKC-δ/θ and -ε/η subclasses,19 although the significance of these differences remains to be elucidated. Most importantly, the full complement of aspartates required to coordinate two to three Ca2+ ions are present in L1 and L3 of type I cPKC C2 domains but not in the type II nPKC domains (Fig. 3). This latter distinction arises from the
20
Protein Kinase C
specific functional adaptations of the domains, rather than from structural constraints imparted by the circular permutation of β-strands, since there are numerous examples of Ca2+-dependent and Ca2+-independent phospholipid-binding type I and type II C2 domains.21
C2 Domain Interactions with Ligands General Recognition Properties of C2 Domains Prototypical C2 domains target proteins in which they are found to membrane surfaces, in many cases delivering enzymatic functions to the membrane where substrates are sequestered.21 As a whole, C2 domains bind to a diverse range of membrane or membrane-derived targets either constitutively or in response to Ca2+, including phospholipids, membrane-associated proteins and inositol polyphosphates.21,22 This feature is illustrated most clearly by the well characterized tandem C2 domains in the synaptic vesicle protein synaptotagmin. Synaptotagmin engages numerous distinct ligands through its C2 domains: the C2A domain binds syntaxin and anionic phospholipids and the C2B domain homodimerizes in a Ca2+-dependent manner, whereas the C2B domain binds the proteins AP-2, β-SNAP, N-type calcium channels, and SNAP-25, as well as inositol polyphosphates and phosphoinositide polyphosphates, in a Ca2+-independent manner.22 Likewise, recent biochemical studies show that C2 domains in the PKC family bind phospholipids or membrane-associated proteins, serving to either localize or activate the enzyme in response to cellular agonists.
Anionic Membrane Binding by cPKC C2 Domains
Isolated cPKC C2 domains bind anionic phospholipid membranes in a Ca2+-dependent manner in vitro,16,28-34 recapitulating the Ca2+ requirement for binding to anionic membranes characteristic of full-length cPKC isotypes (see Chapter 4). The isolated PKC-β C2 domain binds vesicles that contain phosphatidylserine (PS), phosphatidylglycerol (PG), phosphatidic acid (PA) or phophatidylinositol (PI) with only a slight preference for PS over other anionic lipids.32 Similar findings have been reported for the PKC-α C2 domain.29 Notably, the phospholipid preference of cPKC C2 domains contrasts strikingly with that of the cPLA2-α C2 domain, which binds primarily zwitterionic phosphatidylcholine (PC) or phosphatidylethanolamine (PE) vesicles in response to Ca2+,35 but parallels exactly that observed for numerous C2A domains of the synaptotagmins,13,33,36,37 which structurally are very similar to the cPKC C2 domains. In synaptotagmin, Ca2+ binding to the C2A domains has been proposed to act as an electrostatic switch that triggers anionic membrane binding.27,37 According to this model, in the absence of bound Ca2+ ions, the aspartate-lined cleft formed from L1-L3, the Ca2+-binding loops (CBLs), repels the anionic phospholipid surface, but upon the binding of multiple Ca2+ ions, the newly formed basic face formed by the loops and the bound Ca2+ is strongly attracted to the acidic surface of the phospholipid membrane, facilitating membrane docking.27 This mechanism is likely to govern the interaction of cPKC C2 domains with anionic membranes.27 As is the case for the full-length enzyme,38 Ca2+-induced binding of the PKC-β C2 domain to anionic membranes is inhibited in high ionic strength solutions, indicating that electrostatic interactions facilitate membrane docking.33 This mechanism contrasts strikingly with that underlying the Ca2+-driven interaction of the cPLA2-α C2 domain with zwitterionic PC or PE membranes, which is mediated largely through hydrophobic interactions and is insensitive to ionic strength.25,33,39-43 The magnitude of the dependence of PKC-β C2 domain binding on ionic strength suggests that, at a minimum, three ionic interactions are involved in formation of the domain-membrane ternary complex.33 Further illustrating the nonspecific electrostatic nature of phospholipid recognition by cPKC C2 domains is the finding that the slight preference for PS is not stereospecific (see Chapter 4): the PKC-β C2 domain binds in a Ca2+-dependent manner equally well to vesicles containing the natural PS isomer 1,2-sn-phosphatidyl-L-serine as to vesicles composed of the artificial PS
Structural and Functional Specialization of C2 Domains in Protein Kinase C
21
enantiomer 2,3-sn-phosphatidyl-D-serine.32 This property contrasts with the binding to anionic vesicles containing DAG or phorbol ester exhibited by the isolated PKC-β C1B domain, which stereoselectively recognizes phospholipid components, preferring the natural L-isomer of PS two-fold over the D-isomer and PS ten-fold over other anionic phospholipids.32 Although the isolated PKC-α C1B domain apparently displays no preference for PS over PG in binding vesicles containing DAG,29 together these findings indicate that the C2 domain does not, in general, dictate the stereoselective activation of full-length cPKC isotypes by PS.32,44 Thus, the main structural feature of the phospholipid head group recognized by cPKC C2 domains is its negative charge. Several residues believed to contact the anionic membrane surface have been identified in cPKC C2 domains. Functional studies have indicated that the Ca2+-binding loops L1-L3 constitute the primary determinants for membrane docking in numerous C2 domains, including those in cPLA2-α,25,41-43,45 synaptotagmin,46,47 and cPKC isotypes.48,49 Indeed, the structure of the Ca2+-bound PKC-α C2 domain crystallized in the presence of 1,2-dicaproylsn-phosphatidyl-L-serine (DCPS) revealed that the CBLs form several hydrogen bonds and hydrophobic interactions with this PS analogue.16 As functional studies showed,50 this PS-binding site does not correspond to one previously identified using anti-idiotypic antibodies.51 Model building suggested that PKC-α residues Asn-189, Arg-216, Thr-251 and Arg-249 make direct contact with PS.16 The importance of these residues in activating full-length PKC-α has been substantiated by findings that mutation of these residues results in functional impairment of PKC-α.49 Multiple basic (Arg-249 and Arg-252) and aromatic (Trp-245 and Trp-247) residues in the CBLs of the PKC-α C2 domain have been implicated in providing electrostatic and hydrophobic interactions, respectively, between the domain and the anionic membrane.48 Increasing the basicity of the CBLs by replacement of two Ca2+-coordinating aspartates with arginine serves to decrease the apparent Ca2+ affinity of full-length PKC-β without increasing its intrinsic affinity for anionic membranes, suggesting that ionic interactions alone may not account for the membrane docking mechanism of cPKC C2 domains.52 Moreover, the presence of a Ca2+ ion bridging the isolated PKC-α C2 domain to a PS analogue in the domain-(Ca2+)2-DCPS ternary complex16 suggests that Ca2+ plays a greater role than merely acting as an electrostatic switch. How much of the membrane docking energy can be attributed to the large change in electrostatic potential induced by Ca2+ binding or formation of the Ca2+ bridge has begun to be addressed through electrostatic calculations.27 Certainly, understanding the central role played by Ca2+ in inducing anionic membrane binding by cPKC is hampered severely by the simple fact that at present, the exact number of Ca2+ ions involved in anionic membrane binding is not known.
Ca2+ Binding by cPKC C2 Domains
Crystallographic studies have yielded exquisite structural details of Ca2+ binding to the cPKC C2 domains, but, as will be illustrated below, provide an incomplete picture that requires complementary mechanistic investigation. Ca2+ ions bind to these domains in a highly negatively charged cleft formed by oxygen ligands provided by five aspartates embedded within CBLs 1 and 3 (Fig. 4). Three Ca2+ ions are present in the crystal of the isolated PKC-β C2 domain17 but only two of these appear in the PKC-α domain, alone or bound in a complex with DCPS.16 These five coordinating positions provide mono- and bidentate coordination with the Ca2+ ions and are conserved in L1 and L3 throughout cPKCs but are not present in nPKCs (Fig. 3). Additional oxygen ligands are provided by several main chain carbonyl groups, and, in the PKC-β C2 domain, by the side chain hydroxyl of Ser-251 in CBL3 (Fig. 4). As expected, substitution of the five coordinating aspartates individually with asparagine impairs the ability of Ca2+ to induce anionic vesicle binding, enzymatic activity, or plasma membrane translocation of full-length PKC-α.30,48,49 Generally, four distinct Ca2+-binding sites appear to be occupied to varying extents in different C2 domains, including those in cPKC. Of these four common sites, numbered I to IV
22
Protein Kinase C
Figure 4. Coordination of Ca2+ ions by the PKC-β C2 domain. Left. Space-filling representation of the PKC-β C2 domain, viewing into the Ca2+-binding cleft. CBLs 1-3 are shaded grey, and Ca2+ ions are numbered II-IV. Coordinating oxygens are black. Structure was rendered by Ras Mac (v 2.6) using coordinates of PDB file 1A25.17 Right. Schematic of observed Ca2+-coordination geometry (adapted from ref. 17). Ca2+ ions are numbered II-IV. Observed coordinating protein oxygen ligands are black; missing ligands are shaded light grey. Protein ligands are labeled to indicate contribution from side chains or, when preceded with “O”, from main chain carbonyls provided by CBL3; those provided by CBL3 are labeled in black and those from CBL1 in grey. Coordination geometry of CaII and CaIII approaches pentagonal-bipyramidal and that of CaIV approaches square-bipyramidal. Missing axial ligands may be provided by phospholipid oxygens in the domain-membrane ternary complex.
using a standardized system,17,53 at a minimum, sites II, III and IV are utilized in the PKC-β C2 domain (Fig. 4).17 Only sites II and III are filled with Ca2+ in the isolated PKC-α domain complexed with DCPS; in site IV, a water molecule replaces the Ca2+ ion (water-301, PDB code 1DSY, ref. 16). In the related type I synaptotagmin C2A domain, at a minimum, these same three Ca2+-binding sites, II, III and IV, are occupied.54 In contrast, a different subset of Ca2+ ions are known to be utilized in the type II C2 domains: sites I, II and III are filled in PLC-δ1, whereas only sites I and II are occupied in the cPLA2-α C2 domain.40,45,55 In the cPLA2-α C2 domain, sites III and IV are likely to have been “inactivated” by natural substitution of asparagine for aspartate at position 95 and valine for serine at position 97 in CBL3; these residues correspond to Asp-248 and Ser-251 in PKC-β, respectively.17 In the cPLA2-α and PLC-δ1 C2 domains, coordination of CaI is assisted by oxygen ligands provided by residues in CBL2 that are not present in the type I C2 domains of cPKC and the synaptotagmins. In these latter C2 domains, the shorter CBL2 has been suggested to play a supporting structural role.18 Thus, occupancy of these four Ca2+ sites is shaped by natural elimination or generation of new potential coordinating ligands. In cPKC C2 domains, it is not clear how the Ca2+ stoichiometry observed crystallographically corresponds to that achieved under physiological conditions, that is, when the domain is bound to target phospholipid membranes. In the case of the cPLA2-α C2 domain, the stoichiometry of two Ca2+ ions observed in crystal structures has been confirmed independently by studies employing equilibrium dialysis and stopped-flow fluorescence spectroscopy33,56 and
Structural and Functional Specialization of C2 Domains in Protein Kinase C
23
calorimetry.45 Importantly, these methods have also been used to show that two Ca2+ ions are bound in the cPLA2-α C2 domain-membrane ternary complex.33,56 In contrast, for the cPKC and synaptotagmin C2 domains, the exact Ca2+ stoichiometry achieved when the domain is bound to vesicles is currently not known. Results of a stopped-flow fluorescence spectroscopy investigation have shown that a minimum of three Ca2+ ions dissociate from the PKC-β C2 domain-membrane ternary complex.33 Notably, this study could not rule out the existence of additional Ca2+ ions that are kinetically “invisible” due to their rapid dissociation from the domain-membrane ternary complex during the stopped-flow apparatus “dead time.” The results of this study provided, at a minimum, biochemical confirmation in solution of the Ca2+ stoichiometry observed crystallographically.17 In many C2 domains, suitable oxygen ligands are present that could potentially coordinate additional physiologically relevant Ca2+ ions22 but do not do so in solution or in crystals, perhaps because the isolated, membrane-free domain binds these Ca2+ ions with too low an affinity. Such may be the case for a hypothetical Ca2+ ion that is “missing” from site IV in the PKC-α C2 domain, which is occupied by only two ions (II and III) in crystals.16 In this domain, a potentially conservative threonine substitutes for Ser-251, whose side chain oxygen assists in coordination of CaIV in the PKC-β C2 domain (Fig. 4). Although postulated to directly contact PS,49 Thr-251 in PKC-α may instead assist in coordination of CaIV, whose coordination geometry approaches an octahedral arrangement in the PKC-β domain.17 For CaIV bound to the PKC-β C2 domain, one apical oxygen ligand is provided by another C2 domain in the asymmetric unit of the crystal.17 It is likely that this Ca2+ ion binds to the free domain only poorly, as it does in the case of the synaptotagmin C2A domain.54 It may be stabilized in the ternary domain-membrane complex by other protein oxygen ligands, as in crystals of the PKC-β C2 domain, or by phosphoryl oxygens provided by the bound phospholipid or phosphorylated residues elsewhere in the full-length protein. Spectroscopic measurements have confirmed that the isolated PKC-β C2 domain binds Ca2+ relatively weakly in the absence of target membranes (Kd ~ 60 µM) and that the apparent Ca2+ affinity increases several-fold in their presence (discussed below).33 Remarkably, even for the much studied synaptotagmin C2 domains, the relevant physiological Ca2+ stoichiometry remains ambiguous: the isolated C2A domain has been observed to bind only one Ca2+ ion in crystals57 and three in solution.33,54 Moreover, it has been suggested that all four Ca2+ ions might bind to the domain in the ternary domain-membrane complex.22 Likewise, as many as four Ca2+ ions have been proposed to bind to the PLC-δ1 C2 domain, even though only two Ca2+ ions (CaI and II) or three La3+ ions (corresponding to CaI, II and III) are observed to bind in crystals53 or in solution.58 Extending this possibility to the cPKC C2 domains, suitable coordinating ligands, conserved throughout the cPKC C2 domains, are present that may provide low affinity binding of CaI in the domain-membrane ternary complex. (The position of CaI is represented by water molecules in the crystal structures of the isolated cPKC C2 domains; these molecules are designated water-328 in the PKC-α domain [PDB code 1DSY, ref. 16] and water-6 and -22 in the A and B chains, respectively, of the PKC-β domain [PDB code 1A25, ref. 17]). These oxygens, provided by the main-chain carbonyl of Ser-217 and the side-chain carboxylates of Asp-193 and Asp-187 and the carbonyl of Asn-189, are favorably positioned such that only slight reorientation could generate vertices of a suitable coordination array. Further biochemical experiments will be required to determine the actual Ca2+ stoichiometry when cPKC C2 domains are bound to anionic phospholipid membranes. Although it is not clear why C2 domain docking to membranes requires the binding of multiple Ca2+ ions to the domain, it is likely that only multiple ions can provide the necessary change in electrostatic potential that is required for stable association of the domain with anionic membranes.27 Secondly, these additional Ca2+ ions may also increase the complexity of the “Ca2+ tuning” properties of the C2 domains, as is often the case for Ca2+ signaling proteins.59,60 For instance, the binding of multiple Ca2+ ions creates the opportunity for positive and negative cooperativity,
24
Protein Kinase C
binding properties that might modulate the activity of the C2 domain in response to changes in Ca2+ levels. Specifically, the domain may tuned as a high sensitivity “on-off ” switch by binding Ca2+ in a positively cooperative manner, docking to membranes over a narrow Ca2+ range. As will be discussed below, the step-wise binding of Ca2+ ions to distinct sites also enables the PKC-β C2 domain to display both fast dissociation kinetics when Ca2+ levels are rapidly reduced as well as slow dissociation kinetics when elevated Ca2+ levels are sustained. Finally, each of these Ca2+- and membrane-binding properties may be differentially tuned in various cPKC isotypes to allow functional specialization for various signaling pathways, a general principle displayed by C2 domains in distinct Ca2+ signaling proteins.33 Also remaining to be understood is the mechanism by which the domain remains inactive in the presence of 103- to 104-fold higher concentrations of physiological monovalent K+ and Na+ and divalent Mg2+ ions. Although Mg2+ is very similar in size to Ca2+-effective ionic radii are 0.89 Å and 1.09 Å for Mg2+ and Ca2+, respectively (see ref. 59)—it does not effectively trigger target membrane binding of most C2 domains, including the PKC-β C2 domain.32 Detailed investigation into the cation preferences of the cPLA2-α C2 domain for metal binding and membrane docking has led to a size exclusion and charge selectivity model for cation binding,61 certain features of which may be applicable to cPKC C2 domains. According to this model, in the absence of bound Ca2+ ions, strong electrostatic repulsion between several aspartates in the Ca2+-binding cleft of the cPKC C2 domains increases the inherent flexibility of the CBLs. The lower charge of monovalent ions, such as K+ and Na+, is not able to overcome the electrostatic repulsion between the oxygens; however, the greater charges of divalent, and probably trivalent spherical cations can. Mg2+ ions are too small to functionally replace Ca2+: steric constraints prevent the cation binding site from contracting sufficiently to engulf this smaller ion. Investigation into the metal ion charge and size selectivity of cPKC C2 domains promises to yield greater insight into the central role played by Ca2+ in membrane binding.
Constitutive Anionic Phospholipid Binding by nPKC C2 Domains In comparison to the cPKC C2 domains, considerably less is known about the ligand-binding properties of nPKC C2 domains, perhaps because of the relative ease in which reversible Ca2+-regulated binding of the cPKC C2 domains to anionic membranes can be demonstrated in vitro. Of the five aspartates present in the Ca2+-coordinating positions of the PKC-α and -β C2 domains, also shared with many of the C2 domains of the synaptotagmins, only one is present in the C2 domains of nPKC isotypes (see Fig. 3).18 The loss of these aspartates presumably accounts for the failure of the latter C2 domains to be regulated by Ca2+.18 Nevertheless, given the differences in the conformations of loops L1-L3 in cPKC and nPKC C2 domains (see Fig. 1), it is likely an over-simplification to regard nPKC C2 domains as cPKC C2 domains in which the coordinating aspartates have been excised. To date, interactions between nPKC C2 domains and anionic phospholipids have been best characterized for the PKC-ε C2 domain. These studies have relied exclusively on the “classical” approach used to measure binding of C2 domains to phospholipid vesicles in vitro, whereby fusion proteins consisting of glutathione S-transferase linked to the C2 domain are first bound to glutathione-conjugated beads and then the ability of these beads to cosediment vesicles containing trace amounts of radioactive PC is measured.13 Using this assay, the PKC-ε C2 domain has been shown to bind anionic phospholipid vesicles containing PIP2, PI, PA, PG or PS in a Ca2+ independent manner, with slight preference for PA.62 It has been reported that interstrand loops L1 and L3 are required for this interaction, since three engineered variants in these loops displayed reduced anionic phospholipid binding: the deletion of residues Trp-23 to Pro-34, the triple substitution of Trp-23, Arg-26 Arg-32 to alanine, or the double substitution of Trp-23 and Arg-26 to alanine each reduced the amount of vesicle binding compared to the wild-type PKC-ε C2 domain.19 For the three binding variants, half-maximal phospholipid binding was achieved at nearly the same percentage of PA as for the wild-type C2 domain, suggesting that the mutations had only little affect on phospholipid binding.
Structural and Functional Specialization of C2 Domains in Protein Kinase C
25
It will be important to test the ability of isolated C2 domains from other nPKC isotypes to bind anionic phospholipids and to relate their binding affinities to those of the cPKC C2 domains and the full-length enzymes in which they are found. Differences are expected, given that, for instance, the binding of PKC-α in the presence of 50 µM Ca2+ to mixed PC/PS vesicles is 18-fold tighter than that of PKC-ε.63 Electrostatic calculations suggest that the affinity of the PKC-δ C2 domain for PS should be weak but that membrane binding might be facilitated by a combination of hydrophobic and electrostatic interactions.27 Moreover, the striking conformational differences in the ligand-binding loops of the PKC-δ/θ and -ε/η C2 domains may reflect functional specialization within these subclasses. Still unexplained is the presence of a bound Mg2+ ion in the crystal of the PKC-ε C2 domain, which suggests that divalent cations might somehow regulate nPKC C2 domains, although apparently not their interaction with PA vesicles.19 Since translocation of certain nPKC isotypes in vivo is regulated by phosphorylation of the C2 domain,64 it will be of interest to test how this modification affects interactions with anionic vesicles. Such comparisons will allow a deeper understanding of the differentiation between cPKC and nPKC isotypes.
Protein Binding by PKC C2 Domains The enzymatic activities of several PKC isotypes are regulated by interactions with other intracellular proteins (see Chapter 5). As is the case for other C2 domains, exemplified by the Ca2+-dependent anionic membrane-binding C2 domains of synaptotagmin, PKC C2 domains have been reported to bind to other proteins in vitro and in vivo. Some of these protein interactions may play a role in cellular localization of PKC: for example, the PKC-δ C2 domain forms a binding site for actin, which serves to localize PKC-δ to the leading edge of cells.65 Other interactions may be directed towards PKC substrates themselves: the C2 domain of PKC-δ mediates an interaction with GAP-43, one of its substrates.66 The binding sites for these proteins may reside only partly within the C2 domain. For instance, at least part of the binding site for the membrane-associated receptor of activated C-kinase (RACK1) resides in the C2 domain of PKC-βII.67 The V5 region of PKC also contributes to this interaction.68 RACK1 also associates with the PH and C2 domains of the p120 GTPase-activating protein.69 Furthermore, the cytoskeletal protein calponin, which activates both PKC-ε and -α in vitro, binds to the C2 domains of these isotypes70 and may also interact with the C1B domain. These findings illustrate two important properties of PKC regulatory domains. First, just as PKC engages phospholipid membranes though multiple contacts mediated by both the C1 and C2 domains,71 PKC is likely to interact with cellular proteins through discontinuous determinants spanning surfaces of these same domains. Secondly, although they are functional modules that have been independently shuffled into several diverse signaling proteins and bind distinct ligands, C1 and C2 domains are likely to function interdependently in PKC. Future experiments that address the affinity and kinetics of PKC binding to these proteins will assist in our understanding of the physiologically role played by these interactions in vivo.
Mechanism of Membrane Targeting of PKC by C2 Domains Recent investigations into the mechanism of anionic membrane binding by PKC C2 domains have begun to shed light on the central role played by the domain in targeting PKC to anionic membranes in vivo. For the cPKC isotypes, studies employing equilibrium and rapid kinetic techniques have provided a level of understanding not achievable previously. The following will highlight these findings, relate them to the properties of full-length enzymes, and finally these findings will be interpreted in light of properties of PKC isotypes in vivo.
Ca2+-Triggered Anionic Membrane Binding by cPKC
Results of mechanistic investigations into the process by which Ca2+ triggers membrane binding by PKC-βΙΙ can serve to illustrate this class of enzyme. At first glance, it might
Protein Kinase C
26
appear that the pathway for binding of its isolated C2 domain to anionic membranes in response to Ca2+ would be simple. However, since at least three Ca2+ ions are trapped within the PKC-β C2 domain-membrane ternary complex, there are at least four possible macroscopic pathways leading to membrane (L) binding (and considerably more microscopic pathways as well), depicted by: C2 • Ca 0 • L
→ ←
↓↑ C2 • Ca 0
C2 • Ca 1 • L
→ ←
C2 • Ca 2 • L
↓↑ → ←
→ ←
↓↑
C2 • Ca 1
→ ←
C2 • Ca 2
C2 • Ca 3 • L ↓↑
→ ←
C2 • Ca 3
Scheme I where C2•Can and C2•Can•L represent membrane-free and -bound states of the C2 domain and n denotes the Ca2+ stoichiometry. A similar minimal complexity would also apply for the full-length enzyme. It is worth noting at the outset that distinct pathways might proceed under different Ca2+ concentrations, depending on the relative rates of the various steps. These differences may be of physiological relevance, considering, for instance, that Ca2+ concentrations near the membrane surface will be considerably larger than that in bulk solution as a consequence of the membrane surface potential generated by anionic phospholipids.72,73
β C2 Domain Ca2+ Binding to the Isolated PKC-β
Ca2+ binding to the PKC-β C2 domain in vitro has been measured indirectly via fluorescence spectroscopy, whereby the increase in intrinsic tryptophan fluorescence emission of the domain resulting from Ca2+ binding is monitored.33,34 In the absence of anionic target membranes, multiple Ca2+ ions bind to the isolated PKC-β C2 domain with low affinity in a positively cooperative manner. The binding of these Ca2+ ions to the isolated domain may be depicted by: C2 + nCa
k on(Ca) → ← k off (Ca)
C2 • Ca n
Scheme II 2+
where n reflects the uncertainty in the Ca stoichiometry, and kon(Ca) and koff(Ca) represent net second-order association and dissociation rate constants for Ca2+ binding, respectively. An apparent dissociation constant Kd equal to 39 - 60 µM for these Ca2+ ions has been measured. Such a constant must be considered an approximation of the average Ca2+ affinity for the binding of multiple Ca2+ ions, since such fluorescence approaches, unlike equilibrium dialysis, do not measure stoichiometric binding of individual Ca2+ ions. Nevertheless, it is noteworthy that this value is similar to the macroscopic dissociation constant observed for the binding of the first Ca2+ ion to the synaptotagmin C2A domain (Kd = 60 - 75 µM, refs. 28, 54, 74) and the cPLA2-α C2 domain (Kd = 60 µM, ref. 56). The similarity in these values may indicate that each of these C2 domains first binds CaII, which is common to all C2 domains that bind Ca2+. However, this value is considerably larger than the net apparent Kd of 11 - 24 µM for the binding of two Ca2+ ions to the isolated cPLA2-α C2 domain measured in the absence of membranes,33,45,56,75 indicating that the PKC-β and cPLA2-α C2 domains are specialized as low and high affinity Ca2+ sensors, respectively.33 Since the presence of Ca2+ increases the affinity of the C2 domain for anionic target membranes, the presence of target membranes increases the apparent Ca2+ affinity, as predicted by laws of mass action: membranes decrease the apparent Ca2+ Kd of the PKC-β C2 domain several-fold,33 an effect also observed for the cPLA2-α C2 domain.56 The association and dissociation kinetics of Ca2+ binding to the isolated PKC-β C2 domain are both rapid, consistent with the hypothesis that no slow conformational changes are required for Ca2+ binding to the domain. The time course of irreversible Ca2+ dissociation from
Structural and Functional Specialization of C2 Domains in Protein Kinase C
27
the membrane-free PKC-β C2 domain has been measured in a Ca2+-trapping approach, whereby the Ca2+-occupied domain is rapidly mixed via stopped-flow with a Ca2+ chelator that rapidly binds free Ca2+ ions, and the loss in intrinsic tryptophan emission is recorded.33,34 The apparent rate constant for Ca2+ dissociation from the membrane-free PKC-β C2 domain (koff(Ca)) is very large, estimated by to be >> 500 s-1. Such rapid Ca2+ dissociation strikingly contrasts with that for the two Ca2+ ions that are released from the membrane-free cPLA2-α domain, in which the apparent koff(Ca) is 110 s-1.56 Ca2+ dissociation from the cPLA2-α C2 domain has been proposed to proceed via an ordered-sequential mechanism,56,61 whereby rate-limiting dissociation of one ion, presumably CaI (koff(CaI) ~ 110 s-1), precedes the rapid dissociation of the next (CaII; koff(CaII) >> 110 s-1). Assuming that Ca2+ ions likewise dissociate from the isolated PKC-β C2 domain in an ordered-sequential manner, the larger koff(Ca) for PKC-β is likely due to faster dissociation of the rate-limiting Ca2+ ion and is responsible its lower Ca2+ affinity in the absence of membranes. The apparent rate constant for Ca2+ association with the isolated PKC-β C2 domain (kon(Ca)) is very large. As is the case for many Ca2+-binding proteins, the rate constant for Ca2+ association cannot be directly measured but may be estimated based on a Ca2+-mixing experiment. In this approach, Ca2+ is rapidly mixed via stopped-flow with the apo-C2 domain and the observed rate at which the binding reaction approaches equilibrium (kobs) is measured by monitoring intrinsic tryptophan emission.34 In principle, if Ca2+ binding is rate-limiting, under pseudo-first order conditions kobs should be dependent on kon(Ca), koff(Ca) and the concentration of Ca2+. However, the time course of the Ca2+-mixing experiment was very short, preventing determination of this dependence: even at very low Ca2+ concentrations, the reaction is completed in less than ~ 1 ms. Considering that the Ca2+ dissociation constant of the PKC-β C2 domain is ~ 60 µM and that the koff(Ca) is estimated to be >> 500 s-1 by the Ca2+-trapping approach, then the apparent kon(Ca) for the free PKC-β C2 domain in solution must be at least 1 x 107 M-1 s-1. This value is comparable to that reported for many other Ca2+-binding proteins (106 to 109 M-1 s-1) where the observed association rate for Ca2+ binding often approaches the diffusion-controlled limit.59,60
β C2 Domain Anionic Membrane Binding by the PKC-β
The time course of Ca2+-initiated membrane binding by the PKC-β C2 domain has been measured by a vesicle-mixing approach, whereby C2 domain preincubated with Ca2+ is rapidly mixed via stopped-flow with anionic phospholipid vesicles.33,34 In this experiment, the observed rate of approach to binding equilibrium (kobs) is measured, either by monitoring increases in intrinsic tryptophan emission elicited by membrane docking or by fluorescence resonance energy transfer (FRET) from donor tryptophans in the domain to trace amounts of acceptor dansyls incorporated into the vesicles.76,77 In the presence of Ca2+, the binding of the C2 domain to target membranes proceeds via a single kinetically observable step. kobs varies linearly with vesicle concentration via the relationship: (1)
k obs = k on [ v]+ k off
where kon and koff represent the apparent membrane association and dissociation rate constants. Since kobs does not depend on whether Ca2+ is preincubated with the domain or provided along with the vesicles in this experiment, Ca2+ binding to the free domain does not appear to be rate-limiting under these conditions. This suggests that the rate-limiting step in membrane docking is the bimolecular interaction between a (partially) Ca2+-occupied C2 domain and the phospholipid vesicle, which may be represented by: C2 • Ca n + L
k
on → ← k off
C2 • Ca n • L
Scheme III
Protein Kinase C
28
For binding to 100 nm vesicles composed of 40 mol % anionic phospholipids in the presence of 200 µM Ca2+, kon and koff values of 1.2 x 1010 M-1 s-1 and 8.9 s-1 have been measured, respectively.34 The resulting calculated Kd value of 0.73 nM is reasonably close to the Kd estimation of 0.40 nM based on the observed amplitudes of the fluorescence changes. Importantly, the koff value was confirmed independently by a C2 domain-trapping approach, whereby a solution containing the C2 domain prebound to fluorescently labeled anionic vesicles in the presence of Ca2+ was rapidly mixed with a solution containing a large molar excess of unlabeled vesicles. The rate constant observed in this experiment, equal to 9.2 s-1, agrees well with the value of 8.9 s-1 from the vesicle-mixing approach. The influence of Ca2+, the mol fraction of anionic phospholipid in vesicles and the presence of Mg2+ in the binding reaction on both kon and koff sheds light on the mechanism of anionic membrane binding by the PKC-β C2 domain.34 Increasing the Ca2+ concentration or the mol fraction of anionic phospholipids both raise the membrane affinity, in each case both by elevating kon and by reducing koff. Conversely, the presence of Mg2+, which reduces the surface electrostatic potential by nonspecifically accumulating near the membrane surface,72,73 reduces the membrane affinity by both diminishing kon and raising koff. The effect of Ca2+, the mol fraction of anionic phospholipid in vesicles and the presence of Mg2+ have been attributed to the favorable role played by electrostatic interactions in membrane binding34 and are consistent with recent electrostatic calculations.27
Theoretical Rate of C2 Domain Association with Vesicles Based on collisional theory, the diffusion-controlled limit for the binding of the C2 domain to large unilamellar vesicles (LUVs) produced by extrusion may be calculated from the Einstein-Smoluchowski relation (see ref. 78):
(
)(
k calc a ≅ 4πN A R p + R v D p + D v
)
(2) 23
-1
where NA represents Avogadro’s number (6.023 × 10 mol ), Rp and Rv and represent the radii of the C2 domain and the vesicle, respectively, and Dp and Dv represent the diffusion constants for the C2 domain and the vesicle, respectively. Rv can be taken as 50 nm, based on the diameter of vesicles produced by extrusion through 0.1 µm membranes (see ref. 79). Assuming its shape is spherical, Rp can be calculated from its volume (Vp) using the relationship: 1
3 3 Rp = Vp 4π
(3)
Based on the molecular weight (MW) of the C2 domain, its volume can be estimated as 1.27 MW Å3 da-1 (see ref. 80), yielding Rp equal to 17 Å. Thus, since Rp 106 clones screened, 10 clones were isolated and all are in vitro PKC substrates; most of these have now been shown to be in vivo substrates as well (individual examples are discussed below). Since these proteins bind to PKC and are also phosphorylated by PKC, we named them STICKs for Substrates That Interact with C-Kinase. What is the function of direct PKC-substrate interactions? PKCs are relatively indiscriminate kinases in vitro. Many proteins that are not physiological substrates can be phosphorylated by PKC in vitro. Given this lack of fidelity, substrate binding may be an important mechanism for restricting substrate accessibility. There are now several examples of direct binding of kinases to their physiological substrates including Mitogen-activated protein kinases, Phosphoinositide-dependent kinase-1 (PDK-1) and Src.6,7,8 Thus, like other kinases, PKC may rely on direct, high affinity interactions between PKCs and its substrates to limit PKC action to only appropriate substrates in vivo. Such interactions may also enhance the efficiency of phosphorylation.
Receptors for Activated C-Kinase (RACKs) A separate class of PKC binding proteins named RACKs was also identified and cloned using an overlay assay.9 These authors used phosphorylation of the proteins bound to PKC (instead of antibody binding) to identify the PKC interacting proteins. Like STICKs, RACKs
52
Protein Kinase C
form Phosphatidylserine (PS)-dependent interactions with PKC in the overlay assay.10 Unlike STICKs, RACKs selectively bind PKC isotypes, but are not PKC substrates. In vivo, RACKs are thought to anchor translocated (active) PKCs in specific membrane domains in the proximity of appropriate substrate proteins, and consequently promote phosphorylation (see below). Mapping studies showed that WD40 motifs in isozyme-selective RACKs for PKC-β (RACK1)11 and PKC-ε (β’-COP10) interact with unique peptide sequences in regulatory domains of individual PKCs. When added to permeabilized cell preparations, these PKC peptides can interfere with PKC redistribution. In more recent studies, additional RACKI interacting sites in the V5 domain of the catalytic domain were also identified.12 The effects of these regulatory and catalytic domain peptides on PKC subcellular location demonstrate their potential to interfere with PKC functions; however, effects on substrate phosphorylation and PKC-mediated biological responses have not yet been characterized. It is not yet known if RACKs are essential for PKC-mediated responses. Nonetheless, the isozyme-selective binding properties of RACKs indicate that they may play a role in isozyme-selective location and function in living cells.
INAD In any signaling event, both activation and deactivation processes are required for proper temporal control. Timing is especially important in phototransduction where signals rapidly activated in response to light must be quickly attenuated in order to reset the system. In order to attain the high speed and specificity needed, key elements of the light activated signaling pathway in Drosophila are organized into transduction complexes through their common affinity for the scaffold protein INAD. Inad was identified in genetic screens for mutants with inactivation no after potential. This unique molecule consists entirely of 5 PDZ domains that directly bind to eye PKC, PLC and the major light activated calcium channel TRP. INAD binding of each of these components is essential for coordinated localization of these signaling molecules and vision in flies.13 The critical dependence on INAD and its organizing functions emphasize the importance of localization as a determinant of PKC activity, as well as other signaling molecules. INAD PDZ2, PDZ3 and PDZ4 can each bind to eye PKC binds through a type 1 PDZ ligand located in the PKC C-terminus.14 Similar interactions between mammalian PKCs and PDZ domains in PICK1, ASIP and PAR-3 indicate that PKC-PDZ domain interactions may be a common targeting strategy. INAD PDZ 1 and PDZ5 interact with the PLC-β homologue NorpA, whereas PDZ3 and PDZ4 are responsible for recruiting the ion channel TRP to the complex. Biochemical and genetic studies indicate that INAD signaling complexes are preassembled and poised for rapid activation. The assembled complexes must then be targeted to the appropriate location. Determinants on TRP as well as INAD appear to contribute to the binding of the complexes to rhabdomeres, where signaling occurs.79 There is some evidence that the INAD complex directly interacts with the membrane associated phoro-receptor rhodopsin. Light-activated rhodopsin interacts with a heterotrimeric G protein that then dissociates into the signaling competent Gαq-GTP and Gβγ complexes. The free Gαq-GTP associates with the INAD signaling complex where it activates the INAD-bound PLC to hydrolyze phosphoinositides. Concurrently, PLC acts as an GTPase-activating protein (GAP) to facilitate GTP hydrolysis and the rapid cycling of the phototransduction pathway.80 Subsequently, the inactive Gαq-GDP dissociates from the complex. Increased PLC activity has been associated with localized gating of the TRP calcium channels, whereas PKC activity is required to attenuate the calcium signal and the visual cycle response.13,79 Both INAD and TRP appear to phosphorylated by eye PKC, although the sites of phosphorylation and functional modification have not yet been described.81 The assembly of eye PKC with other signaling molecules on a protein scaffold provides a paradigm for understanding the role of PKC in spatial and temporal control of signaling event.
Protein Kinase C Regulation by Protein Interactions
53
Myristoylated Alanine-Rich C-Kinase Substrate (MARCKS), MARCKS-Related Protein (MRP) and Adducins MARCKS, was one of the first physiological PKC substrates to be identified. The PKC phosphorylation sites are within a highly positively charged domain (the phosphorylation site domain (PSD)) that is also a site for MARCKS functional activities. These activities include binding to actin, phosphatidylinositol-4,5-bisphosphate (PIP2) and calmodulin. Phosphorylation by PKC decreases each of these activities.15 MARCKS, but not phosphorylated MARCKS, binds PKC in the overlay assay.16 MRP, α-adducin and γ-adducin, each of which contains a homologous PSD sequence, were also cloned as PKC binding proteins by the overlay assay method. Taken together, these results indicate that the PSD is a determinant of PKC binding to MARCKS and related proteins. These proteins are therefore classified as STICKs. Phosphorylation by PKC decreases MARCKS affinity for actin, and this results in substantial inhibition of actin cross-linking and actin bundling.17,18 The effects on actin organization may account for the ability of overexpressed MARCKS to decrease cell adhesion and promote spreading.19 The unphosphorylated PSD peptide also binds PIP2. PIP2 binding is sufficiently strong to sequester PIP2 from PLC and thereby interfere with its hydrolysis by PLC.20 PIP2 binding, along with the high affinity of the N-terminal myristoyl group for lipids, contributes to the membrane-association of unphosphorylated MARCKS. As PKC becomes active and MARCKS becomes phosphorylated, MARCKS affinity for PIP2 is decreased. Increased phosphorylation and decreased PIP2 binding correlate well with redistribution of MARCKS from the membrane to the cytosol in stimulated cells.20 The combination of effects of MARCKS on dynamic cell processes (actin organization, cell spreading, PIP2 binding and inhibition of endogenous PLC) along with the dynamic trafficking of MARCKS between membrane and cytosol cellular compartments argues that the cycles of MARCKS phosphorylation/dephosphorylation play a role in dynamic cellular processes. In fact, studies with phosphatase inhibitors demonstrate that the phosphorylation state of MARCKS is dynamically regulated. In secretory cells, cytosolic MARCKS is rapidly dephosphorylated by a process that is regulated by cGMP.21 It has been suggested that this cycle of phosphorylation/dephosphorylation and the associated changes in cytoplasm/membrane compartmentalization are a means of shuttling MARCKS between internal membrane compartments. PKCs role in this process would be to regulate the rate of cycling and the relative levels of cytosolic and membrane-bound MARCKS. MRP, also known as MacMARCKS contains a homologous effector domain and is also an actin, phospholipid and PKC binding protein. Like its brother, changes in MRP phosphorylation have been linked to changes in cell spreading, possibly through regulation of β2-integrins.22 PKC activation indirectly stimulates diffusion of β2 integrins within cell membranes by modulating cytoskeletal complexes that constrain integrin diffusion within the membranes. The PKC-dependent cytoskeletal remodeling is at least in part due to PKC regulation of interactions between MRP and dynamitin, a cytoskeletal-associated protein that regulates microtubule-dependent motor functions.23,24 Copurification, coimmunoprecipitation and fluorescence resonance energy transfer experiments all demonstrate that the PSD domain of MRP directly interacts with the N-terminal domain of dynamitin. Phosphorylation of MRP (or calcium/calmodulin binding) can disrupt this interaction, facilitate β2-integrin diffusion and promote cell spreading. Thus, similar to MARCKS, PKC appears to regulate MRP’s function in controlling dynamic cytoskeletal processes. Adducins were originally isolated as components of the red blood cell membrane skeleton and named according to their ability to promote interactions between actin and spectrin. Three isoforms of adducins have been isolated (α, β, and γ) and each contains an N-terminal globular head, a linker region and a C-terminal tail which contains the MARCKS homologous PSD sequence. Adducin functions as a heterotetramer, and phosphorylation by PKC interferes with its activities such as actin binding, actin capping, and spectrin recruitment to the actin short filaments. The PSD domain is required for each of these functions. Adducin is also phosphorylated at other sites by rho-kinase, which promotes adducin-actin interactions.25 Direct
54
Protein Kinase C
interactions with the myosin binding subunit of Protein Phosphatase 1 (PP1) have also been described.26 Thus, the relative activities of these kinases and phosphatases are reflected in the net phosphorylation state of adducin. In this way, adducin becomes a determinant of localized cortical actin polymerization dynamics. The PKC-phosphorylated form of adducin can be selectively monitored with phosphorylation state selective adducin antibodies. Levels of phosphorylated adducin (ser 726, 713 or 660 in α, β or γ-adducins, respectively) are increased by activators of PKC but not by activators of cAMP- or calcium-dependent kinases. Agonist-stimulated phosphorylation is not sensitive to nonPKC inhibitors such as wortmannin (Jaken, unpublished results). Thus, adducin phosphorylation can be used in conjunction with other methods as a surrogate to monitor endogenous PKC activity in cells and tissues. Using this criteria, it appears that PKC activity is up-regulated in renal carcinomas.27,28 Levels of phosphorylated adducin are also increased in attaching fibroblasts, where the phospho-adducin colocalizes with PKC-α (Jaken, manuscript in preparation). These studies demonstrate the utility of using substrate phosphorylation to monitor PKC activity in time and space.
Gravin/STICK72/SSECKs Gravin was originally identified as an antigen recognized by serum from myesthenia gravis patients. Gravin and its homologue (known as STICK72 or SSECKs) were subsequently cloned in screens for Protein Kinase A (PKA) and PKC binding proteins.29-31 These proteins are homologous in their PKA binding domains, which are located in the C-terminus. They are also homologous in their 2 PKC binding domains which are located in the middle of these very large (>250 kDa) proteins. Each of the PKC binding domains contains 2 PKC phosphorylation sites. Phosphorylation at these sites in these STICKs has been useful for monitoring endogenous PKC activity. Phosphorylated STICK72 is concentrated in leading lamellipodia of migrating cells, indicating that PKC is also active in these areas and functions in regulation of cell motility.31 Clone72 /SSECKs is downregulated in transformed fibroblasts and there is some evidence that reintroducing expression attenuates the metastatic potential of prostate epithelial cells.32 In addition to binding PKA and PKC, gravin also interacts with β-adrenergic receptors, Protein Phosphatase 2B (PP2B)33 and G-protein-Regulated Kinase-2.34 Gravin provides a scaffold to integrate the activities of these kinases and phosphatases with the receptor to provide dynamic regulation of the β-adrenergic receptor signaling, desensitization and resensitization. Gravin binds to the β-adrenergic receptor cytoplasmic tail in the absence of agonist, and remains associated during receptor sequestration to clathrin pits.35 The implication is that the gravin-organized signaling complex remains intact during desensitization and resensitization, and that gravin’s ability to traffic from plasma membrane to vesicles is critical for modulating β-adrenergic receptor signaling dynamics. A-Kinase-Anchoring-Protein (AKAP)79, another scaffold protein that binds both PKC and PKA, also interacts with the β-adrenergic receptor and PP2B, and may function similarly to gravin in other cells.36 The exact role of PKC in these processes is not yet known; however, both PKC and PP2B are required for recycling of receptors back to the plasma membrane after sequestration. Thus, similar to other PKC substrates, PKC phosphorylation may regulate gravin-membrane interactions. Dynamic, coordinated regulation by phosphorylation/dephosphorylation may be an essential component of β-adrenergic receptor trafficking.
Ezrin Ezrin is a member of the Ezrin, Radixin and Moesin (ERM) family of F-actin binding proteins that regulate cytoskeletal dynamic functions.37,38 The actin binding activity of ERM proteins depends on release of intramolecular constraints which leads to unfolding and transition of inactive multimers to active monomers. Different activation signals, including C-terminal threonine phosphorylation (T567 in ezrin), regulate the affinity of ERM proteins for actin.
Protein Kinase C Regulation by Protein Interactions
55
Consequently, stimuli that influence ERM protein conformation and actin binding are also key regulators of cellular activities that involve cytoskeletal remodeling, e.g., lamellipodia formation and motility. Phosphorylation of ERM proteins was originally thought to be primarily regulated by the Rho-activated kinase (ROCK); however, subsequent experiments which showed that phosphorylation of ERM is insensitive to ROCK inhibition casts doubt on the role of ROCK.39 PKC-α, PKC-δ and PKC-θ activities have each been associated with enhanced cell motility.40-43 In particular, PKC-α overexpression increased motility of breast cancer cells in a process that was in part regulated by Rho-family GTPases.44 Furthermore, PKC-α promoted motility in cell systems in which migration was selectively dependent on β1 integrin ligation.45 These studies, which implicated PKC-α as a primary regulator of integrin intracellular trafficking and motility, prompted further studies on the role of this PKC in ERM phosphorylation. PKC-α overexpression increased the rate of cell movement as well as T567 phosphorylation of ezrin in fibrosarcoma cells.39 Activated PKC-α cosedimented with β1 integrins and ezrin in a semi-purified plasma membrane fraction. Fluorescence lifetime imaging microscopy also demonstrated colocalization of PKC-α and ezrin, and this was further verified in coimmunoprecipitation studies. Both ezrin and PKC-α were localized to plasma membrane protrusions. Phosphorylation of ezrin was required for the PKC-α effect on increasing cell motility, since it was shown that expression of an ezrin phosphorylation site mutant abrogated the PKC-α stimulatory effect. These data clearly establish a dynamic association between PKC-α and ezrin and indicate that the association leads to a productive phosphorylation event that regulates cytoskeletal remodeling events that are determinants of the rate of cell motility. Phosphatidyl Inositol-3-kinase (PI3-kinase) is also well known to be an important regulator of cell motility. Several lines of evidence indicate that the PKC and PI3-kinase pathways are functionally integrated to provide fine control of cytoskeletal remodeling and motility. In addition to phosphorylation, the actin binding activity of ERM proteins is also regulated by polyphosphoinositides. Lipid binding appears to enhance actin binding of ERM proteins when they are C-terminally phosphorylated.46 Calphostin C, which interferes with DAG binding and activation of PKC, blocked PIP3-stimulated motility in NIH3T3 cells.47 These studies demonstrate potential opportunities for cross talk between PKC, PI3-kinase and ERM protein phosphorylation to modulate cell motility. Other isoforms of PKC have also been implicated in integrin-dependent signaling events. β1 integrin regulates spreading of cells plated on collagen or fibronectin. β1 integrin-dependent signaling events that regulate cell spreading can be analyzed in experiments in which a β1 fragment that acts as a dominant negative inhibitor of integrin function is used to inhibit cell spreading. Overexpression of an activated (myristoylated) form of PKC-ε restored cell spreading.48 T cells from PKC-β-deficient mice are relatively nonmotile, and this phenotype can be rescued through overexpression of PKC-β.49 In these assays, locomotory behavior was largely dependent on LFA-1 integrins. PKCs have also been reported to directly bind to transmembrane-4 superfamily proteins (TM4SF) which may then facilitate their selective binding to β1 integrin complexes.50
Protein Interacting with C Kinase (PICK)1 PICK1 was identified in a yeast two-hybrid screen for PKC-α binding partners. Subsequent mapping studies demonstrated that PKC-α contains a PDZ binding domain in its C-terminus that directly interacts with the single PDZ domain contained within PICK1. PICK1 also associates with the AMPA-type glutamate receptor subunit 2,51 the metabotropic glutamte type 7 receptor,52 the dopamine transporter53 and the receptor tyrosine kinase ERBB2/HER2.54 In each case, PICK1 interactions with these receptors were mediated through its single PDZ domain (see also Chapter 10). In addition to the PDZ domain, PICK1 contains a coiled-coil domain and an acidic domain. These coiled-coil domains mediate homotropic interactions of PICK1 that promote
56
Protein Kinase C
oligomerization.55 In contrast, the acidic domain interferes with oligomerization. The ability of PICK1 to oligomerize was directly linked to clustering of receptors bound to PICK1.55 These mapping studies suggest a model in which PICK1 is targeted to synaptic membranes through its PDZ domain-mediated association with receptors. These interactions do not interfere with the ability of PICK1 to self-associate through its coiled-coil domain. Thus, PICK1 can regulate synaptic aggregation and receptor coclustering through these two protein interaction motifs. These studies suggest an important role for PICK1 in synaptic development and plasticity, as well as the function of monoamine transporters. PICK1 interactions with ERBB2/ HER2, which itself has multiple binding partners, have also been noted.54 These results suggest that oligomerization of PICK1 may also participate in regulating the functions and specificity of this important growth regulatory signaling pathway. Several studies have addressed the role of PKC in regulating PICK1 activities. In some cells, the interaction between PKC and PICK1 is promoted by phorbol esters, indicating that activated PKC is the preferred binding partner.51 In this model, activation is needed to recruit PKC to the PICK1-receptor complex. Once recruited, there is evidence that PKC can directly phosphorylate the GluR2 receptor. Receptor phosphorylation correlated with disruption of GluR2 interactions with its synaptic anchoring proteins and GluR2 release from the plasma membrane. Thus, in this system, PICK1 acts as a scaffold to facilitate PKC substrate phosphorylation and downregulation of GluR2 interactions with its targeting proteins. Whole cell recordings also suggest a role for PKC-mediated phosphorylation events in regulation of GluR2 association with synaptic membranes.56 PKC-α also phosphorylates the GluR7 receptor in vitro, and PICK1 inhibits this activity. These in vitro studies demonstrate that PICK1 modulates PKC phosphorylation of this family of receptors, although the role of phosphorylation has not yet been described. The formation of complexes containing oligomerized PICK1, PKC and glutamate receptors that are also PKC substrates suggests that this PICK1 plays an essential role in targeting PKC to its appropriate substrates within cells.
Bruton’s Tyrosine Kinase (Btk)1 Interactions between PKCs and several Pleckstrin Homology (PH) domain containing proteins have been noted. Among these, interactions between PKC-β and Btk are among the most well characterized to date. Initial studies demonstrated that PKC bound to Btk PH domains with high affinity (Kd = 39 nM).57 Moreover, PKC decreased Btk catalytic activity, indicating that this interaction could have functional relevance. In cells, Btk activity is primarily regulated by membrane association, which is required for tyrosine phosphorylation and catalytic competence. In resting cells, Btk is largely recovered in the cytosol. B-cell receptor ligation, which activates PI3-kinase to generate PIP 3, recruits Btk to the membrane through PH domain-dependent interactions. Mutations in the Btk PH domain that decrease PIP3 binding (R28C) impair plasma membrane recruitment, whereas mutations that enhance PIP3 binding (E41K) enhance membrane translocation. Curiously, mice carrying either the inactivation or the activating mutations are immunodeficient.58 These results suggest a critical role for dynamic Btk-membrane association/dissociation in fine tuning the amplitude, threshold and/or duration of B-cell receptor signaling events. PKC-β plays an important role in regulating Btk membrane association. PKC-β phosphorylates Btk in its Tec homology domain at a selective site (ser 180).58 Whereas wt Btk translocates from soluble to membrane fractions with B-cell receptor ligation, the Btk S180A mutant is preferentially recovered in membrane fractions. Btk S180A mutants, which were not phosphorylated by PKC-β, function as hyperactive alleles that enhance agonist-stimulated increases in PLC-γ activity, cytosolic calcium levels and JNK activation.58 These results are consistent with other evidence that Btk membrane association is a determinant of B-cell receptor signals, and that signaling is attenuated by processes that decrease Btk-PH domain membrane association. Hydrolysis of PIP3 as well as PKC-β-dependent Btk phosphorylation decrease PH domain affinity for membrane lipids and contribute to redistribution of activated Btk back to the
Protein Kinase C Regulation by Protein Interactions
57
cytosol. Taken together, these results demonstrate that PKC-β acts as a feedback inhibitor of B-cell receptor signaling by phosphorylating Btk and attenuating Btk PH-domain membrane interactions independent of PIP3. Studies with various activated and inactivated Btk alleles in mice indicate that exaggerated and deficient Btk signals are both detrimental in B-cell lineage development. These studies establish that PKC-β plays an important role, along with various other regulators, as a rheostat to fine tune B-cell functions.
PDK-1 PKCs are themselves phosphorylated on at least 3 residues, and these phosphorylation events regulate catalytic activity, subcellular location and stability59,60(see Chapter 6). The first phosphorylation event is by PDK-1 at a critical threonine in the “activation loop” and is required for catalytic activity. PKCs can then autophosphorylate at two C-terminal sites, which also regulate PKC activity and functions. Phosphorylation by PDK-1 depends on direct interactions between PDK-1 and sequences in the C-terminal tails of the PKCs containing the “hydrophobic” autophosphorylation sites.59,60 Direct binding studies with synthetic peptides demonstrate that both the phosphorylated and the unphosphorylated forms of the C-terminal peptide bind PDK-1. In contrast, kinase-dead mutants (ie, unphosphorylated PKCs) preferentially coimmunoprecipitate with PDK-1. This apparent discrepancy can be explained by the observation that the unphosphorylated C-terminal tails are more exposed and, consequently, more readily available to bind PDK-1. Taken together, these studies suggest that one mechanism by which PKC kinase dead mutants interfere with PKC signaling is through binding and sequestering PDK-1. This interferes with PDK-1 phosphorylation of activation loops of wild type, endogenous PKCs and the subsequent autophosphorylation events needed for production of the fully mature enzymes. This model helps explain why the effects of overexpressed kinase dead mutants are not limited to the cognate PKC isotypes. For example, overexpressed PKC-ζ kinase dead mutant inhibited endogenous PKC-ζ as well as PKC-α and PKC-ε.61 Inhibition by the dominant negative constructs was due to decreased phosphorylation of wild type PKCs. Since PDK-1 phosphorylates critical residues in the activation loop of several kinases besides PKC (i.e., Protein Kinase B and PKA), overexpression of the PDK-1 sequestering PKC mutants may have unforeseen consequences (for example, see62). In some cases, high expression levels of mutant, kinase dead PKCs could potentially interfere with both PKC-dependent and -independent signaling events. Thus, while overexpression of kinase dead mutants has proved useful, some cautionary notes are needed in view of our increased awareness of PKC interactions with other proteins.
Par-3/ASIP, Par-6 and cdc42/Rac1 A role for PKC-ζ in formation of epithelial cell junctional complexes was originally suggested by studies of atypical PKC (aPKC) from C. elegans. These studies established that aPKC plays a critical role in establishing cell polarity by interacting with PAR-3.63 Subsequent studies identified a mammalian PAR-3 homologue referred to as a typical PKC-specific interacting protein (ASIP).63 ASIP colocalizes with PKC-ζ at tight junctions of polarized epithelial cells, indicating conservation of aPKC function in junctional complex formation throughout evolution. Atypical PKCs ζ and λ also interact with a mammalian homologue of PAR-6 resulting in a ternary complex of aPKC-ASIP/PAR-3-PAR-6. Overexpression of mutant, kinase-dead PKC-λ disrupted junctional complexes in MDCK cells,64 indicating that aPKC catalytic activity was required for complex assembly and/or stability. ASIP/PAR-3 and PAR-6 also directly interact with the junctional adhesion molecule (JAM) in tight junctions through its first PDZ domain.65 These studies indicate that interactions with JAM may play an additional role in recruiting the ASIP/PAR-3 - aPKC complex to tight junctions. Recently, several laboratories established that ASIP/PAR-3 and PAR-6 directly interact with the small GTPases cdc42 and Rac1.64,66-69 Association of PKC-ζ with PAR-6 and GTP-cdc42 stimulates PKC-ζ kinase activity. Furthermore, PAR-6 overexpression potentiated disruption
58
Protein Kinase C
of tight junctions and increased cell transformation caused by cdc42/Rac1. This PAR-6 function was PKC-ζ-dependent. In other studies, expression of activated cdc42 mutants caused a loss of stress fibers, and this effect was blocked by coexpression of a kinase-dead, dominant negative PKC-ζ mutant. Taken together, these studies indicate that aPKCs are required for cdc42-dependent actin cytoskeletal remodeling. It is also intriguing to consider that the cooperative effects of PKC-ζ and PAR-3/PAR-6 on promoting growth and transformation are mediated through their disruption of contacts. PKC-ζ phosphorylates PAR-3 but not PAR-6 in vitro; however, the role of phosphorylation in regulating PAR-3 interactions with PKC-ζ has not yet been described. Nonetheless, the accumulated evidence indicates that the formation of Cdc42/PAR-3/PAR-6/PKC-ζ complexes is dynamic. The possibility that PKC-ζ is in part responsible for reversible complex formation and subcellular localization is intriguing.
Caenorhabditis elegans C Kinase Adapter (CKA) 1 and 1S In a separate yeast 2-hybrid screen, another pair of related C. elegans aPKC binding proteins were identified (CKA1 and the splice variant CKA1S).70 Both proteins contain polybasic N-terminal domains similar to the classical phosphorylation site domain (PSD) in MARCKS and MARCKS-related protein. This domain is thought to mediate phospholipid binding and membrane association in these proteins (see below). Both CKA proteins also contain phosphotyrosine binding (PTB) domains that directly interact with aPKC. The aPKC binding site was localized to the V2 “linker” region which connects the N-terminal regulatory and C-terminal catalytic domains of PKCs. The coordinated regulation of the activities of the PSD membrane targeting domain and the aPKC binding PTB domain influence the localization and activity of C. elegans aPKC. Subsequent studies investigated the mechanisms that govern formation and localization of the aPKC-CKA1 complex.71 Activation of DAG-stimulated PKCs promoted phosphorylation of CKA1 within the PSDs. In turn, phosphorylation promoted redistribution of CKA1 (along with the tethered aPKC) from the cell surface to the cytoplasm. Effects of phosphatase inhibitors demonstrated that the net phosphorylation state of CKA1 is dynamically regulated by endogenous phosphatase activities. These results suggest a novel mechanism whereby DAG-stimulated PKCs regulate the localization of aPKC through phosphorylation of PSDs in the aPKC binding protein CKA1.
Zeta-Interacting Protein (ZIP)/p62 Other proteins isolated in yeast 2-hybrid screens for PKC-ζ interacting proteins (ZIPs) include ZIP1 (also known as A170, EBIAP and p62). Initially these proteins were recognized for their abilities to bind tyrosine kinases, to form oligomers, to be induced by oxidative stress and to influence intracellular trafficking.72,73 More recently, ZIP1 and a splice variant named ZIP2 were also identified in a yeast 2-hybrid screen as potassium channel (Kvb2)-interacting proteins.74 The association of PKC-ζ with an ion channel is consistent with physiological evidence for signaling complexes containing ion channels together with their regulatory kinases and phosphatases. Adaptor proteins such as ZIP1 and ZIP2 are thought to play critical roles in establishing associations with specific kinases and phosphatases that regulate the phosphorylation and activity of the ion channel protein. In situ hybridization and coimmunoprecipitation experiments confirmed the close association of PKC-ζ, Kvb2 and ZIP1 in pyramidal and Purkinje cell populations. Immune complex kinase assays demonstrated that PKC-ζ phosphorylates Kvb2 and that ZIP1 potentiates this activity. ZIP1 did not potentiate PKC-ζ phosphorylation of myelin basic protein, which does not bind to ZIP1. These results demonstrate that substrate binding to ZIP1 was required for ZIP1 enhancement of phosphorylation and imply functional significance for the adaptor/binding protein activity. ZIP1 and ZIP2 appear to differ in their potencies to enhance PKC-ζ kinase activity, suggesting that differential expression of these splice variants may fine tune PKC-z phosphorylation of physiological substrates.
Protein Kinase C Regulation by Protein Interactions
59
ZIP1/p62 has also been shown to mediate selective PKC-ζ interactions with components of the Tumor Necrosis Factor (TNF)-α signaling pathway. Although several studies have presented evidence that PKC-ζ is required for TNF-α signaling events leading to Inhibitor of κB kinase (IKK) activation and Nuclear Factor-κB (NF-kB)-dependent changes in gene expression,75,76 the molecular mechanisms defining the role of PKC-ζ have not been well understood. Recently, p62/ZIP1 was shown to directly interact with Receptor Interacting Protein and TNF Receptor- associated Factor 6, which are critical components of the TNF-α signaling process.77 Thus, p62/ZIP1 is thought to act as an adapter that links PKC-ζ to the TNF-α signaling complex. Similarly, the selective involvement of PKC-ζ in the TNF-α signaling pathway can be explained by the selective interactions of PKC-ζ with its binding partner p62/ ZIP1. The critical role of PKC-ζ in TNF signaling has been substantiated in studies with PKC-ζ null mice.78 These animals develop normally; however, upon close inspection, defects in B cell maturation were noted. These defects, which are similar to those seen in TNF-Receptor-1 knock-out mice, correlated with decreased TNF-α activation of IKK and NFκB. In agreement with these results, PKC-ζ null cells were more susceptible to TNF-α-mediated apoptosis ex vivo. Defective TNF-α signaling was most prominent in lung tissue which expresses high levels of PKC-ζ but only low levels of PKC-λ. Somewhat surprisingly, although TNF-mediated NFκB activation was markedly attenuated in fibroblast cell lines from the knock-out mice, this was not due to decreased IKK activation, but was associated with decreased p65 phosphorylation. Thus, PKC-ζ may phosphorylate and regulate the activity of more than one component of the TNF signaling complex. Although the role of p62 was not directly evaluated in these studies, the results do substantiate the importance of PKC-ζ in TNF signaling (for further detail see Chapter 9 in this volume).
Discussion PKC interacts with several types of binding proteins that influence subcellular localization, activation, substrate juxtaposition and activation-dependent relocation of PKC within cells. It appears that a wide range of binding partners is needed to establish the functional diversity and isoform selectivity of PKC actions. Identifying mechanisms that regulate and coordinate PKC interactions with its binding partners will ultimately be needed to fully understand PKC signaling. It is likely that disregulation of these binding events is the molecular basis for PKC involvement in pathological conditions such as tumorigenesis. The overall theme emerging from these studies of PKC-protein interactions is that binding properly positions PKC in relation to upstream activators and downstream substrates, and furthermore, that these interactions are dynamically regulated. There are now several examples demonstrating that PKC plays a significant role in desensitization/resensitization of signaling complexes or as a determinant of membrane/cytosol compartmentalization of substrate proteins. Cycles of phosphorylation by PKC and dephosphorylation by phosphatases that are also included in these protein complexes modulate the amplitude, duration and/or rate of signaling events. In these models, PKCs act primarily as rheostats rather than switches to regulate signaling events and cellular responses. As such, PKCs are attractive targets for development of small molecule inhibitors with minimal toxicity and broad therapeutic application.
References 1. Nishizuka Y. Protein kinases 5: Protein kinase C and lipid signaling for sustained cellular responses. FASEB J 1995; 9:484-496. 2. Cho W. Membrane targeting by C1 and C2 domains. J Biol Chem 2001. 3. Newton AC, Johnson. Protein kinase C: A paradigm for regulation of protein function by two membrane-targeting modules. Biochim Biophys Acta 1998; 1376:155-172. 4. Nishikawa K, Toker, Johannes et al. Determination of the specific substrate sequence motifs of protein kinase C isozymes. J Biol Chem 1997; 272:952-960. 5. Chapline C, Ramsay, Klauck et al. Interaction cloning of protein kinase C substrates. J Biol Chem 1993; 268:6858-6861.
60
Protein Kinase C
6. Sharrocks AD, Yang Galanis. Docking domains and substrate-specificity determination for MAP kinases. Trends Biochem Sci 2000; 25:448-453. 7. Pellicena P, Miller. Processive phosphorylation of p130Cas by Src depends on SH3-polyproline interactions. J Biol Chem 2001; 276:28190-28196. 8. Gao T, Toker Newton. The carboxyl terminus of protein kinase c provides a switch to regulate its interaction with the phosphoinositide- dependent kinase, PDK-1. J Biol Chem 2001; 276:19588-19596. 9. Mochly-Rosen D, Gordon. Anchoring proteins for protein kinase C: A means for isozyme selectivity. FASEB J 1998; 12:35-42. 10. Csukai M, Chen, De Matteis et al. The coatomer protein b’-COP, a selective binding protein (RACK) for protein kinase Ce. J Biol Chem 1997; 272:29200-29206. 11. Ron D, Chen, Caldwell et al. Cloning of an intracellular receptor for protein kinase C: A homolog of the beta subunit of G proteins. Proc Natl Acad Sci USA 1994; 91:839-843. 12. Stebbins EG, Mochly-Rosen. Binding specificity for RACK1 resides in the V5 region of beta II protein kinase C. J Biol Chem 2001; 276:29644-29650. 13. Tsunoda S, Sun, Suzuki et al. Independent anchoring and assembly mechanisms of INAD signaling complexes in Drosophila photoreceptors. J Neurosci 2001; 21:150-158. 14. Kumar R, Shieh BH. The second PDZ domain of INAD is a type I domain involved in binding to eye protein kinase C. Mutational analysis and naturally occurring variants. J Biol Chem 2001; 276:24971-24977. 15. Seykora JT, Myat, Allen et al. Molecular determinants of the myristoyl-electrostatic switch of MARCKS. J Biol Chem 1996; 271:18797-18802. 16. Hyatt SL, Liao, Aderem et al. Correlation between protein kinase C binding proteins and substrates in REF52 cells. Cell Growth Diff 1994; 5:495-502. 17. Yarmola EG, Edison, Lenox et al. Actin filament cross-linking by MARCKS: Characterization of two actin -binding sites within the phosphorylation site domain. J Biol Chem 2001; 276:22351-22358. 18. Wohnsland F, Schmitz, Steinmetz et al. Influence of the effector peptide of MARCKS-related protein on actin polymerization: A kinetic analysis. Biophys Chem 2000; 85:169-177. 19. Spizz G, Blackshear. Overexpression of the myristoylated alanine-rich c-kinase substrate inhibits cell adhesion to extracellular matrix components. J Biol Chem 2001. 20. Wang J, Arbuzova, Hangyas-Mihalyne et al. The effector domain of myristoylated alanine-rich C kinase substrate binds strongly to phosphatidylinositol 4,5-bisphosphate. J Biol Chem 2001; 276:5012-5019. 21. Li Y, Martin, Spizz et al. MARCKS protein is a key molecule regulating mucin secretion by human airway epithelial cells in vitro. J Biol Chem 2001; 276:40982-40990. 22. Zhou X, Li. Macrophage-enriched myristoylated alanine-rich C kinase substrate and its phosphorylation is required for the phorbol ester-stimulated diffusion of beta 2 integrin molecules. J Biol Chem 2000; 275:20217-20222. 23. Yue L, Lu, Garces et al. Protein kinase C-regulated dynamitin-macrophage-enriched myristoylated alanine-rice C kinase substrate interaction is involved in macrophage cell spreading. J Biol Chem 2000; 275:23948-23956. 24. Jin T, Yue Li. In vivo interaction between dynamitin and MacMARCKS detected by the fluorescent resonance energy transfer method. J Biol Chem 2001; 276:12879-12884. 25. Kimura K, Fukata, Matsuoka et al. Regulation of the association of adducin with actin filaments by rho-associated kinase (Rho-kinase) and myosin phosphatase. J Biol Chem 1998; 273:5542-5548. 26. Fukata Y, Oshiro, Kinoshita et al. Phosphorylation of adducin by rho-kinase plays a crucial role in cell motility [In Process Citation]. J Cell Biol 1999; 145:347-361. 27. Fowler L, Everitt, Stevens et al. Localization, expression levels and phosphorylation state of alphaand gamma-adducins in progressive herediatary renal cell carcinoma. Cell Growth & Diff 1998; 9:405-413. 28. Fowler L, Dong, Van de Water et al. Transformation-sensitive changes in expression, localization and phosphorylation of adducins in renal proximal tubule epithelial cells. Cell Growth & Diff 1998; 9:177-184. 29. Nauert JB, Klauck, Langeberg et al. Gravin, an autoantigen recognized by serum from myasthenia gravis patients, is a kinase scaffold protein. Current Biology 1997; 7:52-62. 30. Chapline C, Mousseau, Ramsay et al. Identification of a major protein kinase C-binding protein and substrate in rat embryo fibroblasts—Decreased expression in transformed cells. J Biol Chem 1996; 271:6417-6422. 31. Chapline C, Cottom, Tobin et al. A major, transformation-sensitive PKC binding protein is also a PKC substrate involved in cytoskeletal remodeling. J Biol Chem 1998; 273:19482-19489.
Protein Kinase C Regulation by Protein Interactions
61
32. Xia W, Unger, Miller et al. The Src-suppressed C kinase substrate, SSeCKS, is a potential metastasis inhibitor in prostate cancer. Cancer Res 2001; 61:5644-5651. 33. Shih M, Lin, Scott et al. Dynamic complexes of beta2-adrenergic receptors with protein kinases and phosphatases and the role of gravin. J Biol Chem 1999; 274:1588-1595. 34. Lin F, Wang Malbon. Gravin-mediated formation of signaling complexes in beta 2-adrenergic receptor desensitization and resensitization. J Biol Chem 2000; 275:19025-19034. 35. Fan G, Shumay, Wang et al. The scaffold protein gravin (cAMP-dependent protein kinase-anchoring protein 250) binds the beta 2-adrenergic receptor via the receptor cytoplasmic Arg-329 to Leu-413 domain and provides a mobile scaffold during desensitization. J Biol Chem 2001; 276:24005-24014. 36. Fraser ID, Cong, Kim et al. Assembly of an A kinase-anchoring protein-beta(2)-adrenergic receptor complex facilitates receptor phosphorylation and signaling. Curr Biol 2000; 10:409-412. 37. Niggli V. Structural properties of lipid-binding sites in cytoskeletal proteins. Trends Biochem Sci 2001; 26:604-611. 38. Mangeat P, Roy, Martin. ERM proteins in cell adhesion and membrane dynamics. Trends Cell Biol 1999; 9:187-192. 39. Ng T, Parsons, Hughes et al. Ezrin is a downstream effector of trafficking PKC-integrin complexes involved in the control of cell motility. EMBO J 2001; 20:2723-2741. 40. Platet N, Prevostel, Derocq et al. Breast cancer cell invasiveness: Correlation with protein kinase C activity and differential regulation by phorbol ester in estrogen receptor-positive and -negative cells. Int J Cancer 1998; 75:750-756. 41. Kiley SC, Clark, Duddy et al. Increased protein kinase Cd in mammary tumor cells: Relationship to transformation and metastatic progression. Oncogene 1999; 18:6748-6757. 42. Kiley SC, Clark, Goodenough et al. Protein kinase C d involvement in mammary tumor cell metastasis. Cancer Res 1999; 59:3230-3238. 43. Tang SQ, Morgan, Parker et al. Requirement for protein kinase C q for cell cycle progression and formation of actin stress fibers and filopodia in vascular endothelial cells. J Biol Chem 1997; 272:28704-28711. 44. Sun XG, Rotenberg. Overexpression of protein kinase Calpha in MCF-10A human breast cells engenders dramatic alterations in morphology, proliferation, and motility. Cell Growth Differ 1999; 10:343-352. 45. Ng T, Shima, Squire et al. PKCalpha regulates beta1 integrin-dependent cell motility through association and control of integrin traffic. EMBO J 1999; 18:3909-3923. 46. Nakamura F, Huang, Pestonjamasp et al. Regulation of F-actin binding to platelet moesin in vitro by both phosphorylation of threonine 558 and polyphosphatidylinositides. Mol Biol Cell 1999; 10:2669-2685. 47. Derman MP, Toker, Hartwig et al. The lipid products of phosphoinositide 3-kinase increase cell motility through protein kinase C. J Biol Chem 1997; 272:6465-6470. 48. Berrier AL, Mastrangelo, Downward et al. Activated R-ras, Rac1, PI 3-kinase and PKCepsilon can each restore cell spreading inhibited by isolated integrin beta1 cytoplasmic domains. J Cell Biol 2000; 151:1549-1560. 49. Volkov Y, Long, McGrath et al. Crucial importance of PKC-beta(I) in LFA-1-mediated locomotion of activated T cells. Nat Immunol 2001; 2:508-514. 50. Zhang XA, Bontrager Hemler. Transmembrane-4 Superfamily Proteins Associate with Activated Protein Kinase C (PKC) and Link PKC to Specific beta 1 Integrins. J Biol Chem 2001; 276:25005-25013. 51. Perez JL, Khatri, Chang et al. PICK1 targets activated protein kinase Calpha to AMPA receptor clusters in spines of hippocampal neurons and reduces surface levels of the AMPA-type glutamate receptor subunit 2. J Neurosci 2001; 21:5417-5428. 52. Dev KK, Nakanishi Henley. Regulation of mglu(7) receptors by proteins that interact with the intracellular C-terminus. Trends Pharmacol Sci 2001; 22:355-361. 53. Torres GE, Yao Mohn et al. Functional interaction between monoamine plasma membrane transporters and the synaptic PDZ domain-containing protein PICK1. Neuron 2001; 30:121-134. 54. Jaulin-Bastard F, Saito, Le Bivic et al. The ERBB2/HER2 receptor differentially interacts with ERBIN and PICK1 PSD-95/DLG/ZO-1 domain proteins. J Biol Chem 2001; 276:15256-15263. 55. Boudin H, Craig. Molecular determinants for PICK1 synaptic aggregation and mGluR7a receptor coclustering: Role of the PDZ, coiled-coil, and acidic domains. J Biol Chem 2001; 276:30270-30276. 56. Daw MI, Chittajallu, Bortolotto et al. PDZ proteins interacting with C-terminal GluR2/3 are involved in a PKC- dependent regulation of AMPA receptors at hippocampal synapses. Neuron 2000; 28:873-886.
62
Protein Kinase C
57. Yao LB, Suzuki, Ozawa et al. Interactions between protein kinase C and pleckstrin homology domains - Inhibition by phosphatidylinositol 4,5-bisphosphate and phorbol 12-myristate 13-acetate. J Biol Chem 1997; 272:13033-13039. 58. Kang SW, Wahl, Chu et al. PKCbeta modulates antigen receptor signaling via regulation of Btk membrane localization. EMBO J 2001; 20:5692-5702. 59. Parekh DB, Ziegler Parker. Multiple pathways control protein kinase C phosphorylation. EMBO J 2000; 19:496-503. 60. Newton AC. Protein kinase C: Structural and spatial regulation by phosphorylation, cofactors, and macromolecular interactions. Chem Rev 2001; 101:2353-2364. 61. Garcia-Paramio P, Cabrerizo, Bornancin et al. The broad specificity of dominant inhibitory protein kinase C mutants infers a common step in phosphorylation. Biochem J 1998; 333:631-636. 62. Matsumoto M, Ogawa, Hino et al. Inhibition of insulin-induced activation of Akt by a kinase-deficient mutant of the epsilon isozyme of protein kinase C. J Biol Chem 2001; 276:14400-14406. 63. Ohno S. Intercellular junctions and cellular polarity: The PAR-aPKC complex, a conserved core cassette playing fundamental roles in cell polarity. Curr Opin Cell Biol 2001; 13:641-648. 64. Suzuki A, Yamanaka, Hirose et al. Atypical protein kinase C is involved in the evolutionarily conserved par protein complex and plays a critical role in establishing epithelia- specific junctional structures. J Cell Biol 2001; 152:1183-1196. 65. Ebnet K, Suzuki, Horikoshi et al. The cell polarity protein ASIP/PAR-3 directly associates with junctional adhesion molecule (JAM). EMBO J 2001; 20:3738-3748. 66. Joberty G, Petersen, Gao et al. The cell-polarity protein Par6 links Par3 and atypical protein kinase C to Cdc42. Nat Cell Biol 2000; 2:531-539. 67. Lin D, Edwards, Fawcett et al. A mammalian PAR-3-PAR-6 complex implicated in Cdc42/Rac1 and aPKC signalling and cell polarity. Nat Cell Biol 2000; 2:540-547. 68. Qiu RG, Abo Steven. A human homolog of the C. elegans polarity determinant Par-6 links Rac and Cdc42 to PKCzeta signaling and cell transformation. Curr Biol 2000; 10:697-707. 69. Yamanaka T, Horikoshi, Suzuki et al. PAR-6 regulates aPKC activity in a novel way and mediates cell-cell contact-induced formation of the epithelial junctional complex. Genes Cells 2001; 6:721-731. 70. Zhang L, Wu Rubin. A novel adapter protein employs a phosphotyrosine binding domain and exceptionally basic N-terminal domains to capture and localize an atypical protein kinase C: Characterization of Caenorhabditis elegans C kinase adapter 1, a protein that avidly binds protein kinase C3. J Biol Chem 2001; 276:10463-10475. 71. Zhang L, Wu Rubin. Structural properties and mechanisms that govern association of C kinase adapter 1 with protein kinase C3 and the cell periphery. J Biol Chem 2001; 276:10476-10484. 72. Puls A, Schmidt, Grawe et al. Interaction of protein kinase C zeta with ZIP, a novel protein kinase c-binding protein. Proc Natl Acad Sci USA 1997; 94:6191-6196. 73. Sanchez P, De Carcer, Sandoval et al. Localization of atypical protein kinase C isoforms into lysosome- targeted endosomes through interaction with p62. Mol Cell Biol 1998; 18:3069-3080. 74. Gong J, Xu, Bezanilla et al. Differential stimulation of PKC phosphorylation of potassium channels by ZIP1 and ZIP2. Science 1999; 285:1565-1569. 75. Moscat J, Diaz-Meco. The atypical protein kinase Cs. Functional specificity mediated by specific protein adapters. EMBO Rep 2000; 1:399-403. 76. Moscat J, Sanz Sanchez et al. Regulation and role of the atypical PKC isoforms in cell survival during tumor transformation. Adv Enzyme Regul 2001; 41:99-120. 77. Sanz L, Diaz-Meco, Nakano et al. The atypical PKC-interacting protein p62 channels NF-kappaB activation by the IL-1-TRAF6 pathway. EMBO J 2000; 19:1576-1586. 78. Leitges M, Sanz, Martin et al. Targeted disruption of the zetaPKC gene results in the impairment of the NF-kappaB pathway. Mol Cell 2001; 8:771-780. 79. Li HS, Montell C. TRP and the PDZ protein, INAD, form the core complex required for retention of the signalplex in Drosophila photoreceptor cells. J Cell Biol 2000; 150:1411-1422. 80. Bahner M, Sander P, Paulsen R et al. The visual G protein of fly photoreceptors interacts with the PDZ domain assembled INAD signaling complex via direct binding of activated Galpha(q) to phospholipase c-beta. J Biol Chem 2000; 275:2901-2904. 81. Liu M, Parker LL, Wadzinski BE et al. Reversible phosphorylation of the signal transduction complex in Drosophila photoreceptors. J Biol Chem 2000; 275:12194-12199.
CHAPTER 6
Regulation of Protein Kinase C by Phosphorylation Alex Toker
Preface
R
egulation of protein kinase C (PKC) by phosphorylation is central to its ability to transduce the numerous responses elicited by phospholipid hydrolysis. Thus, phosphorylation acts in concert with the two other key regulatory mechanisms, allosteric activation by lipid second messengers and interaction with targeting proteins, to fine-tune PKC function. All PKC family members are regulated by a series of phosphorylations on conserved serine or threonine residues in the catalytic kinase domain of the enzyme, and studies in the last few years have provided detailed mechanistic insight into regulation of PKC activity and function by phosphorylation. PKC is regulated both by an upstream kinase as well as by autophosphorylation. The finding that phosphoinositide-dependent kinase-1 (PDK-1), originally discovered as the upstream kinase for PKC’s close cousin, the Akt/Protein Kinase B serine/ threonine kinase, is also the upstream kinase for all PKCs suggested a common regulatory mechanism for all family members. However, what has also emerged recently is that important differences exist concerning the precise mechanism of phosphorylation of distinct PKC family members. The fact that PKCs are also phosphorylated on tyrosine residues under certain physiological conditions has added a new dimension to the complex tale of PKC regulation. This chapter will highlight the key findings and review the mechanisms which have been shown to regulate PKC phosphorylation.
Introduction PKCs are members of a larger kinase family typically referred to as the AGC kinase superfamily (so-called because of the prototypes protein kinase A (PKA), protein kinase G (PKG) and protein kinase C (PKC)). The carboxyl-terminal catalytic domain is approximately 45 kDa and comprises the C3 and C4 domains which encompass the ATP and substrate-binding cavities. Although the crystal structure of native PKC has so far remained elusive, the catalytic kinase core is most similar to that of PKA (approximately 40% identity at the amino acid level) only diverging significantly at the very carboxyl-terminal sequence. Indeed, modeling studies based on the PKA crystal structure have suggested that key residues retained in both PKA and PKC are those which maintain the fold of the kinase domain, whereas residues found on the surface of the protein are different.1,2 PKCs phosphorylate Ser and Thr residues and an oriented peptide library screening approach provided valuable information concerning the optimal substrate peptide sequences of nine distinct PKC isozymes3. This study confirmed that PKCs prefer to phosphorylate substrates with basic amino acids (Arg or Lys) both amino- and carboxyl-terminal to the phospho acceptor. However it is worth noting that the selectivity reported was somewhat modest and not as rigorous as that of other AGC kinases such as PKA or Akt/PKB. The substrate binding cavity is also an important determinant in the activation of Protein Kinase C, 2nd Edition, edited by Lodewijk Dekker. ©2004 Eurekah.com and Kluwer Academic / Plenum Publishers.
64
Protein Kinase C
PKC: release of the pseudosubstrate domain upon diacylglycerol binding induces the active conformation allowing catalysis to occur (see Chapter 4). The three PKC subfamilies, conventional (α, βI, βII, γ), novel (δ, ε, η/L, θ) and atypical (ζ, and ι/λ) PKCs have highly conserved catalytic kinase domains. Conversely, the kinase domain of PKC-µ, which was originally identified as a novel member of the PKC family, is more closely related to the Ca2+/calmodulin-dependent protein kinase family. The regulation of PKC-µ is also distinct from that of PKCs, and for these reasons this enzyme is now termed Protein Kinase D (PKD), and together with two additional members, PKD-2 and PKD-3/PKC-ν, comprises an entirely separate and distinct protein kinase family. Regulation of PKD phosphorylation and function will not be further discussed here but is covered in an excellent review by Van Lint and colleagues.4 Similarly, the homology of the catalytic kinase domain extends to another kinase family which comprises PKC-related kinase (PRK)-1, PRK-2 and PRK-4 (also known as Protein Kinase N). These kinases are distinct from PKCs in that they are regulated primarily by their interaction with the small GTPases of the Rho family.5 However, the activation of PRK is also controlled by phosphorylation, and in this regard closely resembles that of PKC.6 The first indication that PKCs are regulated by phosphorylation was when Fabbro and co-workers showed that long-term 12-phorbol 13-myristate acetate (PMA) stimulation of cells induced a fast migrating (dephosphorylated) form of PKC which was inactive.7 More direct evidence came from studies in which in vitro phosphatase treatment of purified PKC rendered the kinase inactive towards exogenous substrates.8 Again, comparison with the PKA crystal structure9 suggested a potential mechanism of PKC regulation by phosphorylation. All AGC kinases require phosphorylation of a segment near the entrance of the active site, the ‘activation loop’, and phosphorylation of a Ser or Thr residue in this loop fulfils two important functions: it correctly aligns residues for subsequent catalysis, and once phosphorylated, unmasks the entrance of the substrate binding cavity. The notion that such a mechanism also exists for PKC was reinforced by the finding that negative charge at activation loop is absolutely required for PKC activity.10,11 As discussed below, the discovery of PKC phosphorylation at the activation loop by the PDK-1 enzyme afforded exquisite mechanistic insight into PKC regulation. However, it was also appreciated that phosphorylation of PKC at two other conserved residues is necessary for proper PKC function. Thus, the concept emerged that PKC is processed by a series of ordered phosphorylations, the first and rate-limiting step being phosphorylation at the activation loop. The second phosphorylation occurs at a conserved residue in what has been termed the ‘turn motif ’, because by analogy to the PKA crystal structure, the sequence and phospho-acceptor residue lie at the apex of a turn. The third and final phosphorylation occurs at the very carboxyl-terminus in a sequence known as the ‘hydrophobic motif ’, so-called because the phospho-acceptor Ser/Thr is flanked by hydrophobic residues. It is also worth noting that additional phosphorylations within the regulatory and C2 domain of novel PKCs from the marine mollusk Aplysia have been described and shown to regulate membrane association, although it is not known if these phosphorylations are conserved in mammalian PKCs12. The activation loops, turn motifs and hydrophobic motifs of PKCs are not only highly conserved within the PKC family, but can also be found in other AGC kinases including PKA, Akt/PKB, p70 S6-kinase, p90 RSK and serum and glucocorticoid-induced kinase (SGK) (Fig. 1). The regulation of PKC activation loop, turn motif and hydrophobic site phosphorylation are discussed below.
Phosphorylation at the Activation Loop Phosphorylation of the PKC activation loop is the rate-limiting step in the processing of the enzyme. The first indication of this was mutagenesis studies which showed that mutation of the PKC-α (Thr497) and PKC-β II (Thr500) activation loop residues to a neutral, non-phosphorylatable residue results in accumulation of an inactive, dephosphorylated protein which accumulates in the detergent-insoluble fraction.10,11 This is not due to increased
Regulation of Protein Kinase C by Phosphorylation
A
65
B
Figure 1. Alignment of the activation loop, turn motif and hydrophobic sites of PKCs and Akt/PKB, S6K1 and PKA. The sequences of the activation loop (top) and turn motif and hydrophobic site (bottom) of conventional (PKC-α, PKC-βI, PKC-βII, PKC-γ), novel (PKC-δ, PKC-ε, PKC-η (L), PKC-θ) and atypical (PKC-ζ, PKC-ι/λ) isozymes are shown. The homology with the corresponding sequences in Akt1 (PKBα), S6K1 and PKA are also shown. All sequences are of human origin. The numbering of the first amino acid in each sequence is also given.
phosphatase sensitivity of the activation loop mutant, rather it is the lack of subsequent carboxyl-terminal phosphorylations which results in a fully dephosphorylated enzyme.13 It is also worth noting that in PKC-βII activation loop mutants, compensating phosphorylation at adjacent Thr residues occurs, such that only a triple Thr497/498/500Ala mutant results in a completely dephosphorylated protein. Replacement of the PKC-βII activation loop with a negatively charged Glu, an effective mimic of phosphorylation, results in a fully-functional, active kinase.11,14 Importantly, although phosphorylation of the PKC activation loop is a pre-requisite for the maturation and activation of the kinase, once phosphorylated, phosphate at this position becomes dispensable.15 Therefore, in the case of conventional PKCs, although phosphorylation of the activation loop is an obligatory first step in the regulation of the kinase, it does not per se directly contribute to increased protein kinase activity. Rather, it initiates subsequent autophosphorylations in the turn motif and hydrophobic site resulting in a fully mature enzyme which is now competent to bind diacylglycerol (Fig. 2A). Consequently, loss of activation loop phosphate does not lead to any appreciable loss of protein kinase activity as long as the carboxyl-terminal sites are phosphorylated (see below). Indeed, mass spectrometric analysis of purified bovine brain PKC has shown that only 50% of PKC molecules are phosphorylated at the activation loop.15 These observations also account for the fact that changes in PKC activity are not readily detectable in immune-complex protein kinase assays following stimulation of cells with agonists which stimulate phospholipid hydrolysis, and which are known to activate PKC. This is because at least in the case of conventional PKCs, phosphorylation does not directly increase catalytic activity, rather it is required for the subsequent activating step, diacylglycerol binding. This is presumably lost during immunoprecipitation allowing the kinase to fold back into the inactive conformation. Both novel and atypical PKCs also have an absolute requirement for activation loop phosphorylation, although the functional consequence of this step is somewhat different than for conventional PKCs. Interestingly, PKC-δ is the only example of a PKC which is functionally active (albeit less than one-tenth of wild-type protein) when heterologously expressed in
66
Protein Kinase C
Figure 2. Sequence of events leading to phosphorylation of conventional, novel and atypical PKCs. A. βII is phosphorylated by PDK-1 at the activation loop Thr500. This event occurs at Conventional PKC-β βII to the membrane the membrane, but does not require membrane binding of PDK-1. Tethering of PKC-β is likely mediated by other interactions (e.g., the pseudosubstrate). Phosphorylation of the activation loop triggers phosphorylation of the two carboxyl-terminal sites, the turn motif (Thr641) and the hydrophobic βII is now released into the cytosol as a fully mature enzyme. site (Ser660). Fully phosphorylated PKC-β βII to the membrane where Agonists which promote phospholipid hydrolysis mediate translocation of PKC-β it interacts with its allosteric activator diacylglycerol (DAG), leading to activation. B. Novel PKC-εε is phosphorylated at the membrane in a PI 3-K-dependent manner such that interaction of PDK-1 with PtdIns-3,4,5-P3 (PIP3) is required for activation loop (Thr566) phosphorylation. This promotes autophosphorylation at the turn motif (Thr710), and at the hydrophobic site (Ser729), although transphosphorylation of Ser729 by a heterologous upstream kinase is possible. Fully phosphorylated PKC-εε is now ζ competent to bind to diacylglycerol, leading to full activation. C. Phosphorylation of the atypical PKC-ζ activation loop (Thr410) is also dependent on PI 3-K and occurs by binding of PtdIns-3,4,5-P3 to PDK-1. ζ autophosphorylates at the turn motif Thr560. The carboxyl terminal hydrophobic Phosphorylated PKC-ζ site (Glu579) contains a negatively charged residue and this motif is not subject to phosphorylation. Other ζ activity in stimulated cells. inputs, such as PtdIns-3,4,5-P3 and ceramide may further regulate PKC-ζ
Regulation of Protein Kinase C by Phosphorylation
67
bacteria; this is due to either compensating phosphorylations or the presence of negatively-charged Glu residues proximal to the activation loop residue (Thr505).16,17 In the case of atypical PKC-ζ, mutation of the activation loop Thr410 results in an inactive protein, whereas a Glu at this position is an effective mimic of phosphorylation.18,19 Interestingly, in the case of both novel and atypical PKCs, there is evidence that phosphorylation at the activation loop does lead to a measurable increase in protein kinase activity as judged by immune-complex kinase assays. For example, stimulation of cells with growth factors (e.g., Platelet-derived Growth Factor (PDGF), Epidermal Growth Factor (EGF)) and G-protein-coupled receptor agonists (e.g., bombesin, Lysophosphatidic acid (LPA)) leads to increased PKC-ε,20 PKC-δ18 and PKC-ζ18,19 activity when measured in an in vitro kinase assay using exogenous substrates. This suggests that phosphorylation of these PKC serves as a direct ON/OFF switch to regulate enzyme function. In the case of novel PKCs, this presumably acts in synergy with diacylglycerol binding, as these isozymes retain the C1 domains which mediate this interaction (Fig. 2B). In the case of atypical PKCs, phosphorylation appears to be the primary, if not sole, mechanism of activation, as these family members are not competent to bind to diacylglycerol. It is worth noting that in certain specialized cases, phosphorylation of atypical PKCs may act in concert with other inputs, such as generation of ceramide which has been shown to induce PKC-ζ activation21 (Fig. 2C). Thus, all PKCs have an absolute requirement for activation loop phosphorylation, but the net consequence of this regulatory event differs between distinct isozymes. Although phosphorylation of PKC was known to occur as far back as the late 1980s, it was not until one decade later that the kinase responsible for mediating activation loop phosphorylation was described. The PDK-1 enzyme was purified and characterized for its ability to directly phosphorylate the activation loop Thr of the AGC kinase Akt/PKB,22 an important effector of the phosphoinositide 3-kinase (PI 3-K) pathway. Because the activation loop sequences of several AGC kinases are highly homologous to that of Akt/PKB (Fig. 1), it was perhaps not surprising to find that PDK-1 is also the activation loop kinase of the p70 S6-kinases S6K1 and S6K2,23 although interestingly this phosphorylation does not require PtdIns-3,4-P2 or PtdIns-3,4,5-P3, at least in vitro.24 Immediately following this discovery, three independent laboratories almost simultaneously reported that PDK-1 is also the long-sought after PKC upstream kinase, for both conventional (PKC-βII)25, novel (PKC-δ and PKC-ε)18 and atypical (PKC-ζ)19 PKCs (Fig. 2). As discussed below, there are important differences in the mechanisms by which PDK-1 regulates phosphorylation of different PKCs, and these reflect the PI 3-K requirement for this event. Regardless of the mechanism, the notion that PDK-1 is the universal PKC upstream kinase is reinforced by the finding that all PKC isotypes form direct complexes with PDK-1 in transfected cells,18 and more importantly, that embryonic stem cells deleted in both PDK-1 alleles are defective in PKC-ζ Thr410 phosphorylation, and show accelerated degradation of all other PKCs.26 Because PDK-1-triggered phosphorylation of PKCs is necessary for protein stability, the inference is that phosphorylation at the activation loop cannot occur in the absence of PDK-1.
Phosphorylation at the Turn Motif Phosphorylation at the activation loop triggers autophosphorylation of the turn motif, a sequence enriched with Pro residues (Fig. 1). It is widely accepted that this site is regulated by autophosphorylation in both conventional, novel and atypical PKCs, and that this is required to yield a fully mature and functionally active kinase. Kinase inactive mutants of PKC-βII27 and PKC-δ are not phosphorylated at this position (Ser641and Ser643 respectively) and accumulate in the detergent insoluble fraction of transfected cells, even when the activation loop residue harbors a negative charge (Glu). Autophosphorylation at the turn motif is necessary for PKC function because selective dephosphorylation of this site abolishes kinase activity.15 Thus, once PKC has fully matured, the only requirement is negative charge at this motif, and studies have shown that this locks the enzyme into a catalytically competent, phosphatase resistant and
68
Protein Kinase C
thermally stable confromation.28,29 In addition, mutation of this residue to an Ala yields an inactive protein which accumulates in the detergent-insoluble fraction.29 As with the activation loop, compensating phosphorylations at other residues in the turn motif have also been reported, and this may account for the low residual activity of single turn motif point mutants of PKC-α (Thr638)28,30 and PKC-δ (Ser643).31 Atypical PKCs are also phosphorylated at their respective turn motifs, and in the case of PKC-ζ there is evidence that this is also an autophosphorylation event.32 Specifically, Thr560 in PKC-ζ is the sole autophosphorylation site and is regulated as a consequence of the PDK-1-dependent step, i.e., activation loop phosphorylation (Fig. 2C). It is also noteworthy that an additional step appears to be required for PKC-ζ activation in response to agonists such as insulin, and which is not dependent on phosphorylation but which leads to release of the pseudosubstrate from the substrate binding cavity.33 Interestingly, phosphorylation of the equivalent site in Akt/PKB is mediated by an unidentified heterologous upstream kinase.34 Whether phosphorylation at this site in Akt/PKB and PKCs also serves to target the kinases in proximity to their substrates at the appropriate cellular location remains to be determined, but would be consistent with the observation that this site conforms to a 14-3-3 protein binding site (see Chapter 5).
Phosphorylation at the Hydrophobic Site The final phosphorylation step in the processing of mature PKC occurs at the hydrophobic site. Again, the sequence surrounding this phosphorylation site is also conserved in all PKCs as well as other AGC kinases (Fig. 1). Curiously, the phosphorylatable residue in the hydrophobic motif of atypical PKC-ζ and PKC-ι/λ is replaced by a negatively charged Glu, and this is also the case in the related PRKs. There is considerable debate as to the precise mechanism of regulation of PKC hydrophobic site phosphorylation, and this extends to the Akt/PKB protein kinase. In the case of conventional PKC-βII, autophosphorylation at Ser660 has been shown to occur by an intramolecular mechanism.27 Similarly, kinase-inactive mutants of PKC-βII are not phosphorylated at this site in vivo, and this cannot be rescued by mutation of the activation loop and turn motifs with negatively-charged residues.27 Unlike the turn motif, PKC-α and PKC-βII mutants with a Ser to Ala substitution in this sequence retain catalytic competence, although with reduced thermal stability and phosphatase sensitivity.30 Therefore, at least in conventional PKCs, autophosphorylation accounts for regulation of the hydrophobic site (Fig. 2A). What is the mechanism for novel PKCs? Kinase inactive mutants of novel PKC-ε are also not phosphorylated at the hydrophobic Ser729 site,35 and again this does not depend on negative charge at the activation loop or turn motif, indicative of autophosphorylation. This is in contrast with one study which has shown that PKC-δ and PKC-ε hydrophobic site phosphorylation is not affected in cells pre-treated with chemical inhibitors of PKC.36 This has suggested that a heterologous upstream kinase exists, distinct from PDK-1, which mediates hydrophobic site phosphorylation (Fig. 2B). This may be the case for the novel PKC-δ. Studies the Parker group have reveled that phosphorylation of PKC-δ at Ser662 is both sensitive to rapamycin, an inhibitor of the Target of Rapamycin (TOR) pathway,37 not sensitive to PKC inhibitors,36 and can be mediated both in vitro and in transfected cells by a protein complex which comprises the atypical PKC-ζ isotype.37 Interestingly, purified recombinant PKC-ζ is not capable of directly phosphorylating PKC-δ at Ser662, suggesting that additional components may be required. It remains to be determined whether phosphorylation of PKC-δ by PKC-ζ at the hydrophobic site can occur under physiological conditions. Because phosphorylation of the hydrophobic site in both Akt/PKB as well as PKC-δ and PKC-ε can be inhibited with antagonists of the PI 3-K pathway, it has been proposed that a distinct kinase, termed ‘PDK-2’ exists.38 To date, no bona-fide protein kinase which fulfils this requirement has been described for either PKCs or Akt/PKB. How might these contrasting observations account for hydrophobic site phosphorylation? This motif also serves as a docking site for PDK-1, and this has been demonstrated for both PKCs39 as well as other AGC kinases
Regulation of Protein Kinase C by Phosphorylation
69
βII phosFigure 3. Model for the role of PIF and PDK-1 binding in the regulation of conventional PKC-β βII is localized to the membrane and constitutively associates with phorylation. Newly synthesized PKC-β PDK-1. This is mediated by interaction of PDK-1 with the unphosphorylated PKC hydrophobic site (Ser660), effectively masking this motif. PDK-1 phosphorylates the activation loop at Thr500 and is released from PKC. This is promoted by proteins which contain the PIF (PDK-1-interacting fragment) βII hydrophobic site is now accessible motif, which compete for binding to PDK-1. The newly exposed PKC-β for autophosphorylation. Fully phosphorylated PKC is now competent to bind to its allosteric activator, diacylglycerol (DAG) leading to activation. Adapted from ref. 46.
such as p90 RSK.40 In a search for PDK-1-interacting proteins, the Alessi group reported that a short peptide sequence derived from the carboxyl-terminus of PRK-1 bound with high affinity to a pocket in the PDK-1 catalytic domain.41 This sequence, which they termed PIF (PDK-1-interacting fragment) comprises the hydrophobic motif of kinases in which the phosphorylatable residue is replaced by a negatively charged Glu, including PRKs and atypical PKC-ζ and PKC-ι/λ. This observation provides a model for the role of PIF-containing proteins in promoting AGC kinase hydrophobic site autophosphorylation. In the case of PKC-βII, this has been directly demonstrated: PDK-1 binds to the unphosphorylated carboxyl-terminus in the inactive conformation of PKC-βII, and binding of PIF competes for the PDK-1-PKC interaction, effectively displacing PDK-1 from the carboxyl-terminus.39 The newly exposed hydrophobic site is now accessible for autophosphorylation (Fig. 3). This would also account for the observation that atypical PKC-ζ is able to mediate PKC-δ hydrophobic site phosphorylation,37 as discussed above. The hydrophobic motif of atypical PKC-ζ would effectively act as a PIF, displacing PDK-1 and allowing PKC-δ autophosphorylation at Ser662. A similar reaction mechanism could also account for the autophosphorylation of the Akt/PKB hydrophobic site (Ser473).34 Thus, PIF-induced hydrophobic site autophosphorylation may be a common regulatory mechanism for many AGC kinases. Whether a true heterologous kinase also exists which fulfils the requirement of a PDK-2 for other AGC kinases under physiological conditions remains to be determined.
The Role of the Phosphoinositide 3-Kinase Pathway As the name implies, PDK-1 was discovered by its ability to phosphorylate Akt/PKB only in the presence of the PI 3-K lipid activators PtdIns-3,4-P2 and PtdIns-3,4,5-P3.22,42 Like Akt/ PKB, PDK-1 has a Pleckstrin Homology (PH) domain which binds with high affinity and selectivity to both PtdIns-3,4-P2 and PtdIns-3,4,5-P3. Binding of these lipids to both PDK-1
70
Protein Kinase C
and Akt/PKB mediates their recruitment to the plasma membrane at the site of lipid synthesis.43 In addition, binding of PtdIns-3,4-P2/PtdIns-3,4,5-P3 to the Akt/PKB PH domain induces a conformational change which exposes the activation loop residue Thr410, which is now accessible to PDK-1.42 Therefore, PDK-1 has an absolute requirement for PtdIns-3,4-P2 and PtdIns-3,4,5-P3 when using Akt/PKB as a substrate. In the case of conventional PKC-βII, PI 3-K is not required for PDK-1-mediated activation loop phosphorylation, because treatment of cells with the PI 3-K inhibitors wortmannin and LY294002 does not lead to any appreciable loss of Thr500 phosphorylation.13 Consistent with this, although phosphorylation of PKC-βII by PDK-1 occurs at the membrane, this does not require the PDK-1 PH domain and therefore there is no requirement for PtdIns-3,4-P2/PtdIns-3,4,5-P3 for PDK-1-mediated phosphorylation of conventional PKCs13 (Fig. 2A). Conversely, treatment of cells with PI 3-K antagonists blocks PKC-δ, PKC-ε and PKC-ζ phosphorylation.18,19 This is in agreement with the observation that PtdIns-3,4,5-P3 increases the rate at which PDK-1 can phosphorylate PKC-δ and PKC-ζ in vitro. This appears to be mediated through the PDK-1 PH domain,18 although there are reports of a direct effect of PtdIns-3,4,5-P3 on PKC-ζ.33,44 Therefore, although PDK-1 is a constitutively active kinase which does not have an absolute requirement for PtdIns-3,4-P2 or PtdIns-3,4,5-P3, in certain cases these lipids are required for PDK-1 function (e.g., Akt/PKB, PKC-ζ; for reviews on PDK-1, see refs. 45,46). What remains to be determined is the precise role of PI 3-K in controlling PKC phosphorylation under physiological conditions, and whether this is a constitutive event, or whether additional requirements exist. PDK-1 is regulated not only by lipid binding, but also by Ser/Thr and Tyr phosphorylation,47,48 and the consequence of this on PKC function has not been explored. It is also likely that perturbation of the PI 3-K/PDK-1 pathway impacts on PKC phosphorylation. A good example of this the phosphorylation of PKC-δ which in adherent cells bypasses the requirement for PtdIns-3,4-P2/PtdIns-3,4,5-P3 but which is necessary in cells in suspension where matrix-integrin interactions are disrupted.49 In summary, initial studies which indicated both a PI 3-K and PLCγ-1 requirement for PKC activation20 in agonist-stimulated cells can now be explained by the fact that (at least) two independent lipid signaling pathways converge on PKC: the PI 3-K/PDK-1 pathway which is necessary to produce a fully phosphorylated, catalytically competent kinase, and the PLCγ-1-diacylglycerol pathway which provides the allosteric activator to induce the active conformation.
Regulation by Dephosphorylation The control of PKC activation by phosphorylation reflects the balance of PDK-1 activity, autophosphorylation as well as protein phosphatase activity, regulating dephosphorylation and ultimately degradation. Although little is known about phosphatase regulation of PKC, the original finding that long-term PMA treatment induces PKC dephosphorylation and degradation suggests that it is the active, allosteric activator-bound form of the kinase which is more sensitive to dephosphorylation.7 Indeed, this treatment results in accumulation of the enzyme in the detergent-insoluble fraction of the cell. Similarly, the PKC hydrophobic site is selectively dephosphorylated in cells which have been serum-starved.37,50 In addition, loss of phosphate at the PKC-ε hydrophobic motif (Ser729) is observed as quiescent fibroblasts are passaged over time, and this is mediated by a protein phosphatase.51 This event requires both serum as well as cell adhesion, and does not appear to involve either protein phosphatases 1, 2A or 2B. Rather, inhibitor studies implicate a rapamycin-sensitive phosphatase as well as the input of the Mitogen activated Protein Kinase (MAPK) pathway and a PKC-dependent phosphatase.52 In vitro, protein phosphatase 1 can dephosphorylate all three sites, whereas protein phosphatase 2A selectively dephosphorylates the activation loop and hydrophobic site but the turn motif.15 Although it is not known if this occurs in vivo, PMA-stimulation of cells leads to accumulation of a complex comprising PKC-α and protein phosphatase 2A in the membranes of cells.50 The precise nature of other protein phosphatases which control activation loop and turn motif dephosphorylation remain to be determined. However, binding of PKC to adapter proteins
Regulation of Protein Kinase C by Phosphorylation
71
such as A Kinase Associated Protein (AKAP)79 brings them into proximity with other associated proteins, and protein phosphatases have been reported to exist in such complexes, including protein phosphatase 2B/calcineurin53 (see Chapter 5). In summary, because dephosphorylation and degradation of PKC is intimately linked with its activated state, this provides the cell with a molecular basis for the integration of lipid signaling events in a temporal manner.
Implications for Cellular Function Because it is the fully phosphorylated, mature species of PKC which binds to diacylglycerol, leading to its activation (in conventional and novel isozymes), it is implied that regulation of PKC by phosphorylation is a pre-requisite for appropriate cellular responses. Indeed, numerous studies investigating the role of PKCs in cell physiology have made use of ‘dominant-negative’ mutants in which the activation loop residue is mutated to a non-phosphorylatable Ala. PKCs have been implicated in a plethora of signaling pathways leading to diverse responses, ranging from cell growth and survival, to programmed cell death (apoptosis), motility and migration as well as differentiation.54 Because phosphorylation and activation of PKC are in some cases (e.g., PKC-ζ) intimately linked with the PI 3-K pathway, one would expect that the PKC effectively serves as an effector of the signal relay. This is perhaps best illustrated for the atypical PKCs and their role in insulin-mediated glucose transport in insulin-responsive tissues and cells. Signaling through PI 3-K has been shown to regulate glucose uptake in response to insulin stimulation, and numerous studies have implicated both atypical PKC-ζ and PKC-ι/λ in this event. Specifically, translocation of PKC-ζ and PKC-ι/λ to the glucose transporter GLUT4 has been reported, and requires phosphorylation of the activation loop and turn motif residues in a PI 3-K-dependent manner.32,55 Expression of PKC-ζ induces GLUT4 translocation and glucose uptake in the absence of insulin stimulation, whereas a dominant negative PKC-ζ mutant blocks this event.55,56 Although the precise mechanism by which atypical PKCs regulate GLUT4 translocation and glucose uptake is not known, PKC-ζ-mediated phosphorylation of Vesicle-associated Membrane Protein-2 which is associated with GLUT4 and which cycles to the membrane in an insulin-dependent manner may represent one such mechanism.57 The regulation of PKC phosphorylation is also likely to have significant impact on other cellular processes. Of particular importance is the consequence of dephosphorylation and loss of PKC function in cell survival responses. In this regard, progressive dephosphorylation of PKC-βI has been reported as cells are passaged over time, and this correlates with increased apoptosis.58 Similarly, a PKC-α allele in which all three phosphorylation sites are mutated to Ala triggers an apoptotic response in transfected cells.58 A reduction in PKC-α protein using antisense oligonucleotides also results in increased apoptosis.59 Although conventional PKCs have been implicated in anti-apoptotic signaling in several cell types, the precise mechanism is not known, although phosphorylation of both the anti-apoptotic protein Bcl-2, c-Raf-1 and mitotic laminin have been proposed (reviewed in refs. 60,61). A major hurdle which has yet to be overcome in the field is the discovery of distinct substrates of PKC isotypes which are phosphorylated in response to various agonists. Studies with reportedly ‘specific’ PKC inhibitors have not been very informative in this regard. Similarly, the rather loose optimal phosphorylation motif preferred by PKCs has also not provided significant progress in this important area of research. The recent development of specific inhibitors of distinct kinases is likely to provide much needed information concerning downstream targets of PKC.62
Regulation of Protein Kinase C by Tyrosine Phosphorylation In addition to Ser/Thr phosphorylation in the catalytic domain, PKCs are also phosphorylated on tyrosine residues under certain conditions. In response to oxidative stress (H2O2) and pervanadate, a global inhibitor of tyrosine phosphatases, several PKC isoforms, particularly atypical PKC-ζ, have been shown to be tyrosine phosphorylated with a concomitant increase in protein kinase activity.63,64 A recent study showed that in nerve growth factor-stimulated
72
Protein Kinase C
PC12 cells, atypical PKC-ι/λ is phosphorylated on up to 5 distinct tyrosine residues by the Src tyrosine kinase.65 Of these, Tyr325 was shown to be required for NGF-dependent survival of these cells indicative of a novel pathway involving the NGF-receptor TrkA, Src and PKC-ι/λ. It is not clear if NGF-stimulated PKC-ι/λ phosphorylation is a specialized event in neuronal cell signaling which is not retained in other cells in response to other stimuli which also activate Src and Abl. Perhaps the best understood example of PKC tyrosine phosphorylation is for the novel PKC-δ isotype, which is also phosphorylated on multiple Tyr residues in response to stimulation of cells with numerous agonists.66 The net effect on PKC activity appears to be cell-type dependent as various studies have reported either an increase, decrease or no change in activity, at least as measured in vitro.67 A number of tyrosine kinases have also been implicated in PKC-δ tyrosine phosphorylation, including Src, Abl, Lyn, and the insulin and PDGF-receptors. Consistent with this, direct association between Abl and PKC-δ has been reported in cells stimulated subjected to oxidative stress.68 Recent studies have also addressed the consequence of PKC-δ tyrosine phosphorylation on cellular responses. Genotoxic agents such as etoposide induce apoptosis, and tyrosine phosphorylation of PKC-δ has been shown to be required for this pro-apoptotic event.69 Src-dependent tyrosine phosphorylation of PKC-δ has also been shown to contribute to the neoplastic phenotype of skin keratinocytes.70 In summary, tyrosine phosphorylation of novel PKC-δ and atypical PKC-ι/λ is an important regulatory mechanism under certain physiological conditions, particularly those which mediate apoptotic signaling.
Conclusions and Perspectives The discovery of PKC as the primary phorbol ester receptor and diacylglycerol sensor heralded a new area in lipid signaling in the early 1980s. The interaction of PKC with both phospholipids and diacylglycerol at the membrane not only provided another example of classic allosteric enzyme activation, but also demonstrated that spatial regulation of signaling events leading to phospholipid hydrolysis are critical for signal relay from the membrane to the cytosol. The subsequent discovery that PKC is also regulated by phosphorylation which controls both the maturation and activation of the kinase showed that integration of multiple lipid signaling pathways converging on PKC is necessary to achieve appropriate spatial as well as temporal signaling events. In addition, the finding the PKC phosphorylation is directly mediated by the PDK-1 enzyme suggested an intimate link with the PI 3-K pathway. While it is clear that this is the case for atypical PKCs which are true PI 3-K effectors in agonist-stimulated cells, this does not apply to conventional PKCs such that allosteric activation by diacylglycerol binding represents the activation step. With these latest discoveries, regulation of PKC has come round full circle, such that the mechanistic basis for the integration of the diacylglycerol and phosphorylation signals are well understood. A number of questions remain unanswered; how does regulated PDK-1 signaling affect PKC phosphorylation and function? Is there a bona-fide heterologous kinase for novel PKCs, particularly PKC-δ, which transphosphorylates the hydrophobic motif? How does PKC transduce signaling events at defined intracellular locations? Perhaps most pressing in the field is the need to characterize specific substrates for PKC isozymes which mediate the multitude of biological responses which have been attributed to its function. The advent of specific inhibitors to individual isozymes, knockout animals and cells, and genetic manipulation of PKC homologues in lower eukaryotes will undoubtedly provide significant progress in this area.
Acknowledgements I thank the members of my laboratory for many insightful discussions. Research in the laboratory is supported in part by grants from the National Institutes of Health (CA75134).
Regulation of Protein Kinase C by Phosphorylation
73
References 1. Orr JW, Newton AC. Intrapeptide regulation of protein kinase C. J Biol Chem 1994; 269(11):8383-8387. 2. Srinivasan N, Bax B, Blundell TL et al. Structural aspects of the functional modules in human protein kinase-C alpha deduced from comparative analyses. Proteins 1996; 26(2):217-235. 3. Nishikawa K, Toker A, Johannes FJ et al. Determination of the specific substrate sequence motifs of protein kinase C isozymes. J Biol Chem 1997; 272(2):952-960. 4. Van Lint J, Rykx A, Maeda Y et al. Protein Kinase D: An intracellular traffic regulator on the move. Trends in Cell Biology 2002; 12(4):193-200. 5. Flynn P, Mellor H, Palmer R et al. Multiple interactions of PRK1 with RhoA. Functional assignment of the Hr1 repeat motif. J Biol Chem 1998; 273(5):2698-2705. 6. Flynn P, Mellor H, Casamassima A et al. Rho GTPase control of protein kinase C-related protein kinase activation by 3-phosphoinositide-dependent protein kinase. J Biol Chem 2000; 275(15):11064-11070. 7. Borner C, Eppenberger U, Wyss R et al. Continuous synthesis of two protein-kinase-C-related proteins after down-regulation by phorbol esters. Proc Natl Acad Sci USA 1988; 85(7):2110-2114. 8. Pears C, Stabel S, Cazaubon S et al. Studies on the phosphorylation of protein kinase C-alpha. Biochem J 1992; 283(Pt 2):515-518. 9. Knighton DR, Zheng JH, Ten Eyck LF et al. Crystal structure of the catalytic subunit of cyclic adenosine monophosphate-dependent protein kinase. Science 1991; 253(5018):407-414. 10. Orr JW, Newton AC. Requirement for negative charge on “activation loop” of protein kinase C. J Biol Chem 1994; 269(44):27715-27718. 11. Cazaubon S, Bornancin F, Parker PJ. Threonine-497 is a critical site for permissive activation of protein kinase C alpha. Biochem J 1994; 301(Pt 2):443-448. 12. Pepio AM, Sossin WS. Membrane translocation of novel protein kinase Cs is regulated by phosphorylation of the C2 domain. J Biol Chem 2001; 276(6):3846-3855. 13. Sonnenburg ED, Gao T, Newton AC. The phosphoinositide-dependent kinase, PDK-1, phosphorylates conventional protein kinase C isozymes by a mechanism that is independent of phosphoinositide 3-kinase. J Biol Chem 2001; 276(48):45289-45297. 14. Orr JW, Newton AC. Interaction of protein kinase-C with phosphatidylserine. 1. Cooperativity in lipid binding. Biochemistry 1992; 31(19):4661-4667. 15. Keranen LM, Dutil EM, Newton AC. Protein kinase C is regulated in vivo by three functionally distinct phosphorylations. Curr Biol 1995; 5(12):1394-1403. 16. Stempka L, Schnolzer M, Radke S et al. Requirements of protein kinase cdelta for catalytic function. Role of glutamic acid 500 and autophosphorylation on serine 643. J Biol Chem 1999; 274(13):8886-8892. 17. Stempka L, Girod A, Muller HJ et al. Phosphorylation of protein kinase Cdelta (PKCdelta) at threonine 505 is not a prerequisite for enzymatic activity. Expression of rat PKCdelta and an alanine 505 mutant in bacteria in a functional form. J Biol Chem 1997; 272(10):6805-6811. 18. Le Good JA, Ziegler WH, Parekh DB et al. Protein kinase C isotypes controlled by phosphoinositide 3-kinase through the protein kinase PDK1. Science 1998; 281(5385):2042-2045. 19. Chou MM, Hou W, Johnson J et al. Regulation of protein kinase C zeta by PI 3-kinase and PDK-1. Curr Biol 1998; 8(19):1069-1077. 20. Moriya S, Kazlauskas A, Akimoto K et al. Platelet-derived growth factor activates protein kinase C epsilon through redundant and independent signaling pathways involving phospholipase C gamma or phosphatidylinositol 3-kinase. Proc Natl Acad Sci USA 1996; 93(1):151-155. 21. Muller G, Ayoub M, Storz P et al. PKC zeta is a molecular switch in signal transduction of TNF-alpha, bifunctionally regulated by ceramide and arachidonic acid. EMBO J 1995; 14(9):1961-1969. 22. Alessi DR, James SR, Downes CP et al. Characterization of a 3-phosphoinositide-dependent protein kinase which phosphorylates and activates protein kinase Balpha. Curr Biol 1997; 7(4):261-269. 23. Pullen N, Dennis PB, Andjelkovic M et al. Phosphorylation and activation of p70s6k by PDK1. Science 1998; 279(5351):707-710. 24. Alessi DR, Kozlowski MT, Weng QP et al. 3-Phosphoinositide-dependent protein kinase 1 (PDK1) phosphorylates and activates the p70 S6 kinase in vivo and in vitro. Curr Biol 1998; 8(2):69-81. 25. Dutil EM, Toker A, Newton AC. Regulation of conventional protein kinase C isozymes by phosphoinositide-dependent kinase 1 (PDK-1). Curr Biol 1998; 8(25):1366-1375. 26. Balendran A, Hare GR, Kieloch A et al. Further evidence that 3-phosphoinositide-dependent protein kinase-1 (PDK1) is required for the stability and phosphorylation of protein kinase C (PKC) isoforms. FEBS Lett 2000; 484(3):217-223.
74
Protein Kinase C
27. Behn-Krappa A, Newton AC. The hydrophobic phosphorylation motif of conventional protein kinase C is regulated by autophosphorylation. Curr Biol 1999; 9(14):728-737. 28. Bornancin F, Parker PJ. Phosphorylation of threonine 638 critically controls the dephosphorylation and inactivation of protein kinase Calpha. Curr Biol 1996; 6(9):1114-1123. 29. Edwards AS, Faux MC, Scott JD et al. Carboxyl-terminal phosphorylation regulates the function and subcellular localization of protein kinase C betaII. J Biol Chem 1999; 274(10):6461-6468. 30. Bornancin F, Parker PJ. Phosphorylation of protein kinase C-alpha on serine 657 controls the accumulation of active enzyme and contributes to its phosphatase- resistant state. J Biol Chem 1997; 272(6):3544-3549. 31. Li W, Zhang J, Bottaro DP et al. Identification of serine 643 of protein kinase C-delta as an important autophosphorylation site for its enzymatic activity. J Biol Chem 1997; 272(39):24550-24555. 32. Standaert ML, Bandyopadhyay G, Perez L et al. Insulin activates protein kinases C-zeta and C-lambda by an autophosphorylation-dependent mechanism and stimulates their translocation to GLUT4 vesicles and other membrane fractions in rat adipocytes. J Biol Chem 1999; 274(36):25308-25316. 33. Standaert ML, Bandyopadhyay G, Kanoh Y et al. Insulin and PIP3 activate PKC-zeta by mechanisms that are both dependent and independent of phosphorylation of activation loop (T410) and autophosphorylation (T560) sites. Biochemistry 2001; 40(1):249-255. 34. Toker A, Newton AC. Akt/Protein kinase B is regulated by autophosphorylation at the hypothetical PDK-2 site. J Biol Chem 2000; 275(12):8271-8274. 35. Cenni V, Doeppler H, Sonnenburg ED et al. Regulation of Novel Protein Kinase C e by Phosphorylation. Biochem J 2002; 363(3):537-545. 36. Parekh D, Ziegler W, Yonezawa K et al. Mammalian TOR controls one of two kinase pathways acting upon nPKCdelta and nPKCepsilon. J Biol Chem 1999; 274(49):34758-34764. 37. Ziegler WH, Parekh DB, Le Good JA et al. Rapamycin-sensitive phosphorylation of PKC on a carboxy-terminal site by an atypical PKC complex. Curr Biol 1999; 9(10):522-529. 38. Alessi DR, Andjelkovic M, Caudwell B et al. Mechanism of activation of protein kinase B by insulin and IGF-1. EMBO J 1996; 15(23):6541-6551. 39. Gao T, Toker A, Newton AC. The carboxyl terminus of protein kinase c provides a switch to regulate its interaction with the phosphoinositide-dependent kinase, PDK-1. J Biol Chem 2001; 276(22):19588-19596. 40. Frodin M, Jensen CJ, Merienne K et al. A phosphoserine-regulated docking site in the protein kinase RSK2 that recruits and activates PDK1. EMBO J 2000; 19(12):2924-2934. 41. Balendran A, Casamayor A, Deak M et al. PDK1 acquires PDK2 activity in the presence of a synthetic peptide derived from the carboxyl terminus of PRK2. Curr Biol 1999; 9(8):393-404. 42. Stokoe D, Stephens LR, Copeland T et al. Dual role of phosphatidylinositol-3,4,5-trisphosphate in the activation of protein kinase B. Science 1997; 277(5325):567-570. 43. Anderson KE, Coadwell J, Stephens LR et al. Translocation of PDK-1 to the plasma membrane is important in allowing PDK-1 to activate protein kinase B. Curr Biol 1998; 8(12):684-691. 44. Nakanishi H, Brewer KA, Exton JH. Activation of the zeta isozyme of protein kinase C by phosphatidylinositol 3,4,5-trisphosphate. J Biol Chem 1993; 268(1):13-16. 45. Vanhaesebroeck B, Alessi DR. The PI3K-PDK1 connection: more than just a road to PKB. Biochem J 2000; 346 Pt 3:561-576. 46. Toker A, Newton AC. Cellular signaling: pivoting around PDK-1. Cell 2000; 103(2):185-188. 47. Casamayor A, Morrice NA, Alessi DR. Phosphorylation of Ser-241 is essential for the activity of 3phosphoinositide-dependent protein kinase-1: identification of five sites of phosphorylation in vivo. Biochem J 1999; 342(Pt 2):287-292. 48. Park J, Hill MM, Hess D et al. Identification of tyrosine phosphorylation sites on 3-phosphoinositide-dependent protein kinase-1 and their role in regulating kinase activity. J Biol Chem 2001; 276(40):37459-37471. 49. Parekh DB, Katso RM, Leslie NR et al. Beta1-integrin and PTEN control the phosphorylation of protein kinase C. Biochem J 2000; 352 Pt 2:425-433. 50. Hansra G, Bornancin F, Whelan R et al. 12-O-Tetradecanoylphorbol-13-acetate-induced dephosphorylation of protein kinase Calpha correlates with the presence of a membrane- associated protein phosphatase 2A heterotrimer. J Biol Chem 1996; 271(51):32785-32788. 51. England K, Rumsby MG. Changes in protein kinase C epsilon phosphorylation status and intracellular localization as 3T3 and 3T6 fibroblasts grow to confluency and quiescence: a role for phosphorylation at ser-729? Biochem J 2000; 352(Pt 1):19-26. 52. England K, Watson J, Beale G et al. Signalling pathways regulating the dephosphorylation of ser729 in the hydrophobic domain of protein kinase cepsilon upon cell passage. J Biol Chem 2001; 276(13):10437-10442.
Regulation of Protein Kinase C by Phosphorylation
75
53. Faux MC, Rollins EN, Edwards AS et al. Mechanism of A-kinase-anchoring protein 79 (AKAP79) and protein kinase C interaction. Biochem J 1999; 343 Pt 2:443-452. 54. Toker A. Signaling through protein kinase C. Front Biosci 1998; 3:D1134-D1147. 55. Kotani K, Ogawa W, Matsumoto M et al. Requirement of atypical protein kinase clambda for insulin stimulation of glucose uptake but not for Akt activation in 3T3-L1 adipocytes. Mol Cell Biol 1998; 18(12):6971-6982. 56. Standaert ML, Galloway L, Karnam P et al. Protein kinase C-zeta as a downstream effector of phosphatidylinositol 3-kinase during insulin stimulation in rat adipocytes. Potential role in glucose transport. J Biol Chem 1997; 272(48):30075-30082. 57. Braiman L, Alt A, Kuroki T et al. Activation of protein kinase C zeta induces serine phosphorylation of VAMP2 in the GLUT4 compartment and increases glucose transport in skeletal muscle. Mol Cell Biol 2001; 21(22):7852-7861. 58. Whelan RD, Parker PJ. Loss of protein kinase C function induces an apoptotic response. Oncogene 1998; 16(15):1939-1944. 59. Haimovitz-Friedman A, Balaban N, McLoughlin M et al. Protein kinase C mediates basic fibroblast growth factor protection of endothelial cells against radiation-induced apoptosis. Cancer Res 1994; 54(10):2591-2597. 60. Cross TG, Scheel-Toellner D, Henriquez NV et al. Serine/threonine protein kinases and apoptosis. Exp Cell Res 2000; 256(1):34-41. 61. Newton AC, Toker A. Cellular regulation of protein kinase C. In: Storey KB, Storey JM, editors. Protein Adaptations and Signal Transduction: Elsevier Science B.V.; 2001. p. 163-173. 62. Bishop AC, Ubersax JA, Petsch DT et al. A chemical switch for inhibitor-sensitive alleles of any protein kinase. Nature 2000; 407(6802):395-401. 63. Konishi H, Tanaka M, Takemura Y et al. Activation of protein kinase C by tyrosine phosphorylation in response to H2O2. Proc Natl Acad Sci USA 1997; 94(21):11233-11237. 64. Konishi H, Yamauchi E, Taniguchi H et al. Phosphorylation sites of protein kinase C delta in H2O2-treated cells and its activation by tyrosine kinase in vitro. Proc Natl Acad Sci USA 2001; 98(12):6587-6592. 65. Wooten MW, Vandenplas ML, Seibenhener ML et al. Nerve growth factor stimulates multisite tyrosine phosphorylation and activation of the atypical protein kinase C’s via a src kinase pathway. Mol Cell Biol 2001; 21(24):8414-8427. 66. Li W, Mischak H, Yu JC et al. Tyrosine phosphorylation of protein kinase C-delta in response to its activation. J Biol Chem 1994; 269(4):2349-2352. 67. Gschwendt M. Protein kinase C delta. Eur J Biochem 1999; 259(3):555-564. 68. Kumar S, Bharti A, Mishra NC et al. Targeting of the c-Abl tyrosine kinase to mitochondria in the necrotic cell death response to oxidative stress. J Biol Chem 2001; 276(20):17281-17285. 69. Blass M, Kronfeld I, Kazimirsky G et al. Tyrosine phosphorylation of protein kinase Cdelta is essential for its apoptotic effect in response to etoposide. Mol Cell Biol 2002; 22(1):182-195. 70. Joseloff E, Cataisson C, Aamodt H et al. Src family kinases phosphorylate PKC delta ontyrosine residues and modify the neoplastic phenotype of skin keratinocytes. J Biol Chem 2002; 277(4):12318-12323.
76
Protein Kinase C
CHAPTER 7
Down-Regulation of Protein Kinase C Nigel T. Goode and Nicola Smart
Abstract
P
KC was discovered in 1977 as a kinase activated by proteolysis. However it soon followed that PKC was also activated by reversible binding to cofactors, and this latter route of activation became the main focus of study. The discovery that proteolysis occurs in response to physiological stimuli in the whole cell renewed interest in PKC down-regulation. Studies over the last five years have demonstrated that proteolytic cleavage, leading to activation or inactivation of PKC depending on the particular stimulus and cell type, plays a role in the control of a number of cellular responses. Over the same period, the role of various intracellular protease systems in PKC down-regulation has been investigated, with a significant increase in our knowledge bank but not necessarily clarification of roles of individual proteases: a number of systems appear to be involved. In this chapter, the mechanism and significance of PKC down-regulation are presented and discussed.
Introduction The mechanics and regulation of the activation of the protein kinase C (PKC) family has been extensively studied, and is well understood, as can be attested from most other chapters in this book. However, the flip side of the coin, the mechanism and regulation of PKC inactivation, is much less understood. This is despite the first reports of PKC in the late 1970s being linked with proteolytic activation,1 which is now considered an element of PKC inactivation. Reports of activation through the reversible binding of cofactors (phosphatidylserine [PS], diacylglycerol [DAG] and Ca2+) rather than irreversible proteolytic cleavage followed in the early 1980s (see refs. 2, 3 and previous chapters). Proteolysis was therefore considered an artefact of purification and the PKC research focus moved mainly to study reversible allosteric activation. However, interest in PKC proteolyis was renewed by reports in the mid 1980s which showed that the potent tumor promoters, phorbol esters, which exert many of their effects via the activation of PKC, led to a decrease in the amount of PKC in the cell.4,5 This effect, which was shown to occur at least in part through an increase in the rate of proteolysis of PKC, 6,7 became known as down-regulation. Down-regulation was considered a pharmacological response to the phorbol esters which, although mimicking DAG in their action as activating cofactors, were metabolised at a much slower rate within the cell. The net result of phorbol ester treatment was chronic PKC activation rather than the transient activation resulting from the classical short burst of DAG arising from physiological agonists. Chronic activation was associated with increased susceptibility to proteolytic attack and therefore down-regulation. Several groups continued to study down-regulation, not least because of the intellectual justification of the process—chronic activation of a signalling molecule should lead to an adaptive response whereby the cell’s sensitivity to the stimulant is decreased. This process is recognised to modulate the number of cell surface receptors. Occupancy of a specific receptor leads to a Protein Kinase C, 2nd Edition, edited by Lodewijk Dekker. ©2004 Eurekah.com and Kluwer Academic / Plenum Publishers.
Down-Regulation of Protein Kinase C
77
reduction in the number of receptors at the cell surface as a consequence of internalisation via receptor-mediated endocytosis. This response has presumably evolved as a mechanism to destroy the ligand, and therefore terminate the primary signalling event, by sorting to the lysosome. However, a number of the internalised receptors are also targeted to the lysosome where they are destroyed. Thus fewer receptors are available to be recycled to the cell surface after ligand dissociation. In the case of PKC, cell desensitization appears, likewise, to involve a decrease in the number of PKC molecules. Thus a larger percentage of the remaining molecules would need to be activated (by a more potent or concentrated stimulus) to elicit a PKC-mediated effect—the cell is therefore desensitized. Down-regulation may also be involved in switching off the PKC signal in the short term. Signalling is a dynamic process—binding of ligand to a receptor initiates a short burst of activation of the appropriate signalling molecules. This activity is terminated rapidly and effectively through negative feedback mechanisms. At the receptor level these mechanisms include transmodulation (desensitization) via binding of inhibitory proteins to receptors or feedback phosphorylation of the receptor itself. Signalling molecules may also be inhibited through feedback phosphorylation or through other mechanisms such as translocation away from their active environment or dissociation of signalling complexes. Alternatively, mechanisms for signal termination are built into the signal transducers themselves. For example, GTP-binding proteins contain intrinsic GTPase activity. The end result is that the binding of ligand to its receptor initiates a rapid but transient burst of signalling activity. Following termination of the signal, the cell is reprimed to perceive and respond to subsequent signals. Proteolytic destruction may be one mechanism for terminating PKC kinase activity and therefore switching off the PKC signal. Down-regulation has been exploited widely as an experimental tool. Pretreatment of cells with phorbol esters for 24h or more leads to removal of PKC from the cell, as judged by Western blotting and in vitro and in vivo kinase assays.8 Subsequent treatment is used to determine if a particular response is mediated by the PKC since the PKC element of the response should be ablated in these down-regulated cells. Conversely, if the tested response remains intact, PKC is thought not to be necessary to signal that response. This approach was reasonably informative but relied upon the preconceptions that down-regulation was a neutral process and did not occur under physiological conditions. Both of these assumptions have, or are likely to be, proven indefensible. Certainly PKC down-regulation, especially of the novel PKCs, has since been seen in cells treated with a number of different agonists9-11—the process has relevance under physiological conditions and correlates with chronic production of DAG. There have also been reports suggesting that PKC down-regulation is far from a neutral process which simply terminates the PKC signal. The contrary view, that down-regulation and/or the proteolytic fragments generated may have signalling functions in their own right, can be proposed. Down-regulation of PKC may generate alternative forms of PKC which may perform additional or supplementary functions to the intact holoenzyme, leading to the induction of a distinct cellular response when compared to that induced by PKC activation alone. This model is particularly appropriate to the discussion of the role of PKC proteolysis in programmed cell death (see below). It is also important to consider that down-regulation follows activation: any PKC protein arising from continued translation while proteolysis is occurring would be activated before being down-regulated - therefore active PKC, albeit at low levels, is present during the down-regulation process. Down-regulation should not be equated with an absolute lack of PKC activity. Down-regulation occurs in physiological situations and may have specific functional signalling roles—further investigation is warranted. This chapter will focus on two aspects of PKC down-regulation. Firstly, the current consensus regarding the mechanism(s) of down-regulation will be presented. Secondly, the function and consequence of PKC down-regulation will be discussed.
78
Protein Kinase C
The Proteolytic Mechanism PKC is subjected to proteolytic attack in the cell. In vitro studies established that proteases such as trypsin preferentially cleave PKC in the hinge region separating the catalytic and regulatory domains. The resulting kinase domain fragment (also known as PKM) possesses constitutive kinase activity since it has been physically separated from its regulatory domain.12-15 An increase in the rate of proteolysis can account for the down-regulation of PKCs in most circumstances,6 although some effects on expression have also been reported.16-18 The conformational change associated with activation and membrane association increases the susceptibility of the hinge region to proteolysis.15,19,20 Mutations designed to mimic activation also lead to increased proteolysis, both in vitro21 and in the whole cell.22 Activation is required for down-regulation,23,24 although in some cases kinase-dead PKC mutants can down-regulate, presumably drawn into the process which is stimulated by endogenous active PKCs.25-27 PKC must initially be activated and translocated from the cytosol to become associated with membranes, where down-regulation is initiated. Early research focused on the Ca2+-activated neutral proteases (calpains) as the PKC proteinase. The cofactor requirement of these proteases tallied with the Ca2+ requirement of the classical PKCs. PKCs were shown to colocalise with calpains in cells following increases in intracellular Ca2+ concentrations.28-32 In most reported cases, down-regulation is inhibited by calpain inhibitors32,33 although this effect is not always seen24,34 and inhibition is at best partial. The mM calpain isoform (m-calpain) has been shown to cleave PKC more efficiently than the µM calpain (µ-calpain) but µ-calpain is more effective in the presence of phospholipids.35 One problem with the calpain model is that it has been difficult to detect the free catalytic domain by Western blotting. This is probably due to the proteolytic fragments being extremely sensitive to subsequent cleavage by calapin.35 The more sensitive in vitro kinase assay has demonstrated increased constitutive PKC activity in extracts from cells treated with phorbol esters, which probably arises from the PKM fragment.12,36,37 Indeed, more recent experiments, taking advantage of the increased sensitivity through ECL detection, have detected fragments by Western blotting, especially in the presence of protease inhibitors (e.g., ref. 38). The development of specific antibodies which only detect the cleaved catalytic domain39,40 also point to proteolysis occurring in the cell in the hinge region at putative calpain consensus sites, thereby exposing a novel antibody epitope at the new amino-terminus. Thus calpains colocalise with active cPKCs and PKC down-regulation can be partially inhibited by calpain inhibitors. During down-regulation, PKC cleaved at or near the calpain site can be detected in whole cells.40 However, evidence is accumulating from recent reports that calpain mediated down-regulation is probably only one of several proteolytic mechanisms which function in PKC down-regulation. That other proteases can be involved was shown by experiments in which the hinge region of PKC-α was replaced with equivalent mobile ‘hinge’ regions of other proteins.41 These other protein donors were chosen since they were not known to be susceptible to proteolysis although they all resided at the membrane when activated. The chimaeric proteins displayed markedly different sensitivities to m-calpain in vitro but all down-regulated with indistinguishable kinetics in cells treated with phorbol esters. Thus a different down-regulation mechanism was suggested. The possible candidates are other cytosolic proteases, lysosomal proteases or other proteasome-associated proteases. Cases for all of these mechanisms have been proposed, as will be discussed now and are summarised in Figure 1. No firm evidence links lysosomal degradation with PKC down-regulation. PKC can associate with lysosomal membranes42,43 but presumably on the outer cytosolic face. This association may have more to do with increasing endocytic rate and vesicle traffic42,44 or participation in a signalling complex45 than with down-regulation per se. However, it has been shown that siting in a noncytosolic, detergent-insoluble environment occurs during PKC down-regulation - association with lysosomes may be required for subsequent degradation. This suggestion is supported by findings that temperature affects down-regulation, presumably via inhibition of membrane traffic,44 although Lee et al34 reported that inhibition of vesicle trafficking has no effect
Down-Regulation of Protein Kinase C
79
Figure 1. Simplified schemes for calpain, caspase-3 and ubiquitin/proteasome roles in PKC down-regulation. In all 3 cases, activation precedes proteolysis. Calpain and caspase cleave in the hinge region, generating separate regulatory and catalytic domains. Ubiquitination has been observed on the intact holoenzyme as well as PKC fragments following initial proteolysis. No attempt has been made to depict the involvement of phosphorylation status, transit through the detergent-insoluble compartment or vesicle traffic in down-regulation. PSS, pseudosubstrate site; *, active site; ubq, ubiquitin.
80
Protein Kinase C
on the rate of down-regulation. We have some evidence in support of a role for vesicle trafficking in PKC down-regulation. Coexpression of a GTPase-defective mammalian dynamin with mammalian PKC-γ in fission yeast led to a slower rate of PKC-γ down-regulation when compared with coexpression with wild type mammalian dynamin (Fig. 2). Thus dynamin-dependent vesicle trafficking is implicated in PKC down-regulation, although this effect was somewhat inconsistent (data not shown). Further reports discount any lysosomal involvement in PKC degradation. For example, a decrease in lysosomal pH had no effect on the rate of down-regulation37 nor did inhibitors of lysosomal hydrolases.34 These findings contrast with those relating to many cell surface receptors where down-regulation occurs in lysosomes and is inhibited by NH3 treatment. This is not unexpected since the extracellular domain of cell surface receptor destined to lysosomes following receptor-mediated endocytosis would reside in the lumen of the lysosome and therefore be exposed to attack by lysosomal hydrolases in contrast to PKC which would not be in contact with the lysosome lumen. In parallel with the increased knowledge about proteolysis in the proteasome, and ubquitin tagging of proteins destined for this route of destruction, several groups have shown that ubiquitinated PKCs are detectable in phorbol ester-treated cells where down-regulation is occurring.24,34,46 Furthermore, inhibition of the proteasome with lactacystin leads to the accumulation of ubiquitinated PKCs and also to inhibition of down-regulation.22,24,34 A role for ubiquitin and proteasomes in the down-regulation of PKC is convincing. It has also been shown that ubiquitinated PKC can be detected in cells treated with DAG or physiological stimulants when the proteasome is inhibited,22 suggesting that destruction in the proteasome is a consequence of activation and that down-regulation occurs as an exaggeration of this response when PKC is chronically activated. Interestingly, ubiquitination is dependent on PKC activity,46 in agreement with the activity dependency for PKC down-regulation. A broader role for PKC in the control of protein destruction is suggested by the finding that PKC can phosphorylate and up-regulate ubiquitin activity.47,48 This PKC effect on the rate of protein destruction may serve as a negative feedback loop, increasing the rate of destruction (and therefore inactivation) of active PKC itself. Additional mechanisms are also involved in the destruction of classical PKCs. A lower molecular weight form of PKC-α has been detected in cells chronically treated with phorbol esters49 and an intermediate form in the destuction process has been shown to be dephosphorylated.34,43,50 Dephosphorylation may predispose PKC to proteolysis and ubiquitinisation.34 Hansra et al43 propose the view that proteolysis and dephosphorylation are separate although related routes to terminate PKC signalling. Dephosphorylation temporally precedes down-regulation44 but the inter-relationship between these two paths has yet to be determined conclusively. However it is clear that phosphorylation status can affect the rate of ubiquitination and destruction of a number of proteins, 51 and PKC may be likewise affected. The down-regulation of novel PKCs may also be affected by phosphorylation status, but in a different manner. For example, tyrosine phosphorylation of PKC-δ protects it from down-regulation.52 Proteolytic cleavage of PKC, especially novel and atypical isotypes, is linked to the control of programmed cell death in a wide range of cell types in response to a number of different apoptotic stimuli. Initial reports linked PKC-δ cleavage to UV-induced apoptosis in U937 cells.53 Similarly, PKC-δ cleavage, but not other PKCs, followed UV treatment of human keratinocytes, and the apoptotic response was inhibited by PKC inhibitors.54 Caspase-3 was shown to be the protease responsible for PKC-δ cleavage and the apoptotic response was inhibited by PKC-δ-selective inhibitors and not by inhibitors of classical PKCs in keratinocytes as well as U937 cells and primary human neutrophils.54-57 Furthermore, the PKM form of PKC-δ was a potent apoptotic signal. Thus cleavage of PKC-δ is associated with apoptosis which is dependent on the kinase activity generated. Inhibition of the proteasome allied with PKC activation is also an apoptotic signal in U937 cells, whereas PKC inhibition reduces apoptosis, suggesting that PKC activity is the important signal rather than loss of activity.58 In other cells, such as Jurkat, it appears that the loss of nPKC activity is the important determinant for apoptosis,
Down-Regulation of Protein Kinase C
81
Figure 2. Coexpression of a mutant dynamin I decreases the rate of PKC-γ down-regulation. Mammalian PKC-γ was expressed in fission yeast cotransformed with a control vector (A) or the same vector containing wild-type (B) or GTPase-defective (C) mammalian dynamin I. The transformed S.pombe cells were treated with the phorbol ester 12-O-tetradecanoylphorbol 13-acetate at the indicated concentration for 24hours. The presence of either dynamin protein decreased the sensitivity of PKC-γ to down-regulation, but the effect was more marked with the mutant dynamin (C). The effect of the wild type dynamin may be explained by it acting as a partial antagonist to the yeast dynamin. The yeast cells expressed similar amounts of both the wild-type and mutant dynamin (data not shown).
since PKC activation inhibits apoptosis.57 The aPKCs are considered as anti-apoptotic kinases and the cleavage, once again catalysed by caspase-3, and subsequent destruction of kinase appears to be the important apoptotic signal.59,60 Interestingly, phorbol esters caused PKC activation and down-regulation in keratinocytes but this was not an apoptotic signal whereas UV-induced cleavage of PKC was.54 Other groups also claim a distinction between apoptosis and phorbol ester-induced down-regulation.57,61 Interestingly, the atypical PKCs-ζ and -λ are proteolysed during apoptosis in contrast to the behaviour of these isotypes in response to phorbol esters where they are considered to be down-regulation-resistant.62 This resistance may reflect their unresponsiveness to phorbol ester rather than a general insensitivity to down-regulation of these isotypes. Conversely, two
82
Protein Kinase C
distinct processes may exist and aPKCs are truly resistant to down-regulation whilst being sensitive to caspase-mediated proteolysis. The latter possibility is supported by the lack of known agonists that stimulate aPKC down-regulation. However, recent reports have shown that PKC-ζ can be ubiquitinated in cells.63,64 Only the active form of the enzyme is modified which is a small percentage of total PKC-ζ in the cell: the general lack of observable aPKC down-regulation may reflect a low level of activation rather than resistance to the down-regulation process per se. However, the functional and mechanistic similarities and distinctions between classical down-regulation and apoptotic cleavage are yet to be determined. Thus exact roles played by individual protease systems in PKC down-regulation are yet to be defined fully. It is relevant to note that not all isotypes behave equally; even amongst the phorbol ester-sensitive PKCs9,65 different mechanisms or sensitivities of down-regulation are suggested. Processing through several alternative proteolytic systems is one possible explanation.
Role of Down-Regulation Down-regulation has been an established fact of PKC biology for 18 years and proteolytic cleavage has been known even longer. Progress has been made on elucidating the mechanism of down-regulation, as discussed in the previous section, but less is known of the functional consequences except a general proposed role in cell desensitization. The catalytic domain (PKM) is a functional kinase. It is in fact deregulated and may have a different substrate specificity when compared to the full length PKC.66,67 The altered substrate specificity suggests that PKM may phosphorylate some proteins which are not substrates for the holoenzyme—specific functional consequences are hypothesised. Few reports suggest that the regulatory domain may also play a signalling role (see ref. 8) and there has been no confirmation of these ideas. Theoretically, the regulatory domain could act as a dominant negative factor, soaking up activating cofactors. However, it has proven difficult to detect isolated regulatory domains in whole cells, even in the presence of protease inhibitors. The half life of the fragment is very short, bringing into question whether the protein fragment could serve any role. Calpain inhibitors have been used widely to attempt to dissect the role of calpain and PKC down-regulation in a number of functional responses. For example, protease inhibitors block ionophoreinduced Ca2+ influx into neuroblastoma cells and this effect is mimicked by PKC inhibitors,68 suggesting that the cleavage of PKC is important in generating active PKC which triggers the response. Similarly, PKC activation induces the opposite effect to caplain inhibitors (proliferation versus neurite outgrowth, respectively, in neuroblastoma cells), implying that proteolysis equates with activation of PKC. The calpain inhibitors do inhibit PKC down-regulation in these cells. However, the PKC activation is dominant to calpain inhibitors, suggesting that the processes are not interdependent.69 In contrast, proteolysis is equated with PKC inactivation which inhibits cell-mediated cytotoxicity by NK cells.70 Similarly, loss of PKC activity is associated with anchorage-independent growth in partially transformed , c-src expressing, fibroblasts.71 The crucial isotype appears to be PKC-δ since protection of PKC-δ from down-regulation with bryostatin I prevents the effect whereas selective inhibition of this isotype with rottlerin induces the effect. These findings are supported by later experiments which showed that prevention of PKC-δ down-regulation with the protease inhibitor lactacystin also prevented the transformed state. This effect occurred despite PKC-δ being activated, as judged by translocation to the membrane, confirming that loss of activity rather than activation via the generation of PKM is responsible for the effect.24 Other isotypes are also designated specific roles through studies of differential down-regulation. For example, specific protection of PKC-ε from cleavage leads to more transformed phenotype in partially transformed NIH3T3 fibroblasts.72 Down-regulation of PKC-ε was associated with a less transformed phenotype but it is not clear whether the effect was due to a down-regulation-induced increase or decrease in PKC-ε kinase activity. Ras-associated transformation has also been associated with selective decrease in the expression in PKC-α
Down-Regulation of Protein Kinase C
83
and- ε, again correlating down-regulation with transformation.16 However it is not clear that the down-regulation seen is causative or simply a consequence of prolonged activation of the relevant isotypes in the face of sustained signalling activity. On the other hand PKC down-regulation does not always lead to a more transformed phenotype: down-regulation has also been correlated with accelerated differentiation in other cell lines, such as murine erythroleukemia and neuroblastoma cell lines.69,73 It is interesting to note that down-regulation may be prevented during certain stages of the cell cycle:17 the propensity of PKC to down-regulate in the face of sustained activation may reflect passage through the cell cycle rather than the choice between differentiation and proliferation.
Perspectives Clearly the jury is still out with regard to PKC down-regulation. Both of the important questions remain to be answered conclusively. Firstly, does down-regulation serve any additional purpose beyond terminating the signal and attenuating the cell’s responsiveness to PKC-mediated signalling? The supplementary mechanistic question, by what molecular mechanism does PKC down-regulation occur, is also unresolved, although a body of evidence is accumulating. Knowledge to address the first question will arise when the supplementary question is resolved, permitting specific and effective interference with the down-regulation process. The resulting ability to inhibit or stimulate down-regulation will be crucial to address the functional role of down-regulation. Alternatively, the generation and use of down-regulation-resistant forms of PKC may address these questions.
References 1. Inoue M, Kishimoto A, Takai Y et al. Studies on a cyclic nucleotide-independent protein kinase and its proenzyme in mammalian tissues. J Biol Chem 1977; 252:7610-7616. 2. Nishizuka Y. The molecular heterogeneity of protein kinase C and its implications for cellular regulation. Nature 1988; 334:661-665. 3. Dekker LV, Parker PJ. Protein kinase C—a question of specificity. Trends Biochem Sci 1994; 19:73-77. 4. Rodriguez-Pena A, Rozengurt E. Disappearance of calcium-sensitive, phospholipid-dependent protein kinase activity in phorbol ester-treated 3T3 cells. Biochem Biophys Res Commun 1984; 120:1053-1059. 5. Ballester R, Rosen OM. Fate of immunoprecipitable protein kinase C in GH3 cells treated with phorbol 12-myristate 12-acetate. J Biol Chem 1985; 26:15194-15199. 6. Stabel S, Rodriguez-Pena A, Young S et al. Quantitation of protein kinase C by immunoblot— expression in different cell lines and response to phorbol esters. J Cell Physiol 1987; 130:111-117. 7. Young S, Parker PJ, Ullrich A et al. Down-regulation of protein kinase C is due to an increased rate of degradation. Biochem J 1987; 244:775-779. 8. Parker PJ, Bosca L, Dekker L et al. Protein kinase C (PKC)-induced PKC degradation; a model for down-regulation. Biochem Soc Trans 1995; 23:153-155. 9. Kiley S, Schaap D, Parker PJ et al. Protein kinase C heterogeneity in GH4C1 rat pituitary cells. J Biol Chem 1990; 265:15704-15712. 10. Kiley SC, Parker PJ, Fabbro D et al. Differential regulation of protein kinase C isozymes by thyrotropin-releasing hormone in GH4C1 cells. J Biol Chem 1991; 266:23761-23768. 11. Olivier AR, Parker PJ. Bombesin, platelet-derived growth factor, and diacylglycerol induce selective membrane association and down-regulation of protein kinase C isotypes in Swiss 3T3 cells. J Biol Chem 1994; 269:2758-2763. 12. Kishimoto A, Kajikawa N, Shiota M et al. Proteolytic activation of calcium-activated phospholipid-dependent protein kinase by calcium-dependent neutral protease. J Biol Chem 1983; 258:1156-1164. 13. Huang KP, Huang FL. Conversion of protein kinase C from a calcium-dependent to an independent form of phorbol ester-binding protein by digestion with trypsin. Biochem Biophys Res Commun 1986; 139:320-326. 14. Melloni E, Ponteremoli S, Michetti M et al. The involvement of calpain in the activation of protein kinase C in neutrophils stimulated by phorbol myristic acid. J Biol Chem 1986; 261:4101-4105. 15. Kishimoto A, Mikawa K, Hashimoto K et al. Limited proteolysis of protein kinase C subspecies by calcium-dependent neutral protease (calpain). J Biol Chem 1989; 264:4088-4092.
84
Protein Kinase C
16. Fernandes ER, Ashendel CL. Ras downregulation of protein kinase C mRNA in CH3 10T1/2 fibroblasts. Mol Cacinog 1996; 17:23-34. 17. Olivier AR, Hansra G, Pettitt TR et al. The comitogenic combination of transforming growth factor β1 and bombesin protects protein kinase c-δ from late phase down-regulation, despite synergy in diacylglycerol accumulation. Biochem J 1996; 318:519-525. 18. Shih NY, Floyd-Smith G. Protein kinase C δ mRNA is down-regulated transcriptionally and post-transcriptionally by 12- O -tetradecanoylphorbol-13 -acetate. J Biol Chem 1996; 271:16040-16046. 19. Young S, Rothbard J, Parker PJ. A monoclonal antibody recognising the site of limited proteolysis of protein kinase C. Inhibition of down-regulation in vivo. Eur J Biochem 1988; 173:247-252. 20. Huang FL, Yoshida Y, Cunha-Melo JR et al. Differential down-regulation of protein kinase C isozymes. J Biol Chem 1989; 264:4238-4243. 21. Pears CJ, Kour G, House C et al. Mutagenesis of the pseudosubstrate site of protein kinase C leads to activation. Eur J Biochem 1990; 194:89-94. 22. Kang BS, French OG, Sando JJ et al. Activation-dependent degradation of protein kinase C η. Oncogene 2000; 19:4263-4272. 23. Ohno S, Konno Y, Akita Y et al. A point mutation at the putative ATP-binding site of protein kinase C α abolishes the kinase activity and renders it down-regulation insensitive. J Biol Chem 1990; 265:6296-6300. 24. Lu Z, Liu D, Hornia A et al. Activation of protein kinase C triggers its ubiquitination and degradation. Mol Cell Biol 1998; 18:839-845. 25. Pears C, Parker PJ. Down-regulation of a kinase-defective PKC α. FEBS Lett 1991; 284:120-122. 26. Freisenwinkel I, Rietmacher D, Stabel S. Downregulation of protein kinase C γ is independent of a functional kinase domain. FEBS Lett 1991; 280:262-266. 27. Goode NT, Hajibagheri MAN, Parker PJ. Protein kinase C (PKC)-induced PKC down-regulation; association with up-regulation of vesicle traffic. J Biol Chem 1995; 270:2669-2673. 28. Melloni E, Pontremoli S, Michetti M et al. Binding of protein kinase C to neutrophil membranes in the presence of calcium and its activation by calcium-requiring proteinase. Proc Natl Acad Sci USA 1985; 82:6435-6439. 29. Mellgren RL. Calcium-dependent proteases: An enzyme system active at cellular membranes? FASEB J 1987; 2:110-115. 30. Suzuki K, Imajoh S, Emori Y et al. Calcium-activated neutral protease and its endogenous inhibitor. Activation at the cell membrane and biological function. FEBS Lett 1987; 220:271-277. 31. Savart M, Letard P, Bultel S et al. Induction of protein kinase C down-regulation by the phorbol ester TPA in a calpain/protein kinase C complex. Int J Cancer 1992; 52:399-403. 32. Hong DH, Huan J, Ou BR et al. Protein kinase C isoforms in muscle cells and their regulation by phorbol ester and calpain. Biochim Biophys Acta 1995; 1267:45-54. 33. Adachi Y, Maki M, Ishii K et al. Possible involvement of calpain in down-regulation of protein kinase C. Adv Second Messenger Phosphoprotein Res 1990; 24:478-484. 34. Lee HW, Smith L, Petit GR et al. Bryostatin I and phorbol ester down-modulate protein kinase C-α and ε via the ubiquitin/proteasome pathway in human fibroblasts. Mol Pharmacol 1997; 51:439-447. 35. Cressman CM, Mohan PS, Nixon RA et al. Proteolysis of protein kinase C: mM and µM calcium-requiring calpains have different abilities to generate, and degrade the free catalytic subunit, protein kinase M. FEBS Lett 1995; 367:223-227. 36. Tapley PM, Murray AW. Modulation of calcium-activated, phospholipid-dependent protein kinase in platelets treated with a tumor-promoting phorbol ester. Biochem Biophys Res Commun 1984; 122:158-164. 37. Chida K, Kato N, Kuroki T. Down regulation of phorbol diester receptors by proteolytic degradation of protein kinase C in a cultured cell line of fetal rat skin keratinocytes. J Biol Chem 1986; 261:13013-13018. 38. Shea TB, Beermann ML, Griffin WR et al. Degradation of protein kinase C α and its free catalytic subuniy, protein kinase M, in intact human neuroblastoma cells and under cell-free conditions. Evidence that PKM is degraded by mM calpain-mediated proteolysis at a faster rate than PKC. FEBS Lett 1994; 350:223-229. 39. Kikuchi H, Imajoh-Ohmi S, Kanegasaki S. Novel antibodies specific for proteolyzed forms of protein kinase C: Production of anti-peptide antibodies available for in situ analysis of intracellular limited proteolysis. Biochim Biophys Acta 1993; 1162:171-176. 40. Kikuchi H, Imajoh-Ohmi S. Antibodies specific for proteolyzed forms of protein kinase C α. Biochim Biophys Acta 1995; 1269:253-259.
Down-Regulation of Protein Kinase C
85
41. Junco M, Webster C, Crawford C et al. Protein kinase C V3 domain mutants with differential sensitivities to m-calpain are not resistant to phorbol ester-induced down-regulation. Eur J Biochem 1994; 223:259-263. 42. Goode NT, Hajibagheri MAN, Warren G et al. Expression of mammalian protein kinase C is Schizosaccharomyces pombe; isotype-specific induction of growth arrest, vesicle formation and endocytosis. Mol Biol Cell 1994; 5:907-920. 43. Hansra G, Bornancin F, Whelan R et al. 12-O-Tetradecanoylphorbol-13-acetate-induced dephosphorylation of protein kinase C α correlates with the presence of a membrane-associated protein phosphatase 2A heterotrimer. J Biol Chem 1996; 271:32785-32788. 44. Hansra G, Garcia-Paramio P, Prevostel C et al. Multisite dephosphorylation and desensitization of conventional protein kinase C isotypes. Biochem J 1999; 342:337-344. 45. Kholodenko BN. MAP kinase cascade signalling and endocytic trafficking: A marriage of convenience? Trends Cell Biol 2002; 12:173-177. 46. Lee H-W, Smith L, Pettit GR et al. Ubiquitination of protein kinase C-α and degradation by the proteasome. J Biol Chem 1996; 271:20973-20976. 47. Kong SK, Chock PB. Proein ubiquitination is regulated by phosphorylation. An in vivo study. J Biol Chem 1992; 267:14189-14192. 48. Wall NR, Mohammad RM, Reddy KB et al. Bryostatin 1 induces ubiquitination and proteasome degradation of Bcl02 in the human acute lymphoblastic leukemia cell line, Reh. Int J Mol Med 2000; 5:165-171. 49. Borner C, Eppenberger U, Wyss R et al. Continuous synthesis of two protein kinase C-related proteins after down-regulation by phorbol esters. Proc Natl Acad Sci USA 1988; 85:2110-2114. 50. Lee HW, Smith L, Petit GR et al. Dephosphorylation of activated protein kinase C contributes to downregulation by bryostatin. Am J Physiol 1996; 271:C304-311. 51. OrforK, Crockett C, Jensen JP et al. Serine phosphorylation-regulated ubiquitination and degradation of β-catenin. J Biol Chem 1997; 272:24735-24738. 52. Wooten MW, Seibenhener ML, Heikkila JE et al. Delta protein kinase C phosphorylation parallels inhibition of nerve growth factor-induced differentiation independent of changes in Trk A and MAP kinase signalling in PC12 cells. Cell Signal 1998; 10:265-276. 53. Emoto Y, Manome Y, Meinhardt G et al. Proteolytic activation of protein kinase C δ by an ICE-like protease in apoptotic cells. EMBO J 1995; 14:6148-6156. 54. Denning MF, Wang Y, Nickoloff BJ et al. Protein kinase C δ is activated by caspase-dependent proteolysis during ultraviolet radiation-induced apoptosis of human keratinocytes. J Biol Chem 1998; 273:29995-30002. 55. Khwaja A, Tatton L. Caspase-mediated proteolysis and activation of protein kinase C δ plays a central role in neutrophil apoptosis. Blood 1999; 94:291-301. 56. Koriyama H, Kouchi Z, Umeda T et al. Proteolytic activation of protein kinase C δ and ε by caspase-3 in U937 cells during chemotherapeutic agent-induced apoptosis. Cell Signal 1999; 11:831-838. 57. Gomez-Angelats M, Bortner CD et al. Protein kinase C inhibits Fas receptor-induced apoptosis through modulation of the loss of K+ and cell shrinkage. J Biol Chem 2000; 275:19609-19619. 58. Vrana JA, Grant S. Synergistic induction of apoptosis in human leukaemia cells (U937) exposed to bryostatin 1 and the proteasome inhibitor lactacystin involves dysregulation of the PKC/MAPK cascade. Blood 2001; 97:2105-2114. 59. Frutos S, Moscat J, Diaz-Meco MT. Cleavage of ζPKC but not λ/τPKC by caspase-3 during UV-induced apoptosis. J Biol Chem 1999; 274:10765-10770. 60. Smith L, Chen L, Reyland ME et al. Activation of atypical protein kinase C ζ by caspase processing and degradation by the ubiquitin-proteasome system. J Biol Chem 2000; 275:40620-40627. 61. Mizuno K, Hoda K, Araki T et al. The proteolytic cleavage of protein kinase C isotypes, which generates kinase and regulatory fragments, correlates with Fas-mediated and TPA-induced apoptosis. Eur J Biochem 1997; 250:7-18. 62. Ways DK, Cook PP, Webster C et al. Effect of phorbol esters on PKC ζ. J Biol Chem 1992; 267:4799-4805. 63. Okuda H, Hirai S-I, Takaki Y et al. Direct interaction of the β-domain of VHL tumor suppressor protein with the regulatory domain of atypical PKC isotypes. Biochem Biophys Res Commun 1999; 263:491-497. 64. Okuda H, Saitoh K, Hirai S-I et al. The von Hippel-Lindau tumor suppressor protein mediates ubiquitination of activated atypical protein kinase C. J Biol Chem 2001; 276:43611-43617. 65. Olivier AR, Parker PJ. Identification of multiple protein kinase C isoforms in Swiss 3T3 cells— differential down-regulation by phorbol ester. J Cell Physiol 1992; 152:240-244.
86
Protein Kinase C
66. Nakabayashi H, Sellers JR, Huang KP. Catalytic fragment of protein kinase C exhibits altered substrate specificity toward smooth muscle myosin light chain. FEBS Lett 1991; 294:144-148. 67. Kochs G, Hummel R, Meyer D et al. Activation and substrate specificity of the human protein kinase C α and ζ isoenzymes. Eur J Biochem 1993; 216:597-606. 68. Shea TB, Spencer MJ, Beermann ML et al. Calcium influx into human neuroblastoma cells induces ALZ-50 immunoreactivity: Involvement of calpain-mediated hydrolysis of protein kinase C. J Neurochem 1996; 66:1539-1549. 69. Shea TB, Cressman CM, Spencer MJ et al. Enhancement of neurite outgrowth following calpain inhibition is mediated by protein kinase C. J Neurochem 1995; 65:517-527. 70. Shenoy AM, Brahmi Z. Inhibition of the calpain-mediated proteolysis of protein kinase C enhances lytic activity in human NK cells. Cell Immunol 1991; 138:24-34. 71. Lu Z, Hornia A, Jiang Y-W et al. Tumor promotion by depleting cells of protein kinase Cδ. Mol Cell Biol 1997; 17:3418-3428. 72. Hiwasa T, Nakata M, Nakata M et al. Regulation of transformed state by calpastatin via PKCε in NIH3T3 mouse fibroblasts. Biochem Biophys Res Commun 2002; 290:510-517. 73. Sparatore B, Passalacqua M, Pessino A et al. Modulation of the intracellular calcium-dependent proteolytic system is critically correlated with the kinetics of differentiation of murine erythroleukemia cells. Eur J Biochem 1994; 225:173-178.
Protein Kinase C in Yeast: The Cell Wall Integrity Pathway
87
CHAPTER 8
Protein Kinase C in Yeast: The Cell Wall Integrity Pathway Pilar Perez, Beatriz Santos and Pedro M. Coll
Introduction
B
roadly defined, cell integrity is the ability of a cell to keep its viability in its normal environment. In case of yeast and fungi, with a rigid cell wall, cell integrity includes processes such as correct cell wall biosynthesis, maintenance of a functional plasma membrane and development of responses to different stress conditions such as high temperature or hypotonic shock. Yeasts undergo polarized growth at different stages of the life cycle, such as apical growth, cytokinesis, mating or sporulation. To support polarized growth during the mitotic cycle or the mating process, remodelling of the cell wall must take place at specific growing sites. Because growth in yeast is always limited by the existing cell wall, the cells must organize regions where the wall will be weakened to allow the incorporation of new material.1 In these regions, a localized cell wall defect and hipo-osmolarity might originate the activation signal for the cell integrity pathway, that provide the coordinated regulation of cell wall biosynthetic enzymes and actin organization, necessary to keep cell viability.1-6 In this article we will discuss the role of yeast protein kinase C (PKC) homologues as main regulators of cell integrity, their mechanism of activation and the signalling pathways that they activate in the two model yeasts, Saccharomyces cerevisiae and Schizosaccharomyces pombe.
Yeasts PKC Homologues PKC1 from S. cerevisiae was identified by homology to proteins from the PKC family in animal cells.7 It is the only gene of that family in the budding yeast. Cells lacking Pkc1p lysed with a small bud.8,9 The phenotype is partially suppressed by an osmotic stabilizer suggesting a defective pkc1 cell wall, as corroborated by transmission electron microscopy.9 Pkc1p is a 132 kDa kinase with several regulatory domains. At the N-terminus of the protein there are two copies of the HR1 domain which is also present in other Rho-binding proteins and responsible for the binding of PKN/PRK1 and PRK2 to RhoA GTPase.10,11 Additionally, it has a C2 domain involved in calcium dependent phospholipid binding, the pseudosubstrate site, a C1 domain that might serve for diacylglycerol (DAG) binding, the kinase domain and the V5 region at the C-terminus. Biochemical data indicate that Pkc1p is neither activated by calcium nor by DAG.12 However, changing four cysteine residues to serines in the C1 domain leads to sensitivity to caffeine and low SDS concentrations when the mutated protein was expressed in cells lacking endogenous Pkc1p.13 The fission yeast has two PKC homologues, Pck1p and Pck2p,14 which are very similar to Pkc1p but they do not have a C2 calcium-binding domain.14,15 Both pck1+ and pck2+ genes share overlapping roles in cell viability and partially complement each other.14 The lytic phenotype of pck1∆ pck2∆ double deletion mutant indicates that the cell wall is co-ordinately regulated by both kinases.16 Protein Kinase C, 2nd Edition, edited by Lodewijk Dekker. ©2004 Eurekah.com and Kluwer Academic / Plenum Publishers.
88
Protein Kinase C
However, pck1∆ and pck2∆ strains show different behaviour. Whereas pck1∆ cells do not have an apparent phenotype, pck2∆ cells display abnormal morphology, polarity defects, and a very thin cell wall. In addition, pck2∆ cells are thermosensitive and hypersensitive to cell wall lytic enzymes,14,17,18 and pck2∆ protoplasts are unable to regenerate the cell wall in an osmotically stabilized liquid medium.19 Moreover, pck2+ overexpression is lethal and cells have very thick cell walls, whereas pck1+ overexpression does not cause any apparent phenotype.16
Rho GTPases and PKC Activation Genetic and biochemical studies have shown that, in S. cerevisiae, Pkc1p is directly activated by Rho1p.1,20,21 Besides, it has recently been shown that active Rho1p is required for the dynamic spatial and temporal localization of Pkc1p to sites of polarized growth and cell wall damage.22 In S. pombe, GTP-bound Rho1p and Rho2p interact with both Pck1p and Pck2p.16,23 The two HR1 motifs at the N-terminus of the kinases bind to Rho1p or Rho2p, and only deletion of both motifs completely abolishes the interaction detected.16 The N-terminus of Pck1p and Pck2p also contains possible PEST sequences, involved in degradation. In fact, both kinases are very unstable, and the interaction with GTP-Rho1p increases their stability.16,23 The stabilization of the Pck proteins in the vicinity of the activated Rho GTPases, would ensure the existence of a higher local concentration of these kinases, precisely in the areas where the cell is growing.24-27 Consistent with this idea, Pck1p and Pck2p co-localize with Rho1p in S. pombe.23 Similarly, it has been shown the interaction of the small GTPase Rho1p with the N-terminal HR1A region of S. cerevisiae Pkc1, in addition to the reported interaction of this protein with the C1 region of Pkc1p.27a This interaction is important for proper regulation of protein kinase C activity in vivo although it does not result in an activation of the kinase cascade.27a Although it is clear that Rho GTPases are required for PKC activation in yeast cells, the precise mechanism of activation has yet to be elucidated. In animal cells, the activities of the PKC and PRK family of kinases are controlled by phosphorylation of a conserved threonine residue within the activation loop of the kinase domain that is essential for activity.10,28 On the other hand, the pseudosubstrate site, conserved in all PKCs, is proposed to maintain the kinase in an inactive state. Thus, a conformational change is apparently involved in PKC activation. In fact, the Pkc1p mutation R398P, located in the pseudosubstrate site, results in constitutive activation of Pkc1p.20 In animal cells, the conservation of the activation loop within PKCs, PRKs and other kinases, such as PKB, led to the identification of the role of PDK1 in PKC and PRK phosphorylation.11,29-31 To achieve efficient phosphorylation, PDK1 and PKC need to be recruited to the membrane via allosteric activators. In the case of PRKs, the interaction between these kinases and PDK1 is dependent on active Rho. The binding of GTP-Rho leads to a conformational change in PRKs that allows PDK1 binding and phosphorylation of the PRK activation loop.11 A similar mechanism may also explain PKC activation in yeast. S. cerevisiae Pkh1p and Pkh2p are homologous to PDK1,32,33 and share an essential function that may well be the activation of Pkc1p. Pkh1p and Pkh2p can be substituted by human PDK1, suggesting a well conserved mechanism of action. In fact, Pkh2p in vitro phosphorylates Pkc1p at the threonine 983 residue in the activation loop. Moreover, mutant strains lacking Pkh2p and defective in Pkh1p (Pkh1D398G), display phenotypic similarities to mutants with low Pkc1p activity or with defects in the integrity pathway.32 Therefore, it is possible that Pkc1p activation is similar to that of PRKs in animal cells. Pkc1p is stabilized and localized by activated Rho1p to growth areas, where Pkh1p- or Pkh2p-mediated Pkc1p phosphorylation leads to the activation of the cell integrity pathway (Fig. 1). A single S. pombe gene, ksg1+, codes for a kinase that shows structural homology to the human PDK1.34 ksg1+ is an essential gene involved in signalling processes that controls S. pombe life cycle, and can be substituted by the human PDK1. It is possible that Ksg1p kinase activates Pck1p and Pck2p. In fact, physical interaction between those proteins has been already observed (E. Schweingruber, personal communication).
Protein Kinase C in Yeast: The Cell Wall Integrity Pathway
89
Figure 1. Schematic representation of S. cerevisiae cell integrity pathway. Rho1p is activated through Rom2p by either Tor2p or by sensors of the osmotic equilibrium, Wsc1-4p and Mid2p, that detect signals coming from the cell wall. Pkc1p is stabilized and localized by activated Rho1 to the growth areas where Pkh1p- or Pkh2p-mediated Pkc1p phosphorylation leads to the activation of the MAPK module and the transcription factors, Rlm1p and SBF, required to maintain the cell viability. A possible Pkc1 activation by inositol kinases, Stt4p and Mss4p, and Plc1p is also shown. Activations that may be indirect are shown as broken arrows, and those that are presumed have a question mark (?).
Another aspect of the yeast PKC activation that needs further investigation, is the involvement of inositol kinases. In mammalian cells, PKC activation is dependent on phosphatidylinositol-3-kinase (PI3K) activity as phosphatidyl-inositol 3,4,5-triphosphate (PIP3), the product of this activity, is required for an effective PDK1 relocalization from the cytosol to the plasma membrane.35 29,30 However, the only yeast protein with PI3K activity, Vps34p,36 seems to be essential for membrane traffic and has not been directly implicated in Pkc1p activation or cell integrity. Even so, other inositol kinases have been identified that might be related to Pkc1p. Thus, PKC1 and MSS4 were isolated as suppressors of the temperature-sensitive lethality caused by the stt4-1 mutation.37,38 STT4 codes for a phosphatidyl-inositol 4-kinase39 that localizes to the plasma membrane and functions in the Pkc1-mediated MAP kinase cascade.39a MSS4 was also isolated, together with PLC1, encoding a phospholipase C, RHO2, ROM2, and PKC1, as a multicopy suppressor of the growth defect of a tor2ts strain.40 Overexpression of MSS4 restores both growth and actin organization of the tor2ts strain. Mss4p, identified as a membrane phosphatidyl-inositol-4-phosphate-5-kinase (PIP5K), is required for the organization of the actin cytoskeleton,41 and cell morphogenesis.42 PIP5K generates phosphatidyl-inositol 4,5-biphosphate (PIP2) that can serve as substrate for PI3K and can also be hydrolized by phospholipase C to produce IP3 and diacylglycerol (DAG), a PKC activator in animal cells. However, as mentioned above, it is not clear if Pkc1p is activated by DAG.12 PIP2 can also serve to mediate the attachment to the membrane of ROM2, a Rho1 GEF that contains a PH domain essencial for its activity.42a
90
Protein Kinase C
S. cerevisiae Pkc1p and the Cell Integrity Pathway The involvement of Pkc1p in the regulation of S. cerevisiae cell wall biosynthesis is well known.2-4 Rho1p exerts a key function in this pathway since, as mentioned, participates in Pkc1p activation and is also a positive regulator of the (1,3)β-D-glucan synthase, responsible for the synthesis of the S. cerevisiae main cell wall polymer.43-45 Rho1p is positively regulated by the guanosine nucleotide exchange factors (GEFs) such as Rom1p, Rom2p,46 or Tus1p,46a and negatively regulated by the GTPase-activating proteins (GAPs), Bem2p, Sac7p, Bag7 and Lrg1,47-50a and by the GDP dissociation inhibitor (GDI), Rdi1p.51 Rho1p is activated by cell wall defects.52 Signals coming from the cell wall are detected by a family of membrane receptors that act as sensors of the osmotic equilibrium.53,53a These proteins, Wsc1 to 4, and Mid2p, share some characteristic motifs such as serine/threonine-rich N-terminal regions and C-terminal regions with charged residues.54 Wsc1p and Mid2p act during cell wall remodelling required for vegetative growth. Additionally, Mid2p acts during the mating process.55,56 Both proteins interact and activate Rho1p through Rom2p.57 Targets of rapamycin (TORs) are conserved phosphatidylinositol kinases involved in the coordination between nutritional or mitogenic signals and cell growth. TOR2 encodes an essential 282 kDa phosphatidylinositol-kinase while TOR1 is a nonessential homologue. The TOR signalling pathway broadly controls nutrient metabolism by sequestering several nutrient-regulated transcription factors in the cytoplasm.58 tor1∆ tor2∆ double disruption confers G1 arrest.59 Interestingly, Tor2p also regulates actin organization, acting on Rho1p and Pck1p through Rom2p.60 Thus, Tor2p might act in the cell integrity pathway as a nutrient sensor that connects protein synthesis, required for active growth, with cell wall biosynthesis and cell integrity. The major role in the activation of Rho1p in the cell integrity pathway is played by Rom2p. rom1 null mutant strain does not display the thermosensitive lytic phenotype typical from rom2∆, but is synthetically lethal with rom2∆, suggesting a crucial shared function. The main target of Pck1p activation is the MAPK cascade module (Fig. 1), consisting on Bck1p (MAPKKK), Mkk1p and Mkk2p, two redundant MAPKKs and the MAPK Slt2/Mpk1p. Mutants in the MAPK signalling cascade undergo cell lysis because of a deficiency in cell wall construction.61-63 Activation of Rho1p and Pck1p leads to sequential phosphorylation that is responsible for signal transduction through this protein kinase cascade. Thus, simultaneous elimination of the two Rho1-GAPs, Sac7p and Bem2p, causes constitutive Slt2/Mpk1p activation.64 BCK1, identified independently by three groups,61,63,65 was related to Pkc1p signalling because a dominant mutant allele was an extragenic suppressor of a pkc1 null mutant. BCK1 encodes a 163 kDa protein kinase, sharing 45% amino acid identity with its relative, STE11. bck1∆ resulted in a temperature-sensitive (ts) cell lysis defect, which was suppressed by osmotic stabilizing agents.66 MKK1 and MKK2 were isolated as genes that, when overexpressed, suppressed the cell lysis defect of a temperature-sensitive pkc1 mutant.63 Deletion of either MKK gene does not cause any apparent phenotypic defect, but deletion of both results in a thermosensitive cell lysis defect that is suppressed by osmotic stabilizers. Overexpression of MKK1 suppressed the ts growth defect of bck1∆ mutants, whereas an activated allele of BCK1 (BCK1-20) did not suppress the defect of the mkk1∆ mkk2∆ double mutant. SLT2/MPK1 was first cloned by complementation of an autolytic mutation,67 and was later isolated as a dosage-dependent suppressor of the cell lysis defect associated with bck1∆.65 Overexpression of SLT2/MPK1, also suppressed the defect of the mkk1∆ mkk2∆ double mutant.63 Thus, Slt2/Mpk1p represents the MAPK mediating the Pkc1p signalling.65 Components of the MAPK cascade also interact with proteins involved in polarity. It has been proposed that the formation of a multi protein complex localizes the MAPK module to sites of growth.68 Genetic and biochemical data support the fact that this MAPK pathway is specifically activated during periods of polarized growth (e.g., bud emergence or mating projection formation),69 and in stress conditions of hyperthermia or hypo-osmosis.21,70
Protein Kinase C in Yeast: The Cell Wall Integrity Pathway
91
Rlm1p is the main transcription factor responsible for the Slt2/Mpk1p cell wall regulation.71,72 Rlm1p is a MADS-box protein with a DNA-binding domain similar to that of mammalian MEF2,73 and is directly phosphorylated by Slt2/Mpk1p.71 RLM1 mutations confer resistance to the toxic effects of constitutive activity of the Mpk1p pathway.73 At least 25 genes regulated by Slt2/Mpk1p and Rlm1p have been described,72 and most of them are involved in the biosynthesis of the cell wall. The yeast two-hybrid system has identified Mlp1p, another MAPK-like protein, as an Rlm1-associated protein.71 However, mlp1∆ has no obvious phenotype and there is not genetic interactions between MLP1 and SLT2/MPK1, suggesting that Rlm1p is independently regulated by Mpk1p and the Mlp1p kinase. The phenotypes of the rlm1∆ mutant are less severe than those associated with loss of upstream components of the pathway. Unlike MAPK mutants, rlm1∆ cells are not temperature sensitive. This might be because Slt2/Mpk1p also activates a second transcription complex, SBF.74 Composed of Swi4p and Swi6p, SBF induces the expression of G1 cyclins Cln1p and Cln2p, and activates certain genes involved in bud formation and cell wall biosynthesis.75,76 Mutants lacking Pkc1p lyse at all temperatures, whereas the MAPK pathway is only essential at high temperatures, suggesting that Pkc1p also acts through other effectors. In fact, some signalling events that require Pkc1p, including actin depolarisation,77 repression of ribosome synthesis and rRNA in secretion defective cells,78 or transient relocation of many nuclear proteins to the cytoplasm during hypertonic shock conditions,79 do not require the downstream MAPK cascade. Another possible Pkc1p effector is Bck2p. The BCK2 gene was originally characterized as a multicopy suppressor of cell lysis defects in slt2/mpk1∆ or pck1∆ mutants,80 and is essential for normal growth in the absence of CLN3.81 BCK2 encodes a putative 92-kDa serine/threonine protein of unknown function that might signal in a branch of the PKC1-activated integrity pathway different from the MAPK pathway. The CWH43 product, a protein with putative sensor and transporter domains, is upstream of this parallel Pkc1-Bck2 signalling pathway.82 Interestingly, Pkc1p also directly interacts with several subunits of the oligosaccharyltransferase, and is required for maximal N-glycosylation activity, but the MAPK pathway is not involved in this function.83
Cell Integrity Pathway and Cell Cycle In order to retain cell viability and growth polarity, the integrity pathway must be perfectly coordinated with the cell cycle. Thus, there is a cell cycle regulation of the Pkc1-pathway, and in turn, there is a morphogenetic checkpoint regulating the cell cycle. The formation of a new bud in S. cerevisiae occurs at the G1/S transition of the cell cycle, and some genes involved in cell wall biosynthesis are periodically expressed at the G1/S phase.75,76 Several lines of evidence suggest that Cdc28p, the kinase regulating cell cycle progression, regulates the cell integrity pathway, which in turn promotes the transcription of cell wall biosynthesis genes. Slt2/Mpk1p activity is stimulated at the G1-S transition of the cell cycle.84 This activation is partially dependent on Cdc28-Cln1 or Cdc28-Cln2,69 and Pkc1p is required. However, the mechanism by which Cdc28p stimulates Pkc1p has yet to be elucidated. Possibly, the Cdc28p-induced hydrolysis of phosphatidylcholine to choline phosphate and DAG might activate Pkc1p.85 However, as already indicated, the regulation of Pkc1p by DAG has not been established. The transcription complexes SBF and MBF mediate the G1-S transition in the cell cycle of S. cerevisiae by regulating the expression of G1 cyclins. As mentioned, SBF also regulates the transcription of some cell wall biosynthesis genes expressed at the G1/S phase 75,76. Activation of SBF depends on Slt2/Mpk1p,74 the cyclin Cln3p, and the Pkc1p target protein Bck2p. Unlike Cln3p, Bck2p is capable of inducing SBF in the absence of functional Cdc28p.86 Taken together, these data suggest a multiple regulation of the cell integrity pathway by components of the cell cycle, with the SBF complex and the PKC1-MAPK pathway acting in concert to maintain cell integrity during bud formation at the G1/S transition.
92
Protein Kinase C
Figure 2. Schematic model of S. cerevisiae morphogenetic checkpoint pathway. Environmental and cell wall stresses, cause a transient depolarisation of the actin cytoskeleton that delays cell cycle progression in G2. The signal to actin is mediated by the plasma membrane sensor proteins, Rho1p and Pkc1p. G2 arrest caused by actin perturbation involves the inhibition of Cdc28p by Swe1p accumulation. Simultaneously, the cell integrity pathway components Rho1p, Pkc1p and MAPKs, are required for repolarization of the actin cytoskeleton and downregulation of the phosphatase Mih1p, that is mediated by activated Slt2/Mpk1p.
Saccharomyces cerevisiae cells sense many environmental and cell wall stresses, and respond by transiently depolarizing the actin cytoskeleton.41,77 A morphogenetic checkpoint delays cell cycle progression in G2 when the actin cytoskeleton is perturbed. This pause allows time for cells to complete bud formation prior to mitosis (Fig. 2), and couples cell cycle progression to proper bud formation. Possibly as a mechanism to repair general cell wall, actin depolarization also leads to a transient depolarized distribution of the biosynthetic enzyme (1,3)β-D-glucan synthase, Fks1p, and Rho1p.77 The signal for depolarization of the actin cytoskeleton is mediated by the plasma membrane sensor proteins, Rho1p, and Pkc1p. The PKC1-activated MAPK cascade is not required for depolarization, but is necessary for repolarization of the actin cytoskeleton and redistribution of Fks1p.77 The morphogenetic checkpoint-induced G2 arrest caused by actin perturbation, involves the inhibition of Cdc28p by the Swe1p kinase. The phosphatase Mih1p, homologous to S. pombe Cdc25p, counteracts Swe1p by dephosphorylating and activating Cdc28p. Swe1p kinase is normally stable during G1 and S phases but is unstable during G2 and M phases, due to ubiquitination and subsequent degradation. Perturbations of actin organization lead to stabilization and accumulation of Swe1p.87 It has recently been described that cell integrity pathway components Rho1p, Pkc1p and MAPKs, are required for an effective checkpoint response mediated by activated Slt2/Mpk1p.88 This kinase seems to function down-regulating the phosphatase Mih1p, in a different pathway to that leading to Swe1p stabilization. The membrane sensors and Rom2p are not necessary and the transcription factors Rlm1p and SBF do not participate either.
Protein Kinase C in Yeast: The Cell Wall Integrity Pathway
93
Interaction with Other Pathways Genetic evidence indicate that the cell integrity pathway participates in the regulation of the spindle pole body (SPB) duplication.89 Mutations in SLT2/MPK1 aggravate the G2/M arrest of mutants in SPB components, caused by the failure in SPB duplication. In contrast, overexpression of different PKC1 pathway components suppresses the G2/M arrest. Moreover, the phosphorylation state of Spc110p, an SPB component, depends on the presence of Slt2/ Mpk1p.89 Thus, the PKC1 pathway is a central coordinator of two major events, bud emergence and SPB duplication, that occurs at the G1/S transition. The PKC integrity pathway is also activated in response to pheromones. It is probably an indirect effect, due to the cell wall re-organization required for mating. Initially, Pkc1p is activated through the Mid2p sensor, to ensure cell wall integrity during projection formation.55 Cell wall degradation and remodelling occurs quickly during mating, and must be highly controlled in order to prevent lysis. The osmotic state of the cells regulates cell fusion and Pkc1p is not necessary for this process. By contrast, an activated allele of PKC1 blocks cell fusion.90 The cell integrity pathway also shares some functions with calcium signalling pathways. Increases in cell calcium concentration are always parallel to cell integrity activation in situations such as G1/S transition, mating or hypotonic stress. Moreover, conditional alleles of PKC1 are suppressed by calcium.8 The calcium-dependent phosphatase, calcineurine, is not required for normal cell growth but a null mutation is lethal in combination with pkc1∆ or slt2∆.91 Additionally, overexpression of constitutively active calcineurine is able to suppress the thermosensitive phenotype of pkc1∆ or slt2∆.91 This is probably due to the dual regulation of some cell wall biosynthesis genes such as FKS2.92 It has not been established whether cells require coordinated activation of both signalling pathways and/or cytosolic calcium is required to activate the cell integrity pathway. Pkc1p has a C2 domain but its kinase activity is not affected by calcium in vitro.12 Pkc1 pathway is also required for viability in quiescence. pkc1 and mpk1 mutants rapidly die by cell lysis upon carbon or nitrogen starvation.92a
Function of Pck1p and Pck2p in S. pombe Cell Integrity and Cell Wall Biosynthesis The fission yeast PKCs are very close to Pkc1p, and the Rho1p activation mechanism seems to be conserved. Additionally, in S. pombe, Rho2p also activates Pck2p.27 However, very little is known about the signalling mechanism from the cell surface to the Rho GTPases. Two S. pombe TOR homologues, tor1+ and tor2+ have recently been described. tor2+ is an essential gene, whereas tor1+ is only required under starvation and other stress conditions.93 It is not known if these inositol kinases signal through Pck1p or Pck2p. Two possible homologues to Rom1p and Rom2p also exist in S. pombe genome but have not been studied yet. Surprisingly, in S. pombe, no signalling through a MAPK cascade has been described for either Pck1p or Pck2p. Fission yeast has three different MAPK pathways. 3 The Mkh1p-Pek1p-Spm1p module regulates cell integrity and antagonizes chloride homeostasis.18,94-96 Mutants in either mkh1, pek1 or pmk1/spm1, show abnormal morphology and sensitivity to β-glucanases. Furthermore, Pmk1/Spm1p is a structural and functional homologue of Slt2/Mpk1p.18 However, in contrast with S. cerevisiae, this MAPK module is activated under hypertonic and heat shock conditions,95 and is not regulated by either Pck1p or Pck2p. It has recently been described that Mkh1p-Pek1p-Spm1p is activated by Shk2/Pak2p, a kinase of the PAK family,97,98 which is in turn regulated by the Cdc42p GTPase.99 Based on these results it has been proposed that the Cdc42-Pak2-Mkh1 pathway regulates cell integrity during cytokinesis (Fig. 3A). Pck1p and Pck2p kinases are essential and required for the maintenance of (1-3)β-D glucan synthase activity.16 pck1∆ or pck2∆ single mutants have less cell wall although they maintain the normal proportions in all three major polymers. Moreover, (1-3)β-D-glucan synthase activity level was not decreased in pck1∆ or pck2∆ cells. It seems that both kinases can substitute
94
Protein Kinase C
Figure 3. A) Proposed model for the Cdc42-Pak2-MAPK pathway regulating cell integrity during cytokinesis. B) Current model for Pck-mediated regulation of cell wall biosynthesis in S. pombe. Rho1p regulates (1-3)β-D-glucan synthase directly, and through Pck2p and Pck1p. Rho2p regulates the biosynthesis of α-D-glucan exclusively through Pck2p. Rho proteins stabilize and localize Pck2p and Pck1p to the growth areas and Ksg1p phosphorylates these kinases. Activations that may be indirect are shown as broken arrows and those that are presumed have a question mark (?).
for each other and that a minimum level of kinase C activity is required for the maintenance of (1-3)β-D glucan synthase. However, overexpression of pck2+, but not pck1+, causes a general increase in cell wall biosynthesis that correlates with an increment in the β-glucan content, and activation of the (1-3)β-D glucan synthase activity.16 Additionally, Pck2p plays a crucial role in the regulation of mok1+, a gene encoding the putative α-D-glucan synthase.100 Pck2p is required for Mok1p localization to the growth areas, and this pathway is mainly regulated by Rho2p. Interestingly, none of the genetic interactions observed between mok1+ and rho2+ or pck2+ could be reproduced with pck1+,27 indicating that Pck1p is not involved in α-D-glucan biosynthesis. Taking together all the available data, we propose the model presented in Figure 3B. Rho1p directly regulates (1-3)β -D-glucan biosynthesis as well as through Pck2p and Pck1p.16,23 On the other hand, Rho2p regulates the biosynthesis of α-D-glucan exclusively through Pck2p. Rho1p interacts with, stabilizes and localizes Pck2p and Pck1p to the growth areas.16,23 Rho2p, also localizes to the growth areas, interacts with, and signals to Mok1p through Pck2p. It seems therefore, that both GTPases mainly use Pck2p to coordinately regulate the biosynthesis of the two main S. pombe cell wall polymers. The role of Pck1p in cell wall integrity remains to be established. It might be involved in cell integrity through a different pathway, since pck1+ shows genetic interactions with ras1+ and ral1+.16 This might reflect a possible functional pathway controlled, directly or indirectly, by pck1+ and ras1+. pck1+ also showed genetic interaction with ppe1+, a gene encoding a type 2A phosphatase. ppe1∆ cells are short and pear-shaped,101 and this phenotype is suppressed by pck1+ overexpression. Pck2p, but not Pck1p, also shares some functions with calcium signalling pathways. A genetic interaction between pkc2+ and ehs1+ has been established. ehs1+ codes for a protein similar to S. cerevisiae Mid1p, and is involved in intracellular calcium accumulation.102 Mid1p has been proposed to be part of a calcium channel.103 Overexpression of pck2+ causes a strong calcium accumulation that depends on the presence of Ehs1p, suggesting a regulation of the channel mediated by Pck2p. On the other hand, the phenotypes caused by overexpression of pck2+ are suppressed in ehs1∆ cells.102 These results suggest that cytosolic calcium is required to activate Pck2p, although this kinase does not have a C2 domain.
Protein Kinase C in Yeast: The Cell Wall Integrity Pathway
95
Conclusion Extensive work on S. cerevisiae PKC1-regulated integrity pathway has revealed how the complicated process of cell wall biosynthesis is regulated at different levels during different stages of the life cycle. By contrast, in the fission yeast, Pck1p, Pck2p and the MAPK cascade Mkh1p-Pekp-Pmk1/Spm1p have been implicated in cell integrity, but a specific pathway for cell integrity has yet to be established. Rho GTPases, PKCs and MAPKs seem to be very conserved among different fungi,3,104,105 but their cell integrity pathways are not known yet. Clarification of the relationship between these proteins and their direct effectors will help to unveil general mechanisms of cell wall assembly and maintenance of cell integrity.
Acknowledgements We thank J.C. Ribas, A. Duran and A. Castellino, for their help with the manuscript and for stimulating discussions.
References 1. Cabib E, Drgonova J, Drgon T. Role of small G proteins in yeast cell polarization and wall biosynthesis. Annu Rev Biochem 1998; 67:307-333. 2. Gustin MC, Albertyn J, Alexander M, Davenport K. MAP Kinase pathways in the Yeast Saccharomyces cerevisiae. Microbiol Mol Rev 1998; 62:1264-1300. 3. Banuett F. Signalling in the yeast: an informational cascade with links to the filamentous fungi. Microbiol Mol Biol Rev 1998; 62:249-274. 4. Heinisch JJ, Lorberg A, Schmitz HP et al. The protein kinase C-mediated MAP kinase pathway involved in the maintenance of cellular integrity in Saccharomyces cerevisiae. Mol Microbiol 1999; 32:671-680. 5. Arellano M, Coll PM, Pérez P. Rho GTPases in the control of cell morphology, cell polarity, and actin localization in fission yeast. Microsc Res Tech 1999; 47:51-60. 6. Pruyne D, Bretscher A. Polarization of cell growth in yeast. I. Establishment and maintenance of polarity states. J Cell Sci 2000; 113:365-375. 7. Levin DE, Fields FO, Kunisawa R et al. A candidate protein kinase C gene, PKC1, is required for the S.cerevisiae cell cycle. Cell 1990; 62:213-224. 8. Levin DE, Bartlett-Heubusch E. Mutants in the S. cerevisiae PKC1 gene display a cell cycle-specific osmotic stability defect. J Cell Biol 1992; 116:1221-1229. 9. Paravicini G, Cooper M, Friedli L et al. The osmotic integrity of the yeast cell requires a functional PKC1 gene product. Mol Cell Biol 1992; 12:4896-4905. 10. Mellor H, Parker PJ. The extended protein kinase C superfamily. Biochem J 1998; 332:281-292. 11. Flynn P, Mellor H, Casamassima A et al. Rho GTPase control of protein kinase C-related protein kinase activation by 3-phosphoinositide-dependent protein kinase. J Biol Chem 2000; 275:11064-11070. 12. Antonsson B, Montessuit S, Friedli L et al. Protein kinase C in yeast. Characteristics of the Saccharomyces cerevisiae PKC1 gene product. J Biol Chem 1994; 269:16821-16828. 13. Jacoby JJ, Nilius SM, Heinisch JJ. A screen for upstream components of the yeast protein kinase C signal transduction pathway identifies the product of the SLG1 gene. Mol Gen Genet 1998; 258:148-155. 14. Toda T, Shimanuki M, Yanagida M. Two novel protein kinase C-related genes of fission yeast are essential for cell viability and implicated in cell shape control. EMBO J 1993; 12:1987-1995. 15. Mazzei GJ, Schmid EM, Knowles JK et al. A Ca2+-independent protein kinase C from fission yeast. J Biol Chem 1993; 268:7401-7406. 16. Arellano M, Valdivieso MH, Calonge TM et al. Schizosaccharomyces pombe protein kinase C homologues, pck1p and pck2p, are targets of rho1p and rho2p and differentially regulate cell integrity. J Cell Sci 1999; 112:3569-3578. 17. Shiozaki K, Russell P. Counteractive roles of protein phosphatase 2C (PP2C) and a MAP kinase kinase homolog in the osmoregulation of fission yeast. EMBO J 1995; 14:492-502. 18. Toda T, Dhut S, Superti-Furga G et al. The fission yeast pmk1 + gene encodes a novel mitogen-activated protein kinase homolog which regulates cell integrity and functions coordinately with the protein kinase C pathway. Mol Cell Biol 1996; 16:6752-6764. 19. Kobori H, Toda T, Yaguchi H et al. Fission yeast protein kinase C gene homologues are required for protoplast regeneration: a functional link between cell wall formation and cell shape control. J Cell Sci 1994; 107:1131-1136.
96
Protein Kinase C
20. Nonaka H, Tanaka K, Hirano H et al. A downstream target of RHO1 small GTP-binding protein is PKC1, a homolog of protein kinase C, which leads to activation of the MAP kinase cascade in Saccharomyces cerevisiae. EMBO J 1995; 14:5931-5938. 21. Kamada Y, Jung US, Piotrowski J et al. The protein kinase C-activated MAP kinase pathway of Saccharomyces cerevisiae mediates a novel aspect of the heat shock response. Genes Dev 1995; 9:1559-1571. 22. Andrews PD, Stark MJ. Dynamic, Rho1p-dependent localization of Pkc1p to sites of polarized growth. J Cell Sci 2000; 113:2685-2693. 23. Sayers LG, Katayama S, Nakano K et al. Rho-dependence of Schizosaccharomyces pombe Pck2. Genes Cells 2000; 5:17-27. 24. Arellano M, Duran A, Perez P. Localization of the Schizosaccharomyces pombe rho1 GTPase and its involvement in the organization of the actin cytoskeleton. J Cell Sci 1997; 110:2547-2555. 25. Nakano K, Arai R, Mabuchi I. The small GTP binding protein Rho1 is a multifunctional protein that regulates actin localization, cell polarity, and septum formation in the fission yeast Schizosaccharomyces pombe. Genes Cells 1997; 2:679-694. 26. Hirata D, Nakano K, Fukui M et al. Genes that cause aberrant cell morphology by overexpression in fission yeast: a role for a small GTP-binding protein Rho2 in cell morphogenesis. J Cell Sci 1998; 111:149-159. 27. Calonge TM, Nakano K, Arellano M et al. Schizosaccharomyces pombe Rho2 GTPase regulates the cell wall α-glucan biosynthesis, through the protein kinase Pck2p. Mol Biol Cell 2000; 11:4393-4401. 27a. Schmitz HP, Lorberg A, Heinisch JJ. Regulation of yeast protein kinase C activity by interaction with the small GTPase Rho1p through its amino-terminal HR1 domain. Mol Microbiol 2002; 44:829-840. 28. Cazaubon S, Bornancin F, Parker PJ. Threonine-497 is a critical site for permissive activation of protein kinase C alpha. Biochem J 1994; 301:443-448. 29. Chou MM, Hou W, Johnson J et al. Regulation of protein kinase C zeta by PI 3-kinase and PDK-1. Curr Biol 1998; 8:1069-1077. 30. Le Good J, Ziegler WH, Parekh DB et al. Protein kinase C isotypes controlled by phosphoinositide 3-kinase through the protein kinase PDK1. Science 1998; 281:2042-2045. 31. Toker A. The synthesis and cellular roles of phosphatidylinositol 4,5-biphosphate. Curr. Op. Cell Biol. 1998; 10:254-261. 32. Inagaki M, Schmelzle T, Yamaguchi K et al. PDK1 homologs activate the Pkc1-mitogen-activated protein kinase pathway in yeast. Mol Cell Biol 1999; 19:8344-8352. 33. Casamayor A, Torrance PD, Kobayashi T et al. Functional counterparts of mammalian protein kinases PDK1 and SGK in budding yeast. Curr Biol 1999; 9:186-197. 34. Niederberger C, Schweingruber ME. A Schizosaccharomyces pombe gene, ksg1, that shows structural homology to the human phosphoinositide-dependent protein kinase PDK1, is essential for growth, mating and sporulation. Mol Gen Genet 1999; 261:177-183. 35. Filippa N, Sable CL, Hemmings BA et al. Effect of phosphoinositide-dependent kinase 1 on protein kinase B translocation and its subsequent activation. Mol Cell Biol 2000; 20:5712-5721. 36. Stack JH, Emr SD. Vps34p required for yeast vacuolar protein sorting is a multiple specificity kinase that exhibits both protein kinase and phosphatidylinositol-specific PI 3-kinase activities. J Biol Chem 1994; 269:31552-31562. 37. Yoshida S, Ohya Y, Goebl M et al. A novel gene, STT4, encodes a phosphatidylinositol 4-kinase in the PKC1 protein kinase pathway of Saccharomyces cerevisiae. J Biol Chem 1994; 269:1166-1172. 38. Yoshida S, Ohya Y, Nakano A et al. Genetic interactions among genes involved in the STT4-PKC1 pathway of Saccharomyces cerevisiae. Mol Gen Genet 1994; 242:631-640. 39. Cutler NS, Heitman J, Cardenas ME. STT4 is an essential phosphatidylinositol 4-kinase that is a target of wortmannin in Saccharomyces cerevisiae. J Biol Chem 1997; 272:27671-27677. 39a. Audhya A, Emr SD. Stt4 PI 4-kinase localizes to the plasma membrane and functions in the Pkc1-mediated MAP kinase cascade. Dev Cell 2002; 2:593-605. 40. Helliwell SB, Schmidt A, Ohya Y et al. The Rho1 effector Pkc1, but not Bni1, mediates signalling from Tor2 to the actin cytoskeleton. Curr Biol 1998; 8:1211-1214. 41. Desrivieres S, Cooke FT, Parker PJ et al. MSS4, a phosphatidylinositol-4-phosphate 5-kinase required for organization of the actin cytoskeleton in Saccharomyces cerevisiae. J Biol Chem 1998; 273:15787-15793. 42. Homma K, Terui S, Minemura M et al. Phosphatidylinositol-4-phosphate 5-kinase localized on the plasma membrane is essential for yeast cell morphogenesis. J Biol Chem 1998; 273:15779-15786. 42a. Lorberg A, Jacoby JJ, Schmitz HP et al. The PH domain of the yeast GEF Rom2p serves an essential function in vivo. Mol Genet Genomics 2001; 266:505-513.
Protein Kinase C in Yeast: The Cell Wall Integrity Pathway
97
43. Drgonova J, Drgon T, Tanaka K et al. Rho1p, a yeast protein at the interface between cell polarization and morphogenesis. Science 1996; 272:277-279. 44. Qadota H, Python CP, Inoue SB et al. Identification of yeast Rho1p GTPase as a regulatory subunit of 1,3-β-glucan synthase. Science 1996; 272:279-281. 45. Mazur P, Baginsky W. In vitro activity of 1,3-β-D-glucan synthase requires the GTP-binding protein Rho1. J Biol Chem 1996; 271:14604-14609. 46. Ozaki K, Tanaka K, Imamura H et al. Rom1p and Rom2p are GDP/GTP exchange proteins (GEPs) for the Rho1p small GTP binding protein in Saccharomyces cerevisiae. EMBO J 1996; 15:2196-2207. 46a. Schmelzle T, Helliwell SB, Hall MN. Yeast protein kinases and the RHO1 exchange factor TUS1 are novel components of the cell integrity pathway in yeast. Mol Cell Biol 2002; 22:1329-1339. 47. Bender A, Pringle JR. Use of a screen for synthetic lethal and multicopy suppressee mutants to identify two new genes involved in morphogenesis in Saccharomyces cerevisiae. Mol Cell Biol 1991; 11:1295-1305. 48. Zheng Y, Hart MJ, Shinjo K et al. Biochemical comparisons of the Saccharomyces cerevisiae Bem2 and Bem3 proteins. J Biol Chem 1993; 269:24629-24634. 49. Schmidt A, Bickle M, Beck T et al. The yeast phosphatidylinositol kinase homolog TOR2 activates RHO1 and RHO2 via the exchange factor ROM2. Cell 1997; 88:531-542. 50. Watanabe D, Abe M, Ohya Y. Yeast Lrg1p acts as a specialized RhoGAP regulating 1,3-beta-glucan synthesis. Yeast 2001; 18:943-951. 50a. Schmidt A., Schmelzle T, Hall MN. The RHO1-GAPs SAC7, BEM2 and BAG7 control distinct RHO1 functions in Saccharomyces cerevisiae. Mol Microbiol 2002; 45:1433-1441. 51. Masuda T, Tanaka K, Nonaka Y et al. Molecular cloning and characterization of yeast rho GDP dissociation inhibitor. J Biol Chem 1994; 269:19713-19718. 52. Bickle M, Delley PA, Schmidt A et al. Cell wall integrity modulates Rho1 activity via the exchange factor ROM2. EMBO J 1998; 17:2235-2245. 53. Verna J, Lodder A, Lee K et al. A family of genes required for maintenance of cell wall integrity and for the stress response in Saccharomyces cerevisiae. Proc Natl Acad Sci USA 1997; 94:13804-13809. 53a. Zu T, Verna J, Ballester R. Mutations in WSC genes for putative stress receptors result in sensitivity to multiple stress conditions and impairment of Rlm1-dependent gene expression in Saccharomyces cerevisiae. Mol Genet Genomics 2001; 266:142-155. 54. Lodder AL, Lee TK, Ballester R. Characterization of the Wsc1 protein, a putative receptor in the stress response of Saccharomyces cerevisiae. Genetics 1999; 152:1487-1499. 55. Rajavel M, Philip B, Buehrer BM et al. Mid2 is a putative sensor for cell integrity signaling in Saccharomyces cerevisiae. Mol Cell Biol 1999; 19:3969-3976. 56. Ketela T, Green R, Bussey H. Saccharomyces cerevisiae mid2p is a potential cell wall stress sensor and upstream activator of the PKC1-MPK1 cell integrity pathway. J Bacteriol 1999; 181:3330-3340. 57. Philip B, Levin DE. Wsc1 and Mid2 are cell surface sensors for cell wall integrity signaling that act through Rom2, a guanine nucleotide exchange factor for Rho1. Mol Cell Biol 2001; 21:271-280. 58. Beck T, Hall MN. The TOR signalling pathway controls nuclear localization of nutrient-regulated transcription factors. Nature 1999; 402:689-692. 59. Kunz J, Hall MN. Cyclosporin A, FK506 and rapamycin: more than just immunosuppression. Trends Biochem Sci 1993; 18:334-338. 60. Helliwell SB, Howald I, Barbet N. TOR2 is part of two related signaling pathways coordinating cell growth in Saccharomyces cerevisiae. Genetics 1998; 148:99-112. 61. Costigan C, Gehrung S, Snyder M. A synthetic lethal screen identifies SLK1, a novel protein kinase homolog implicated in yeast cell morphogenesis and cell growth. Mol Cell Biol 1992; 12:1162-1178. 62. Martin H, Arroyo J, Sánchez M et al. Activity of the yeast MAP kinase homologue Slt2 is critically required for cell integrity at 37°C . Mol Gen Genet 1993; 241:177-184. 63. Irie K, Takase M, Lee KS et al. MKK1 and MKK2, which encode Saccharomyces cerevisiae mitogen-activated protein kinase-kinase homologs, function in the pathway mediated by protein kinase C. Mol Cell Biol 1993; 13(5):3076-83. 64. Martin H, Rodríguez-Pachon JM, Ruiz C et al. Regulatory mechanisms for modulation of signalling through the cell integrity. J Biol Chem 2000; 275:1511-1519. 65. Lee KS, Irie K, Gotoh Y et al. A yeast mitogen-activated protein kinase homolog (Mpk1p) mediates signalling by protein kinase C. Mol Cell Biol 1993; 13:3067-3075. 66. Lee KS, Levin DE. Dominant mutations in a gene encoding a putative protein kinase (BCK1) bypass the requirement for a Saccharomyces cerevisiae protein kinase C homolog. Mol Cell Biol 1992; 12:172-182.
98
Protein Kinase C
67. Torres L, Martin H, García-Sáez MI et al. A protein kinase gene complements the lytic phenotype of Saccharomyces cerevisiae lyt2 mutants. Mol Microbiol 1991; 5:2845-2854. 68. Sheu YJ, Santos B, Fortin N et al. Spa2p interacts with cell polarity proteins and signaling components involved in yeast cell morphogenesis. Mol Cell Biol 1998; 18(7):4053-69. 69. Zarzov P, Mazzoni C, Mann C. The SLT2 (MPK1) MAP kinase is activated during periods of polarized cell growth in yeast. EMBO J 1996; 15:83-91. 70. Davenport KR, Sohaskey M, Kamada Y et al. A second osmosensing signal transduction pathway in yeast. Hypotonic shock activates the PKC1 protein kinase-regulated cell integrity pathway. J Biol Chem 1995; 270:30157-30161. 71. Watanabe Y, Takaesu G, Hagiwara M et al. Characterization of a serum response factor-like protein in Saccharomyces cerevisiae, Rlm1, which has transcriptional activity regulated by the Mpk1 (Slt2) mitogen-activated protein kinase pathway. Mol Cell Biol 1997; 17: 2615-2623. 72. Jung US, Levin DE. Genome-wide analysis of gene expression regulated by the yeast cell wall integrity signalling pathway. Mol Microbiol 1999; 34:1049-1057. 73. Dodou E, Treisman R. The Saccharomyces cerevisiae MADS-box transcription factor Rlm1 is a target for the Mpk1 mitogen-activated protein kinase pathway. Mol Cell Biol 1997; 17:1848-1859. 74. Madden K, Sheu YJ, Baetz K et al. SBF cell cycle regulator as a target of the yeast PKC-MAP kinase pathway. Science 1997; 275:1781-1784. 75. Igual JC, Johnson AL, Johnston LH. Coordinated regulation of gene expression by the cell cycle transcription factor Swi4 and the protein kinase C MAP kinase pathway for yeast cell integrity. EMBO J 1996; 15(18):5001-5013. 76. Iyer VR, Horak CE, Scafe CS et al. Genomic binding sites of the yeast cell-cycle transcription factors SBF and MBF. Nature 2001; 409:533-538. 77. Delley PA, Hall MN. Cell wall stress depolarizes cell growth via hyperactivation of RHO1. J Cell Biol 1999; 147:163-174. 78. Nanduri J, Mitra S, Andrei C et al. An unexpected link between the secretory path and the organization of the nucleus. J Biol Chem 1999; 274:33785-33789. 79. Nanduri J, Tartakoff AM. Perturbation of the nucleus: A novel Hog1p-independent, Pkc1pdependent consequence of hypertonic shock in yeast. Mol Biol Cell 2001; 12:1835-1841. 80. Lee KS, Hines LK, Levin DE. A pair of functionally redundant yeast genes (PPZ1 and PPZ2) encoding type 1-related protein phosphatases function within the PKC1-mediated pathway. Mol Cell Biol 1993; 13:5843-5853. 81. Epstein CB, Cross FR. Genes that can bypass the CLN requirement for Saccharomyces cerevisiae cell cycle START. Mol Cell Biol 1994; 14:2041-2047. 82. Martin-Yken H, Dagkessamanskaia A, De Groot P et al. Saccharomyces cerevisiae YCRO17c/ CWH43 encodes a putative sensor/transporter protein upstream of the BCK2 branch of the PKC1-dependent cell wall integrity pathway. Yeast 2001; 18:827-840. 83. Park H, Lennarz WJ. Evidence for interaction of yeast protein kinase C with several subunits of oligosaccharyl transferase. Glycobiology 2000; 10:737-744. 84. Gray JV, Ogas JP, Kamada Y et al. A role for the Pkc1 MAP kinase pathway of Saccharomyces cerevisiae in bud emergence and identification of a putative upstream regulator. EMBO J 1997; 16:4924-4937. 85. Marini NJ, Meldrum E, Buehrer B et al. A pathway in the yeast cell division cycle linking protein kinase C (Pkc1) to activation of Cdc28 at START. EMBO J 1996; 15:3040-3052. 86. Wijnen H, Futcher B. Genetic analysis of the shared role of CLN3 and BCK2 at the G1-S transition in Saccharomyces cerevisiae. Genetics 1999; 153:1131-1143. 87. Sia RA, Bardes ES, Lew DJ. Control of Swe1p degradation by the morphogenesis checkpoint. EMBO J 1998; 17:6678-6688. 88. Harrison JC, Bardes ES, Ohya Y et al. A role for the Pkc1p/Mpk1p kinase cascade in the morphogenesis checkpoint. Nat Cell Biol 2001; 3:417-420. 89. Khalfan W, Ivanovska I, Rose MD. Functional interaction between the PKC1 pathway and CDC31 network of SPB duplication genes. Genetics 2000; 155:1543-1559. 90. Philips J, Herskowitz I. Osmotic balance regulates cell fusion during mating in Saccharomyces cerevisiae. J Cell Biol 1997; 138:961-974. 91. Garrett-Engele P, Moilanen B, Cyert MS. Calcineurin, the Ca2+/calmodulin-dependent protein phosphatase, is essential in yeast mutants with cell integrity defects and in mutants that lack a functional vacuolar H(+)-ATPase. Mol Cell Biol 1995; 15:4103-4114. 92. Mazur P, Morin N, Baginsky W et al. Differential expression and function of two homologous subunits of yeast (1,3)- β-D-glucan synthase. Mol Cell Biol 1995; 15:5671-5681. 92a. Krause SA., Gray JV. The protein kinase C pathway is required for viability in quiescence in Saccharomyces cerevisiae. Curr Biol. 2002; 12:588-593.
Protein Kinase C in Yeast: The Cell Wall Integrity Pathway
99
93. Weisman R, Choder M. The fission yeast TOR homolog, tor1+, is required for the response to starvation and other stresses via a conserved serine. J Biol Chem 2001; 276:7027-7032. 94. Sengar AS, Markley NA, Marini NJ et al. Mkh1, a MEK kinase required for cell wall integrity and proper response to osmotic and temperature stress in Schizosaccharomyces pombe. Mol Cell Biol 1997; 17:3508-3519. 95. Zaitsevskaya-Carter T, Cooper JA. Spm1, a stress-activated MAP kinase that regulates morphogenesis in S. pombe. EMBO J 1997; 16:1318-1331. 96. Sugiura R, Toda T, Dhut S et al. The MAPK kinase Pek1 acts as a phosphorylation-dependent molecular switch. Nature 1999; 399:479-483. 97. Sells MA, Barratt JT, Caviston J et al. Characterization of Pak2p, a pleckstrin homology domain-containing, p21-activated protein kinase from fission yeast. J Biol Chem 1998; 273:18490-18498. 98. Yang P, Kansra S, Pimental RA et al. Cloning and characterization of shk2, a gene encoding a novel p21-activated protein kinase from fission yeast. J Biol Chem 1998; 273:18481-18489. 99. Merla A, Johnson DI. The Schizosaccharomyces pombe Cdc42p GTPase signals through Pak2p and the Mkh1p-Pek1p-Spm1p MAPK pathway . Curr Genet 2001; 39:205-209. 100. Katayama S, Hirata D, Arellano M et al. Fission yeast α-glucan synthase Mok1 requires the actin cytoskeleton to localize the sites of growth and plays an essential role in cell morphogenesis downstream of protein kinase C function. J Cell Biol 1999; 144:1173-1186. 101. Shimanuki M, Kinoshita N, Ohkura H et al. Isolation and characterization of the fission yeast protein phosphatase gene ppe1+ involved in cell shape control and mitosis. Mol Biol Cell 1993; 4:303-313. 102. Carnero E, Ribas JC, Garcia B et al. Schizosaccharomyces pombe ehs1p is involved in maintaining cell wall integrity and in calcium uptake. Mol Gen Genet 2000; 264:173-183. 103. Fischer M, Schnell N, Chattaway J et al. The Saccharomyces cerevisiae CCH1 gene is involved in calcium influx and mating. FEBS Lett 1997; 419:259-262. 104. Paravicini G, Mendoza A, Antonsson B et al. The Candida albicans PKC1 gene encodes a protein kinase C homolog necessary for cellular integrity but not dimorphism. Yeast 1996; 12:741-756. 105. Morawetz R, Lendenfeld T, Mischak H et al. Cloning and characterisation of genes (pkc1 and pkcA) encoding protein kinase C homologues from Trichoderma reesei and Aspergillus niger. Mol Gen Genet 1996; 250:17-28.
100
Protein Kinase C
CHAPTER 9
Specificity in Atypical Protein Kinase C Signaling: The NF-κκB Paradigm Jorge Moscat and María T. Diaz-Meco
Abstract
T
he potential relevance of Protein Kinase C (PKC) to the control of Nuclear Factor (NF)-κB activation has been the matter of numerous studies. The first indication was provided by the fact that phorbol esters are potent activators of this pathway, at least in some cell systems. More recently, the use of dominant negative mutants and knock out mice of several PKC isoforms and the identification of novel protein adapters and regulators of the atypical PKCs (aPKCs), have provided convincing evidence and mechanistic details of how these kinases may control not only this important signaling cascade but also other pathways essential for cell function.
Introduction Since their discovery, PKC isoforms have been implicated in a large number of different cell functions.1,2 At the same time, the mechanisms of activation of the transcription factor NF-κB have been the focus of great interest. NF-κB plays critical roles in cell growth and apoptosis and consequently in immune responses, development and in human diseases such as cancer and inflammation.3-5 Very early studies implicated some PKC isoforms in NF-κB activation in vitro,6 consistent with the fact that PMA is a relatively potent NF-κB activator in several cell systems including T lymphocytes. However, downregulation of the PMA-sensitive PKCs did not inhibit the activation of NF-κB by physiological stimuli such as the inflammatory cytokines Tumor Necrosis Factor (TNF)-α and Interleukin (IL)-1.7 Thus, if a PKC was critically implicated in this pathway, it would be a PMA-insensitive isoform. The atypical PKC isoforms are widely recognized to be insensitive to PMA or diacylglycerol because their regulatory domains only have one zinc finger whilst classical and the novel isoforms have two zinc-fingers (Fig. 1; see also Chapter 1).2 The catalytic domain of all the PKC isotypes displays a very high homology suggesting that different PKC isotypes may channel distinct upstream signals to common downstream targets. The specificity will then be the responsibility of the regulatory part of the molecule whereas the catalytic function would be shared by all the isoforms. How the aPKCs impinge on the NF-κB pathway and their relationship with the other PKC isotypes is the main focus of this article.
κB Pathway The NF-κ Numerous and excellent reviews have dealt with the details of this important signaling cascade.5,8 A brief overview of this pathway (Fig. 2) unveils the complexity of a network that is activated by numerous challenges and that regulates diverse and important cellular functions. Protein Kinase C, 2nd Edition, edited by Lodewijk Dekker. ©2004 Eurekah.com and Kluwer Academic / Plenum Publishers.
Specificity in Atypical Protein Kinase C Signaling: The NF-κB Paradigm
101
Figure 1. Structural differences in the PKC isoforms. cPKC, classical PKCs; nPKC, novel PKCs; aPKC, atypical PKCs. ZF, zinc-finger.
From a signal transduction point of view the most interesting pathways are the relatively well characterized TNF-α and IL-1 cascades. NF-κB is composed of dimers of different members of the Rel protein family.8-10 The most classical form of NF-κB is a heterodimer of p50 and p65 (Rel A), which is kept in the cytosol by Inhibitor of NF-κB (IκB), thus preventing its nuclear translocation and activity. Upon cell stimulation by inflammatory cytokines, IκBα is phosphorylated on residues 32 and 36 by the IκB kinase (IKK) complex, thereby triggering the ubiquitination and subsequent degradation of IκB through the proteasome pathway.4,5 The IKK complex consists of two catalytic subunits (IKKα and IKKβ) and an adapter protein termed IKKγ or Nemo. Evidence from knock out mice has revealed the role of each IKK isoform in NF-κB activation. Genetic inactivation of IKKα only slightly affects the stimulation of NF-κB and that of the IKK complex in response to physiological stimuli, indicating that IKKα may not be an important player in this pathway.11 However, at least in some cell systems, IKKα may play a role in NF-κB activation.12,13 The deletion of IKKβ or IKKγ completely ablates IKK and NF-κB activities.14-16 The precise mechanism whereby the IKKγ-IKKβ complex responds to the upstream signals leading to IκB phosphorylation is presently unknown. Oligomerization in response to its interaction with adapter molecules may trigger the autoactivation of IKKβ. For example recent studies demonstrate that the simple dimerization of Receptor Interacting Protein (RIP) in cotransfection experiments triggers the activation of IKK, possibly by autophosphorylation.17,18 Upstream of IKK activation, protein-protein interactions are critical events in the signaling cascades activated by TNFα. The main transducer of TNFα actions is the TNF Receptor-1 (TNFR-1), a 55 kDa protein with a death domain (DD) in its intracellular region.19 Upon cell stimulation with TNFα, TNFR1 recruits, through a homotypic interaction, the DD-containing protein TNFR1-associated-death-domain (TRADD) which itself binds TNFR-associated-factor-2 (TRAF2) and RIP.19 TRAF2 interacts with the intermediary domain of RIP, giving rise to an interconnected trimolecular complex involving TRADD, RIP, and TRAF2 (Fig. 3A). Interestingly, although the overexpression of TRAF2 or RIP is sufficient to activate NF-κB and Jun N-terminal kinase (JNK)/Stress activated protein kinase (SAPK),20 only RIP-/-,21 but not TRAF2-deficient cells,22 have impaired NF-κB activation in response to
102
Protein Kinase C
Figure 2. The IKK complex is channels multiple signals that activate NF-κB.
TNF-α. This indicates that RIP is necessary and sufficient for NF-κB signaling by TNF-α. The evidence that RIP binds IKKγ in cotransfection experiments, which seems sufficient under these conditions for IKK activation, would suggest that no other component needs to be invoked to account for the activation of IKK.23 In other words, a chain of interactions and oligomerizations from the TNFR to TRADD-RIP-IKK would be, in this model, sufficient to activate IKK and NF-κB. However, although NF-κB activators, like the Tax protein may trigger the pathway through a protein-protein interaction,4 the physiological evidence of the RIP-IKKγ connection is not so clear yet. For example, Wallach and coworkers have demonstrated that the IKK signalsome can be recruited to the TNFR complex, most likely through the interaction of IKKγ with RIP.23 This recruitment is insufficient for IKK activation, suggesting that other molecules are required for the stimulation of the IKK complex once bound to RIP. In this regard, a series of interesting studies performed in cells lacking either TRAF2 or RIP by genetic inactivation suggest the existence of a two-signal model for IKK activation.24 Thus, Devin and coworkers showed that whilst the recruitment of IKK to the TNFR complex requires TRAF2 but not RIP, the activation of IKK requires both.24 These results should be seen in the context of other evidence questioning the importance of TRAF2 for IKK and NF-κB activation at least in certain cell systems.22 Redundancy with other TRAFs, perhaps TRAF5,25,26 may explain these results. In addition, it is not clear yet how TRAF2 recruits the IKK signalsome. The important corollary of these experiments is that under physiological conditions RIP may be responsible of activating IKK either directly or by recruiting another IKK-activating protein. IKKβ displays a sequence in its activation loop that suggests that it may be phosphorylated by kinases of the type of MAP and ERK kinase kinases (MEKK).27-31 The initial characterizations of the IKK signalsome complex did in fact indicate that it was activated by MEKK1 in vitro.32 In addition, mutations of the residues in the activation loop to alanines completely abrogated IKK activity whereas mutations to acidic amino acids lead to a permanently active
Specificity in Atypical Protein Kinase C Signaling: The NF-κB Paradigm
103
Figure 3. The adapter p62 and NF-κB activation. The interaction of p62 with RIP (A) and TRAF6 (B) serves to locate the aPKCs in the NF-κB signaling cascades activated by TNF-α and IL-1, respectively.
kinase. Although genetic evidence seems to rule out MEKK1 as a physiologically relevant IKK activator, other related kinases may be important in this pathway in vivo.33,34 For example, the over-expression of MEKK3 potently activates IKKβ35 and its inactivation in knock out embryonic fibroblasts (EFs) impairs at least partially IKK and NF-κB stimulation.36 In this regard, we and others have recently reported that IKK can be stimulated by PMA. The stimulation is completely abrogated by a pharmacological inhibitor of the classical PKCs at concentrations that do not affect other isoforms, suggesting that, as for NF-κB, PKC is able to positively impinge on this pathway.37 The activation of IKK by TNF-α is not affected by the inhibitor of the conventional PKCs, suggesting a role of an atypical isoform in this pathway.37 In agreement with this, transfection of dominant negative mutants of the aPKCs severely inhibited TNF-α-induced IKK activity with no effect on the PMA actions, suggesting a model whereby PMA uses a classical or novel PKC isoform whereas TNF-α employs an aPKC to activate IKK37 (Fig. 4). Interestingly, evidence has been presented that recombinant IKKβ can be activated and phosphorylated by recombinant PKC isotypes in vitro.37 Recent data demonstrate that the lack of PKC-θ severely impairs NF-κB activation in mature but not in immature T lymphocytes38 (see also Chapter 12). Therefore, different PKC isotypes may be critically involved in NF-κB signaling in a cell-type specific manner.
κB Activation Adapters for the aPKCs to NF-κ An important theme in cell signaling is how specificity is achieved for apparently promiscuous kinases. There are numerous examples of this in the literature but the aPKCs are a good paradigm.39 Thus, PKC-ζ as well as PKC-λ/ι have been implicated not only in NF-κB activation but also in distinct signaling pathways such as mitogen-activated kinase (MAPK),
104
Protein Kinase C
Figure 4. Different PKCs target the IKK complex.
Extracellular signal-regulated kinase (ERK)-5, Glucose Transporter (GLUT)-4, p70S6-Kinase, cytoskeletal rearrangement, long-term potentiation, or the respiratory burst in neutrophils.39 Since much of this evidence was obtained in overexpression experiments, it could be argued that this apparent promiscuity is due to artificial activation of pathways that are not targeted under physiological conditions by this PKC isoform. Another possibility is that one sole kinase may be involved in several pathways but that partners are required to confer specificity. The seminal work of Mochly-Rosen and coworkers lead to the identification of proteins (termed RACKs) that serve to localize some classical and novel PKC isotypes to a particular place in the cell which makes the enzyme’s effects more selective and efficient.40 Two-hybrid screens have identified p62 as a novel aPKC-interacting protein.41 Downregulation of p62 using an anti-sense plasmid leads to a dramatic inhibition of NF-κB activation in TNFα-treated cells42,43 with no effects on the activation of JNK or ERK. Thus the p62-aPKC complex is essential for activation of NF-κB and p62 may be an adapter of the aPKCs in this pathway. How does p62 help to locate the aPKCs in this signaling cascade? A possible explanation comes from recent observation that p62 selectively interacts with RIP.43 In addition to a DD domain, which mediates binding to TRADD, RIP has an intermediary and a kinase domain. The kinase activity of RIP is dispensable for its function as an NF-κB-activating molecule,19 however the intermediary domain is sufficient to drive NF-κB activation.19 p62 binds directly to the intermediary domain of RIP. Therefore, a model (Fig. 3A) can be proposed in which interaction of p62 with RIP serves to juxtapose the IKK complex, itself recruited to the TNFR complex by TRAF2 or a TRAF2-like molecule, to the aPKCs. These could efficiently activate the IKKβ subunit of the signalsome complex, possibly by phoshorylating it. The aPKCs are not only activated by TNFα but also by IL-1.44 The question is how a single kinase can be connected to two different upstream signaling pathways. Again p62 may offer a plausible explanation, since there is an outstanding parallel between the TNFα and IL-1
Specificity in Atypical Protein Kinase C Signaling: The NF-κB Paradigm
105
Figure 5. The interaction of p62 with RIP, TRAF6 and the aPKCs involves different motifs.
signaling cascades. Thus the IL-1 receptor, together with the receptor accessory protein, interacts with MyD88, a functional analogue of TRADD, which recruits IL-1 receptor associated kinase (IRAK), whose enzymatic activity, like that of RIP, is dispensable for the activation of NF-κB (Fig. 3B). IRAK plays a critical role in NF-κB activation through its interaction with TRAF6.45-48 Both IRAK and TRAF6 are required intermediates for IL-1 signaling since genetic deletion of IRAK or TRAF6 dramatically impairs NF-κB activation in response to that cytokine.45,47 Interestingly, there is evidence that p62 directly interacts not only with RIP but also with TRAF6 in a ligand-dependent manner.42 These contacts take place through different p62 domains (Fig. 5). The interaction with RIP involves the ZZ domain of p62 and the coil-coiled intermediary region of RIP whereas the interaction with TRAF6 implicates a relatively short sequence of amino acids. This sequence is absent in a second p62 isoform recently discovered to be induced during neuronal differentiation.42,49 Therefore, p62 is a scaffold that can accommodate simultaneously the aPKCs, RIP, and TRAF6 and as such acts as a point of convergence for different NF-κB signaling pathways.
106
Protein Kinase C
κB Crossroad Functional Relevance of the aPKCs in the Ras-NF-κ The first evidence for the importance of the aPKCs to proliferative pathways was obtained in studies in which mouse fibroblasts or Xenopus leavis oocytes were micro-injected with pseudosubstrate peptide inhibitors relatively specific for each PKC isoform.50-52 In these experiments it became apparent that blockade of the aPKC activity impaired mitogenic signaling. These studies were complemented with other approaches using antisense oligonucleotides or kinase-inactive dominant negative mutants.52-54 Altogether, these observations suggested that the aPKCs could be critical for mitogenic activation. The mechanisms whereby aPKCs could be involved in these cell functions was initially linked to the Ras signaling pathways. Because dominant negative mutants of the aPKCs blocked oocyte maturation and cell transformation induced by oncogenic Ras,50,54 the possibility existed that Ras was able to communicate with these PKCs. In support of this model several studies identified aPKC in Ras protein complexes in activated cells.55-57 Although there is some homology between the regulatory domain of the aPKCs and that of the Raf proteins, which are well-established Ras-interacting partners,58 it is not clear yet if a direct physical interaction between Ras and the aPKCs could explain the requirement of these PKCs during Ras signaling.55 Oncogenic Ras activates PI 3-kinase,59,60 which through phosphoinositide-dependent kinase (PDK)-1, modulates the phosphorylation state of a critical residue in the aPKC activation loop.61 It is therefore possible that Ras recruits the aPKCs and PI 3-kinase to the same complex and that this is the crucial step in their activation process. Downstream of Ras lies the MEK/ERK cassette.58,62 MEK is phosphorylated by Raf which serves to phosphorylate and activate ERK which in turn translocates to the nucleus and activates the transcriptional machinery required for cell cycle entry.58,62 Several studies demonstrated that the aPKCs are able to activate MEK and ERK in vivo and in vitro.54,63,64 Therefore, downstream of Ras there is a bifurcation of signals toward Raf and the aPKCs which converge again at the level of MEK. Whether the aPKCs regulate MEK directly or indirectly is not fully resolved. The MEK/ERK cascade is the main transducer of the Ras mitogenic actions. However, Ras also activates the CDC42 and Rac proteins, small GTPases that play a critical role in the reorganization of the actin cytoskeleton, a key feature of the morphological changes associated with the induction of the transforming phenotype.58,65 Interestingly, the aPKCs are required for actin remodeling by Ras downstream of either Rac or CDC42.66 Therefore, the aPKCs appear to be important Ras downstream targets in at least two pathways: MEK and cytoskeletal remodeling. This may explain why the aPKCs are necessary for Ras-induced transformation. Based on this evidence one may predict that pharmacological inhibitors of the aPKCs in combination with molecules that could target Raf could be regarded as potential new effective anti-cancer drugs. Particularly relevant for this review is the fact that Ras also activates NF-κB, which is important for cell survival and for Ras-induced transformation.67-69 Thus, when NF-κB is ablated by transfection of an IκB dominant suppressor molecule, Ras transformed cells are more sensitive to the action of pro-apoptotic agents.69 Interestingly, the aPKCs have been reported to be key Ras players in NF-κB activation by Ras.54 Therefore, the aPKCs can be targeted in the NF-κB pathway by Ras as well as the intermediaries of the inflammatory signaling cascade TRAF6 and RIP. However, there is an important difference between Ras and the inflammatory pathways: whereas the latter use the adapter protein p62, Ras may directly target the aPKCs.
Specificity during Cell Signaling Is p62 the only adapter for the aPKCs? Clearly not. Careful mapping of the region where the aPKCs dock on the p62 molecule reveals that a short stretch of amino acids is required and sufficient for the interaction. This suggests that proteins containing a similar amino acid sequence could potentially represent aPKC-interacting partners (Fig. 6). The region, termed atypical PKC-interacting domain (AID), is part of a broader consensus sequence termed octicosapeptide which is present in many proteins but has as of yet no known role.39,43 A
Specificity in Atypical Protein Kinase C Signaling: The NF-κB Paradigm
107
Figure 6. Sequence of the AID motif in different aPKC adapters.
detailed mutagenesis study reveals that only some of the proposed octicosapeptide sequences are AIDs. We have recently identified several other AID-containing proteins. One is the aPKC-interacting scaffold Par-6, a relatively small molecule of 346 amino acids that contains a CDC42 and Rac interacting binding motif (CRIB)-like domain, a short sequence motif representing sites of interaction with the activated forms of CDC42 and Rac.70-73 In addition, C-terminally of the CRIB site in Par-6 lies a PDZ domain, which has been implicated in protein-protein interactions. Deletion of the AID site in Par-6 completely abrogates its interaction with the aPKCs.74 Par-6 binds Par-3 whose mammalian homologue, termed atypical PKC-isotype specific interacting protein (ASIP), had been reported to interact with the catalytic domain of the aPKCs and to be a relatively good substrate for these.75 The C. elegans homologue of the aPKCs as well as Par-3 and Par-6 have been shown to be required for the proper control of embryonic polarity.76,77 In mammalian cells, ASIP/Par-3 and the aPKCs co-localize at cell-cell contacts in epithelial cells, apparently regulating the establishment/maintenance of epithelial cell polarity.75 Since CDC42 interacts with Par-6, and a quaternary complex can be isolated involving CDC42, Par-6, Par-3 and the aPKCs, it is possible that the aPKC interaction with the AID site of Par-6 accounts for the involvement of the aPKCs in Rac signaling. Of potential interest, Rac has been shown to be important for NF-κB activation.78 Whether Par-6 may play a role in connecting the aPKCs and Rac to NF-κB signaling remains to be solved. In addition to the role of Par-6 as a potential scaffold, there is some evidence that CDC42 and Par-6 may play a role in the activation of the aPKCs.66,71 Therefore, the existence of a Cdc42/Rac-Par-6-aPKC axis for NF-κB activation would not be totally surprising. In any event, there is a remarkable parallelism between p62 and Par-6. Thus, p62 brings together RIP and TRAF6 (upstream components of the cytokine signaling pathways) with the aPKCs, which most likely serves to transmit the signal downstream toward the IKK complex.42,43 Par-6 can be the adapter receiving the upstream signals coming down from CDC42 to allow the aPKCs to control the actin cytoskeletal structure perhaps through Par3. Recently, our laboratory has identified the α isoform of the kinase MEK5 as another AID-containing molecule.74 MEK5 is the regulator of the kinase BMK1/ERK5, the downstream target of which is MEF2C, a transcription factor that in turn controls the expression of c-Jun which along with other immediate early genes is important for cell growth.79,80 Therefore, the activation of this pathway is essential for cell proliferation and oncogenesis. We have shown that the aPKCs interact with MEK5 in a mitogen-inducible manner and that this interaction involves the AID of MEK5 (Fig. 7). We also showed that this interaction is important for MEK5 activation but that, in contrast to the activation of MEK1/2, the enzymatic activity of the aPKCs is not required for that function.74 In this regard, the aPKCs join the increasingly number of kinases that are implicated in signal transduction in a manner that does not necessarily require their enzymatic activity, other examples in the NF-κB pathway being RIP, IRAK, and the double-stranded RNA-activated protein kinase (PKR) .
108
Protein Kinase C
Figure 7. The V1-AID interaction serves to locate the aPKCs in the MEK5-ERK5 mitogenic signaling cascade.
A fourth AID-containing protein that has been reported to interact with the V1 domain of the aPKCs is the mammalian homologue of C. elegans UNC-76, fasciculation and elongation protein zeta (FEZ)-1.81 FEZ-1 is a 393 amino acid protein mainly expressed in brain, with no apparent features suggestive of a scaffold role. However, a careful inspection of its sequence reveals the existence of a potential AID region, albeit with lower homology than the corresponding sites in MEK5 or Par-6.81 This may explain why FEZ-1 is not absolutely specific for the aPKCs and also shows interaction with PKC-ε.81 In contrast to p62 or Par-6, FEZ-1 is a relatively good substrate for the aPKCs. Functional analysis showed that overexpression of FEZ-1 along with an active mutant of PKC-ζ stimulates neuronal PC-12 cells and that FEZ-1 complements the function of UNC-76 which is necessary for normal axonal bundling and elongation in C. elegans.81 Whether the AID-like site of FEZ-1 mediates the interaction with the aPKCs and whether other proteins are recruited to the FEZ-1-aPKC complex remains to be clarified.
Specificity in Atypical Protein Kinase C Signaling: The NF-κB Paradigm
109
In summary, the aPKCs interact through a clearly defined sequence to different scaffolds and downstream targets, which serves to confer specificity to an otherwise promiscuous kinase, thereby avoiding undesired cross-talks. In this way a single enzyme would serve different functional purposes depending on the binding partners. Thus the cell has devised mechanisms to impose specificity on the PKCs, to control potential promiscuity of their highly homologous kinase domains. The regulation of PKC isotype specificity by protein interactions is discussed further in the chapter by Jaken.
κB by the Endogenous aPKC Inhibitor Par-4 Blockade of NF-κ The AID-containing adapters interact with the V1 domain of the aPKCs, the region most dissimilar to the other PKC isotypes. However, aPKCs also have a zinc-finger which has been shown in the novel and classical isotypes to be the target of the lipid cofactors that influence their activity. Prostate androgen responsive (Par)-4 is a potent inhibitor of the aPKCs and its overexpression severely impairs NF-κB activation by blocking the aPKC-IKK axis.82 Par-4 had previously been identified in a differential screening of genes induced in prostate cancer cells undergoing apoptosis after androgen withdrawal.83 Subsequent studies have demonstrated that Par-4 is induced in vulnerable neurons of Alzheimer patients and in neuronal cultures forced into apoptosis by withdrawal of neurotrophic factor.84 Consistent with Par-4 inducing apoptosis by blocking NF-κB, the ectopic expression of Par-4 in cell lines normally resistant to the pro-apoptotic actions of TNF-α, makes these cells susceptible to programmed cell death.82 Therefore, the inhibition of NF-κB emerges as the main mechanism of control for apoptosis either through Par-4 or by inhibition of the aPKCs.85 Interestingly, Par-4 levels are not only positively regulated but can also be downregulated. Ras transformation promotes a dramatic depletion of Par-4 protein and RNA levels.86 This has important physiological consequences because, in agreement with the notion that the aPKCs are downstream targets of Ras, the overexpression of Par-4 into these cells severely inhibits Ras-induced NF-κB.82 Perhaps more importantly, restoration of Par-4 to parental levels in the Ras transformants makes these cells more sensitive to pro-apoptotic insults including the action of chemotherapeutic agents such as the topoisomerase inhibitor camptothecin.86 Strikingly, experiments in vivo demonstrate that the progression of tumors derived from Par-4 expressing Ras-transformed cells is much more effectively reduced than those from Ras cells in which Par-4 levels are depleted.86 This indicates that the expression of Par-4 or inhibitors of the aPKCs could be useful sensitizers in cancer therapy. This is reminiscent of experiments in which the blockade of NF-κB in Ras transformed cells makes these more sensitive to the action of camptothecin.
Concluding Remarks How a single kinase can be involved in more than one pathway is one of the most interesting aspects of the studies on cell signaling. The aPKCs are a good paradigm for this. The identification of the AID sequence in at least three proteins involved in pathways in which the aPKCs appear to be implicated, provides a rationale for the understanding of the mechanisms through which specificity is achieved in aPKC signalling. Of particular relevance is p62, depletion of which with antisense strategies leads to the inhibition of NF-κB. Interestingly, the long suspected involvement of PKCs in the control of this transcription factor has received recent support with the data from the knock out of PKC-θ38 and PKC-ζ.87 The lesson learnt from those mice is that the involvement of the PKCs in this pathway is most likely cell-type specific. In the case of PKC-θ it is not clear yet whether it is an IKK kinase or if it acts at a distinct step in the NF-κB pathway. In any event, it seems that the role of this PKC isoform is restricted to mature T lymphocytes and appears also to be involved in the regulation of the AP-1 transcription factor through a still poorly understood mechanism. The knock out of PKC-ζ indicates that this PKC isoform is critically involved in the control of NF-κB at two different steps, depending on the cell type. Thus, in lung, a tissue in which PKC-ζ is particularly abundant, its absence severely impairs IKK activation,87 in keeping with the evidence from over-expression
110
Protein Kinase C
experiments that PKC-ζ can be an IKK kinase.37 By contrast, in EFs PKC-ζ is dispensable for IKK activation in response to TNF-α and IL-1 although it is still required for the transcriptional activity of NF-κB, through a mechanism that most likely involves the direct phosphorylation of p65.87 Future in vivo studies will clarify the physiological implications of inactivating aPKCs and their adapters in the whole animal. This together with a detailed understanding of the interaction of the aPKC V1 domains with the respective AID sequences will help in the design of new therapeutic strategies aimed at blocking NF-κB functions.
References 1. Nishizuka Y. Protein kinase C and lipid signaling for sustained cellular responses. FASEB J 1995; 9:484-496. 2. Hurley JH, Newton AC, Parker PJ et al. Taxonomy and function of C1 protein kinase C homology domains. Protein Sci 1997; 6:477-480. 3. Didonato JA, Saatcioglu F, Karin M. Molecular mechanisms of immunosuppression and anti-inflammatory activities by glucocorticoids. Am J Respir Crit Care Med 1996; 154:S11-15 4. Karin M. The beginning of the end: IkappaB kinase (IKK) and NF-kappaB activation. J Biol Chem 1999; 274:27339-27342. 5. Karin M, Ben-Neriah Y. Phosphorylation meets ubiquitination: The control of NF-[kappa]B activity. Annu Rev Immunol 2000; 18:621-663. 6. Ghosh S, Baltimore D. Activation in vitro of NF-kappa B by phosphorylation of its inhibitor I kappa B. Nature 1990; 344:678-682. 7. Diaz-Meco MT, Berra E, Municio MM et al. A dominant negative protein kinase C zeta subspecies blocks NF-kappa B activation. Mol Cell Biol 1993; 13:4770-4775. 8. Lenardo MJ, Baltimore D. NF-kappa B: a pleiotropic mediator of inducible and tissue-specific gene control. Cell 1989; 58:227-229. 9. Baldwin AS Jr. The NF-kappa B and I kappa B proteins: New discoveries and insights. Annu Rev Immunol 1996; 14:649-683. 10. Thanos D, Maniatis T. NF-kappa B: A lesson in family values. Cell 1995; 80:529-532. 11. Hu Y, Baud V, Delhase M et al. Abnormal morphogenesis but intact IKK activation in mice lacking the IKKalpha subunit of IkappaB kinase [see comments]. Science 1999; 284:316-320. 12. Li Q, Estepa G, Memet S et al. Complete lack of NF-kappaB activity in IKK1 and IKK2 double-deficient mice: additional defect in neurulation. Genes Dev 2000; 14:1729-1733. 13. Kaisho T, Takeda K, Tsujimura T et al. IkappaB kinase alpha is essential for mature B cell development and function. J Exp Med 2001; 193:417-426. 14. Li ZW, Chu W, Hu Y et al. The IKKbeta subunit of IkappaB kinase (IKK) is essential for nuclear factor kappaB activation and prevention of apoptosis. J Exp Med 1999; 189:1839-1845. 15. Li Q, Van Antwerp D, Mercurio F et al. Severe liver degeneration in mice lacking the IkappaB kinase 2 gene. Science 1999; 284:321-325. 16. Rudolph D, Yeh WC, Wakeham A et al. Severe liver degeneration and lack of NF-kappaB activation in NEMO/IKKgamma-deficient mice. Genes Dev 2000; 14:854-862. 17. Poyet JL, Srinivasula SM, Lin JH et al. Activation of the Ikappa B kinases by RIP via IKKgamma /NEMO-mediated oligomerization. J Biol Chem 2000; 275:37966-37977. 18. Inohara N, Koseki T, Lin J et al. An induced proximity model for NF-kappa B activation in the Nod1/RICK and RIP signaling pathways. J Biol Chem 2000; 275:27823-27831. 19. Hsu H, Huang J, Shu HB et al. TNF-dependent recruitment of the protein kinase RIP to the TNF receptor- 1 signaling complex. Immunity 1996; 4:387-396. 20. Liu ZG, Hsu H, Goeddel DV et al. Dissection of TNF receptor 1 effector functions: JNK activation is not linked to apoptosis while NF-kappaB activation prevents cell death. Cell 1996; 87:565-576 21. Kelliher MA, Grimm S, Ishida Y et al. The death domain kinase RIP mediates the TNF-induced NF-kappaB signal. Immunity 1998; 8:297-303. 22. Yeh WC, Shahinian A, Speiser D et al. Early lethality, functional NF-kappaB activation, and increased sensitivity to TNF-induced cell death in TRAF2-deficient mice. Immunity 1997; 7:715-725. 23. Zhang SQ, Kovalenko A, Cantarella G et al. Recruitment of the IKK signalosome to the p55 TNF receptor: RIP and A20 bind to NEMO (IKKgamma) upon receptor stimulation. Immunity 2000; 12:301-311. 24. Devin A, Cook A, Lin Y et al. The distinct roles of TRAF2 and RIP in IKK activation by TNF-R1: TRAF2 recruits IKK to TNF-R1 while RIP mediates IKK activation. Immunity 2000; 12:419-429.
Specificity in Atypical Protein Kinase C Signaling: The NF-κB Paradigm
111
25. Nakano H, Oshima H, Chung W et al. TRAF5, an activator of NF-kappaB and putative signal transducer for the lymphotoxin-beta receptor. J Biol Chem 1996; 271:14661-14664. 26. Ishida TK, Tojo T, Aoki T et al. TRAF5, a novel tumor necrosis factor receptor-associated factor family protein, mediates CD40 signaling. Proc Natl Acad Sci USA 1996; 93:9437-9442. 27. DiDonato JA, Hayakawa M, Rothwarf DM et al. A cytokine-responsive IkappaB kinase that activates the transcription factor NF-kappaB [see comments]. Nature 1997; 388:548-554. 28. Israel A. The IKK complex: an integrator of all signals that activate NF-kappaB? Trends Cell Biol 2000; 10:129-133. 29. Mercurio F, Zhu H, Murray BW et al. IKK-1 and IKK-2: Cytokine-activated IkappaB kinases essential for NF- kappaB activation. Science 1997; 278:860-866. 30. Ninomiya-Tsuji J, Kishimoto K, Hiyama A et al. The kinase TAK1 can activate the NIK-I kappaB as well as the MAP kinase cascade in the IL-1 signalling pathway. Nature 1999; 398:252-256. 31. Woronicz JD, Gao X, Cao Z et al. IkappaB kinase-beta: NF-kappaB activation and complex formation with IkappaB kinase-alpha and NIK. Science 1997; 278:866-869. 32. Lee FS, Peters RT, Dang LC et al. MEKK1 activates both IkappaB kinase alpha and IkappaB kinase beta. Proc Natl Acad Sci USA 1998; 95:9319-9324. 33. Xia Y, Makris C, Su B et al. MEK kinase 1 is critically required for c-Jun N-terminal kinase activation by proinflammatory stimuli and growth factor-induced cell migration. Proc Natl Acad Sci USA 2000; 97:5243-5248. 34. Yujiri T, Ware M, Widmann C et al. MEK kinase 1 gene disruption alters cell migration and c-Jun NH2- terminal kinase regulation but does not cause a measurable defect in NF-kappa B activation. Proc Natl Acad Sci USA 2000; 97:7272-7277. 35. Zhao Q, Lee FS. Mitogen-activated protein kinase/ERK kinase kinases 2 and 3 activate nuclear factor-kappaB through IkappaB kinase-alpha and IkappaB kinase-beta. J Biol Chem 1999; 274:8355-8358. 36. Yang J, Lin Y, Guo Z et al. The essential role of MEKK3 in TNF-induced NF-kappaB activation. Nat Immunol 2001; 2:620-624. 37. Lallena MJ, Diaz-Meco MT, Bren G et al. Activation of IkappaB Kinase beta by Protein Kinase C Isoforms. Mol Cell Biol 1999; 19:2180-2188. 38. Sun Z, Arendt CW, Ellmeier W et al. PKC-theta is required for TCR-induced NF-kappaB activation in mature but not immature T lymphocytes. Nature 2000; 404:402-407. 39. Moscat J, Diaz-Meco MT. The atypical protein kinase Cs. Functional specificity mediated by specific protein adapters. EMBO Rep 2000; 1:399-403. 40. Csukai M, Mochly-Rosen D. Pharmacologic modulation of protein kinase C isozymes: the role of RACKs and subcellular localisation. Pharmacol Res 1999; 39:253-259. 41. Sanchez P, De Carcer G, Sandoval IV et al. Localization of atypical protein kinase C isoforms into lysosome- targeted endosomes through interaction with p62. Mol Cell Biol 1998; 18:3069-3080 42. Sanz L, Diaz-Meco MT, Nakano H et al. The atypical PKC-interacting protein p62 channels NF-kappaB activation by the IL-1-TRAF6 pathway. EMBO J 2000; 19:1576-1586. 43. Sanz L, Sanchez P, Lallena MJ et al. The interaction of p62 with RIP links the atypical PKCs to NF-kappaB activation. EMBO J 1999; 18:3044-3053. 44. Rzymkiewicz DM, Tetsuka T, Daphna-Iken D et al. Interleukin-1beta activates protein kinase C zeta in renal mesangial cells. Potential role in prostaglandin E2 up-regulation. J Biol Chem 1996; 271:17241-17246. 45. Thomas JA, Allen JL, Tsen M et al. Impaired cytokine signaling in mice lacking the IL-1 receptorassociated kinase. J Immunol 1999; 163:978-984. 46. Cao Z, Henzel WJ, Gao X. IRAK: a kinase associated with the interleukin-1 receptor. Science 1996; 271:1128-1131. 47. Cao Z, Xiong J, Takeuchi M et al. TRAF6 is a signal transducer for interleukin-1. Nature 1996; 383:443-446. 48. Adachi O, Kawai T, Takeda K et al. Targeted disruption of the MyD88 gene results in loss of IL-1- and IL- 18-mediated function. Immunity 1998; 9:143-150. 49. Wooten MW, Seibenhener ML, Mamidipudi V et al. The atypical protein kinase C-interacting protein p62 is a scaffold for NF-kappaB activation by nerve growth factor. J Biol Chem 2001; 276:7709-7712. 50. Dominguez I, Diaz-Meco MT, Municio MM et al. Evidence for a role of protein kinase C zeta subspecies in maturation of Xenopus laevis oocytes. Mol Cell Biol 1992; 12:3776-3783. 51. Dominguez I, Sanz L, Arenzana-Seisdedos F et al. Inhibition of protein kinase C zeta subspecies blocks the activation of an NF-kappa B-like activity in Xenopus laevis oocytes. Mol Cell Biol 1993; 13:1290-1295.
112
Protein Kinase C
52. Berra E, Diaz-Meco MT, Dominguez I et al. Protein kinase C zeta isoform is critical for mitogenic signal transduction. Cell 1993; 74:555-563. 53. Bjorkoy G, Overvatn A, Diaz-Meco MT et al. Evidence for a bifurcation of the mitogenic signaling pathway activated by Ras and phosphatidylcholine-hydrolyzing phospholipase C. J Biol Chem 1995; 270:21299-21306. 54. Bjorkoy G, Perander M, Overvatn A et al. Reversion of Ras- and phosphatidylcholine-hydrolyzing phospholipase C- mediated transformation of NIH 3T3 cells by a dominant interfering mutant of protein kinase C lambda is accompanied by the loss of constitutive nuclear mitogen-activated protein kinase/extracellular signal-regulated kinase activity. J Biol Chem 1997; 272:11557-11565. 55. Diaz-Meco MT, Lozano J, Municio MM et al. Evidence for the in vitro and in vivo interaction of Ras with protein kinase C zeta. J Biol Chem 1994; 269:31706-31710. 56. Liao DF, Monia B, Dean N et al. Protein kinase C-zeta mediates angiotensin II activation of ERK1/2 in vascular smooth muscle cells. J Biol Chem 1997; 272:6146-6150. 57. Wooten MW, Seibenhener ML, Matthews LH et al. Modulation of zeta-protein kinase C by cyclic AMP in PC12 cells occurs through phosphorylation by protein kinase A. J Neurochem 1996; 67:1023-1031. 58. Marshall CJ. Ras effectors. Curr Opin Cell Biol 1996; 8:197-204. 59. Downward J. Role of phosphoinositide-3-OH kinase in Ras signaling. Adv Second Messenger Phosphoprotein Res 1997; 31:1-10. 60. Downward J. Ras signalling and apoptosis. Curr Opin Genet Dev 1998; 8:49-54. 61. Le Good JA, Ziegler WH, Parekh DB et al. Protein kinase C isotypes controlled by phosphoinositide 3-kinase through the protein kinase PDK1. Science 1998; 281:2042-2045. 62. Marshall CJ. Cell signalling. Raf gets it together. Nature 1996; 383:127-128. 63. Berra E, Diaz-Meco MT, Lozano J et al. Evidence for a role of MEK and MAPK during signal transduction by protein kinase C zeta. EMBO J 1995; 14:6157-6163. 64. Schonwasser DC, Marais RM, Marshall CJ et al. Activation of the mitogen-activated protein kinase/extracellular signal- regulated kinase pathway by conventional, novel, and atypical protein kinase C isotypes. Mol Cell Biol 1998; 18:790-798. 65. Marshall CJ. Signal transduction. Taking the Rap. Nature 1998; 392:553-554. 66. Uberall F, Hellbert K, Kampfer S et al. Evidence that atypical protein kinase C-lambda and atypical protein kinase C-zeta participate in ras-mediated reorganization of the F-actin cytoskeleton [In Process Citation]. J Cell Biol 1999; 144:413-425. 67. Finco TS, Westwick JK, Norris JL et al. Oncogenic Ha-Ras-induced signaling activates NF-kappaB transcriptional activity, which is required for cellular transformation. J Biol Chem 1997; 272:24113-24116. 68. Wang CY, Mayo MW, Baldwin AS Jr. TNF- and cancer therapy-induced apoptosis: potentiation by inhibition of NF-kappaB. Science 1996; 274:784-787. 69. Mayo MW, Wang CY, Cogswell PC et al. Requirement of NF-kappaB activation to suppress p53-independent apoptosis induced by oncogenic Ras. Science 1997; 278:1812-1815. 70. Joberty G, Petersen C, Gao L et al. The cell-polarity protein par-6 links par3 and atypical protein kinase C to cdc42. Nat Cell Biol 2000; 2:531-539. 71. Lin D, Edwards AS, Fawcett JP et al. A mammalian PAR-3-PAR-6 complex implicated in Cdc42/ Rac1 and aPKC signalling and cell polarity. Nat Cell Biol 2000; 2:540-547. 72. Watts JL, Etemad-Moghadam B, Guo S et al. par-6, a gene involved in the establishment of asymmetry in early C. elegans embryos, mediates the asymmetric localization of PAR-3. Development 1996; 122:3133-3140. 73. Qiu RG, Abo A, Steven Martin G. A human homolog of the C. elegans polarity determinant par-6 links rac and cdc42 to PKCzeta signaling and cell transformation [In Process Citation]. Curr Biol 2000; 10:697-707. 74. Diaz-Meco MT, Moscat J. MEK5, a new target of the atypical protein kinase C isoforms in mitogenic signaling. Mol Cell Biol 2001; 21:1218-1227. 75. Izumi Y, Hirose T, Tamai Y et al. An atypical PKC directly associates and colocalizes at the epithelial tight junction with ASIP, a mammalian homologue of Caenorhabditis elegans polarity protein PAR-3. J Cell Biol 1998; 143:95-106. 76. Tabuse Y, Izumi Y, Piano F et al. Atypical protein kinase C cooperates with PAR-3 to establish embryonic polarity in Caenorhabditis elegans. Development 1998; 125:3607-3614. 77. Noda Y, Takeya R, Ohno S et al. Human homologues of the Caenorhabditis elegans cell polarity protein PAR6 as an adaptor that links the small GTPases Rac and Cdc42 to atypical protein kinase C. Genes Cells 2001; 6:107-119.
Specificity in Atypical Protein Kinase C Signaling: The NF-κB Paradigm
113
78. Jefferies C, Bowie A, Brady G et al. Transactivation by the p65 subunit of NF-kappaB in response to interleukin-1 (IL-1) involves MyD88, IL-1 receptor-associated kinase 1, TRAF-6, and Rac1. Mol Cell Biol 2001; 21:4544-4552. 79. Kato Y, Kravchenko VV, Tapping RI et al. BMK1/ERK5 regulates serum-induced early gene expression through transcription factor MEF2C. EMBO J 1997; 16:7054-7066. 80. Kato Y, Tapping RI, Huang S et al. Bmk1/Erk5 is required for cell proliferation induced by epidermal growth factor. Nature 1998; 395:713-716. 81. Kuroda S, Nakagawa N, Tokunaga C et al. Mammalian homologue of the Caenorhabditis elegans UNC-76 protein involved in axonal outgrowth is a protein kinase C zeta-interacting protein. J Cell Biol 1999; 144:403-411. 82. Diaz-Meco MT, Lallena MJ, Monjas A et al. Inactivation of the inhibitory kappaB protein kinase/ nuclear factor kappaB pathway by Par-4 expression potentiates tumor necrosis factor alpha-induced apoptosis. J Biol Chem 1999; 274:19606-19612. 83. Sells SF, Wood DP Jr., Joshi-Barve SS et al. Commonality of the gene programs induced by effectors of apoptosis in androgen-dependent and -independent prostate cells. Cell Growth Differ 1994; 5:457-466. 84. Guo Q, Fu W, Xie J et al. Par-4 is a mediator of neuronal degeneration associated with the pathogenesis of Alzheimer disease. Nat Med 1998; 4:957-962. 85. Diaz-Meco MT, Municio MM, Frutos S et al. The product of par-4, a gene induced during apoptosis, interacts selectively with the atypical isoforms of protein kinase C. Cell 1996; 86:777-786. 86. Barradas M, Monjas A, Diaz-Meco MT et al. The downregulation of the pro-apoptotic protein Par-4 is critical for Ras-induced survival and tumor progression. EMBO J 1999; 18:6362-6369. 87. Leitges M, Sanz L, Martin P et al. Targeted disruption of the zeta-PKC gene results in the impairment of the NF-kappaB pathway. Mol Cell 2001; 8:771-780.
114
Protein Kinase C
CHAPTER 10
Protein Kinase C: A Molecular Information Storage Device in Neurons Laura A. Schrader, Coleen M. Atkins, Michael Leitges, J. David Sweatt and Edwin J. Weber
Introduction
O
ther chapters in this book have outlined the broad and diverse roles that PKC plays in many different cell types as well as the elegant and complex regulatory mechanisms operating to control the activation of PKC. In this chapter, we will discuss the unique role of PKC in the nervous system, focusing on its critical involvement in the regulation of the strength of synaptic connections between neurons. We will also extrapolate from this cellular role into a discussion of roles for PKC in the behaving animal, particularly in the context of learning and memory. In several instances, we will highlight how the diverse biochemical mechanisms for regulating PKC activity confer a capacity for PKC to play unique roles in triggering and maintaining long-lasting changes, both in cellular function and behavioral modification.
Activity-Dependent Synaptic Plasticity Activity-dependent plasticity of the efficacy of synaptic transmission between neurons is the foundation for most models of development and learning and memory. This idea is captured succinctly in Hebb’s postulate,1 a modern version of which states that memory is subserved by coincident pre- and postsynaptic activity which induces synaptic strengthening, while asynchronous activity causes no change or induces synaptic weakening. The hippocampus exhibits robust synaptic plasticity and has become a focal point for neuroscientists investigating how the brain encodes memories for several reasons. Perhaps the most influential has been the finding that the hippocampus is critical for the acquisition of declarative memories, the knowledge of facts and events (for a review see ref. 2). In accordance with this finding are anatomical studies, which have determined that the hippocampal system is uniquely positioned to receive converging inputs from the brain’s various sensory processing modules.3 In pursuit of the mechanisms and conditions of synaptic plasticity, long-term potentiation (LTP) has emerged as the most common cellular model for learning and memory. LTP is a persistent (i.e., lasting up to days) enhancement of synaptic strength, commonly induced by brief high-frequency stimulation (tetanus) of presynaptic neurons or postsynaptic depolarization coincident with presynaptic neurotransmitter release. Hippocampal LTP was first induced in anesthetized rabbits with brief bursts of high-frequency stimulation4 and later in awake rabbits,5,6 establishing it as a viable model for learning and memory in vivo. The introduction of LTP in the hippocampal slice preparation7 instigated an explosion in the analysis of the molecular mechanisms underlying this increase in synaptic efficacy. While LTP can be induced in various brain areas, most research has been done in the hippocampus, based on its importance in learning and memory and robustness of LTP that can be induced at each of the synapses in the trisynaptic circuit (see Fig. 1). Protein Kinase C, 2nd Edition, edited by Lodewijk Dekker. ©2004 Eurekah.com and Kluwer Academic / Plenum Publishers.
Protein Kinase C: A Molecular Information Storage Device in Neurons
115
Figure 1. Schematic diagram of the hippocampal formation illustrating the trisynaptic circuit. 1) The perforant path-granule cells of the dentate gyrus synapse. 2) The granule cell-CA3 pyramidal cell synapse 3). Schaffer collateral-CA1 pyramidal cell synapse.
The molecular mechanisms for the induction of various forms of LTP in different areas of the brain as well as at the synapses of the hippocampus display various differences (for review see refs. 8,9) and for clarity’s sake we will focus our discussion on N-methyl-D-aspartate receptor (NMDAR)-dependent LTP in hippocampal area CA1. The different variations of LTP, however are all similar in that an increase in postsynaptic Ca2+ is necessary for LTP induction (reviewed by 8,9 but see ref. 10 ). While several sources for postsynaptic Ca2+ exist,11-18 the NMDAR is the most common source of Ca2+ and its activation is necessary for many forms of LTP.8,9,19-22 Paradigms to induce LTP include the depolarization of the postsynaptic neuron by some means along with presynaptic stimulation. The depolarization (either directly through current injection, or indirectly, through repetitive stimulation of presynaptic axons (tetanus), release of glutamate from the presynaptic terminal and activation of postsynaptic glutamate receptors) of the postsynaptic neuron relieves the Mg2+ block of one type of glutamate receptor, the NMDAR, on the postsynaptic neuron, allowing it to be opened by glutamate released from the presynaptic terminal (hence the NMDAR is termed a coincidence detector, sensing pre- and postsynaptic activity). This opening of the NMDAR allows Ca2+ influx into the postsynaptic neuron and this Ca2+ signal triggers a complex cascade of biochemical events in the neuron. This cascade of intracellular signals triggered by the increase in postsynaptic Ca2+ includes activation of various protein kinases. Several protein kinases are required either pre- or postsynaptically for NMDA receptor-dependent LTP at the Schaffer collateral/commissural pathway of the hippocampus, including protein kinase A (PKA), protein kinase C (PKC), calcium/ calmodulin-dependent protein kinase II (CaMKII), protein kinase G, tyrosine kinases and mitogen-activated protein kinase (MAPK).23-25 Based on the fact that these kinases are activated at different times during LTP, LTP has been divided into an early and a late phase, which is dependent upon protein synthesis.26-29 This chapter will focus on the events involving protein kinase C, and we will describe several lines of evidence that demonstrate that PKC plays an important role in LTP. These include: physiological blockade of LTP with PKC inhibitors, mimicry of LTP with PKC activators, and biochemical data showing that PKC activation plays a unique role in LTP.
116
Protein Kinase C
Figure 2. Inhibition of PKC with polymyxin B applied intracellularly through the recording electrode blocks LTP. Peak EPSP amplitude is plotted as a percent of control (pretetanus). LTP obtained in a CA1 pyramidal neuron induced by tetanic stimulation (arrow) of the Schaffer collaterals (open circles) and perforant path (open squares). LTP was blocked when polymyxin B was included in the recording electrode after tetanic stimulation of the Schaeffer collaterals (closed circles), and perforant path (closed squares). Inset- an example of the EPSP recorded before the tetanus (0), just after the tetanus (0.1 min) and 30 minutes after the tetanus (30) in control (open circle) and with polymyxin B in the recording electrode (closed circles). This figure is adapted from ref. 33.
Inhibitors of PKC Block LTP The phenomenon of LTP is monitored electrophysiologically while recording extracellularly to monitor the field excitatory postsynaptic potential (fEPSP) or intracellularly to monitor the EPSP, pre- and post-tetanus. A persistent increase in the slope of the fEPSP or amplitude of the PSP indicates successful potentiation (see Fig. 2). Block of LTP is considered to be no significant change in the amplitude or slope of the PSP from baseline beyond 20-30 minutes after tetanus. Initial studies that implicated protein kinases in LTP used broad-spectrum kinase inhibitors applied to hippocampal slices. Nonspecific inhibitors and confounding effects of the blockers used, however, made interpretation of these results from early studies difficult. Moreover, in some early studies, there was dispute as to phase of LTP that PKC participates in, depending on the kinase inhibitors used. Biochemical studies have shed some light on this issue (see below). In general, bath application of broad-spectrum kinase inhibitors (mellitin, polymyxin B, H-7), to the hippocampal slice prep before LTP induction did not affect initial potentiation, but blocked LTP.30-33 Moreover, PKC inhibitors bath applied 10 and 60 minutes after the tetanus blocked LTP, suggesting a role for PKC in maintenance of LTP.34 In addition, another study using peptide inhibitors injected into the postsynaptic neuron suggested that postsynaptic PKC plays a role in the induction of LTP, but that presynaptic PKC is necessary for the maintenance of LTP.35 Other studies indicated that postsynaptic PKC was essential for both LTP induction and maintenance33,36 (see Fig. 2). Finally, in more contemporary knockout mouse studies, genetic deletion of PKC-γ causes a clear loss of tetanus-induced LTP
Protein Kinase C: A Molecular Information Storage Device in Neurons
117
Figure 3. PKC activation increases synaptic efficacy at Schaffer collateral/CA1 synapse. A 20 minute bath application of the phorbol ester, PDA, to a coronal slice of hippocampus, caused a >100% increase in the slope of the field EPSP recorded in stratum radiatum of CA1. This potentiation persisted after PDA washout. This demonstrates that PKC activation is sufficient to induce ‘LTP’ at this synapse.
in the hippocampal area CA1, although this loss of LTP in PKC-γ knockout mice can be recovered by priming with low-frequency stimulation.37 Despite some vagaries and disagreement in the literature, taken together there is solid evidence for a necessity for PKC in LTP induction and the available evidence strongly implicates PKC as playing a part in maintaining the ongoing expression of LTP.
Activating PKC Mimics LTP
Phorbol esters are used as potent activators of PKC38 and bath application has been shown to induce ‘LTP’ at hippocampal synapses39-43 (see Fig. 3), which occludes normal LTP induced by tetanization. In general, bath application of phorbol esters is thought to facilitate transmitter release from presynaptic terminals.43,44 In one study, iontophoretic application of phorbol esters did not affect sensitivity to glutamate in hippocampal cells,44 suggesting a purely presynaptic action (but see ref. 45 and discussion below on targets of protein kinase C). In addition, injection of PKC into hippocampal CA1 cells produces a long lasting enhancement of the PSP, which occludes LTP,46 suggesting a postsynaptic mechanism. Moreover, PKC activation increases frequency and amplitude of miniature excitatory postsynaptic currents (mEPSCs) in CA1 pyramidal cells,47 which implies both a pre-and postsynaptic action. At present, it is unclear what the presynaptic molecular targets of PKC are in this context, but it is an area of active investigation. These data clearly indicate that PKC activation is sufficient to increase synaptic strength in the hippocampus. Thus, from the data presented in the above sections, we can surmise that PKC activation appears to be both necessary and sufficient for LTP. In the following sections, we will discuss biochemical studies demonstrating that PKC activation indeed occurs with the induction of LTP physiologically.
118
Protein Kinase C
Figure 4. Schematic diagram showing PKC activation and PKM formation during LTP. Ca2+ influx through NMDARs or VGCCs induced by tetanic stimulation induces PKC translocation to the membrane and de novo PKM zeta protein synthesis. In addition, activation of group I mGluRs activates PLC and translocation of PKC. The activated PKC that was translocated to the membrane is cleaved by the enzyme calpain, and a persistently activated catalytic domain, PKM remains in a form of NMDAR-independent LTP. The newly synthesized PKM zeta is increased during LTP, but decrease following LTD induction.
Biochemical Studies While physiological data shows that PKC plays a role in LTP, biochemical studies have led to a better understanding of both the mechanisms through which it acts and the targets on which it acts. PKC activation has been shown within seconds of tetanic stimulation48,49 and protein kinase activity is persistently increased during LTP maintenance and induction.48,50-52 As mentioned above, this persistent activation is believed to be a mechanism that underlies long-term potentiation.53,54 PKC-α and βII autophosphorylate on Thr634 and Thr641 and this autophosphorylation increases during hippocampal LTP,55 additional evidence for PKC activation in LTP. Interestingly, translocation of protein kinase C to the membrane occurs coincidentally with LTP,56 and is dependent upon activation of metabotropic glutamate receptors.57 Several isozymes of PKC are translocated from the inactive state in the cytosol to the membrane during LTP.49 This translocation, however, is not sustained in LTP at the Schaffer collateral/CA1 synapse.50,51,58 However, other forms of synaptic plasticity do show persistent PKC translocation.59-61 An important series of studies by Todd Sacktor’s group has shown that one isotype, PKMζ is synthesized de novo after LTP induction as a second messenger-independent, constitutively active form,52 which lacks the regulatory domain. An increase in PKMζ is seen at least 2 hours after the tetanus,52 while a decrease of PKMζ is seen after induction of long-term depression (LTD)49 (see Fig. 4 for schematic description). Proteolytic activation of the classical isotypes of PKC also appears to occur with NMDAR-independent LTP at these same synapses.62 This is an interesting example of biochemically convergent mechanisms (autonomously active PKMs) in two forms of LTP triggered by different induction mechanisms.
Oxidative Modification of PKC Several protein kinases and phosphatases are regulated by reactive oxygen species. Typically, protein kinases are activated by reactive oxygen species, whereas protein phosphatases are
Protein Kinase C: A Molecular Information Storage Device in Neurons
119
inhibited, potentially enabling a concerted modulation of protein phosphorylation levels within the cell. PKC is activated by reactive oxygen species in a complex manner. Hydrogen peroxide and superoxide increase both autonomous and cofactor-dependent PKC activity.63-66 Several lines of evidence suggest that activation of PKC by reactive oxygen species plays a role in LTP. First, NMDAR activation in area CA1 of hippocampal slices results in superoxide production,64 providing a source of superoxide. Second, the α, βII, ε and ζ isotypes of PKC are autonomously activated by reactive oxygen species by thiol oxidation and release of zinc from cysteine-rich regions of PKC.67 The generation of these persistently activated PKC occurs concomitantly with LTP induction. And finally, superoxide scavengers inhibited LTP induction.66,68
Protein Kinase Substrates Phosphorylated during LTP Of course, PKC has been shown to regulate a wide variety of intracellular processes, including direct effects on channels and receptors, (see ref. 69 for a detailed review), and several other chapters of this book deal with the wide and important variety of cellular targets of PKC. This section will highlight some of PKC’s targets particularly relevant to LTP.
GAP-43 and Neurogranin Two of the first identified PKC substrates phosphorylated during LTP were neuromodulin (known also as GAP43, B50, F170,72 and neurogranin (known also as RC370-73). GAP-43 is localized presynaptically74-76 and neurogranin is localized postsynaptically.74 Interestingly, GAP-43 and neurogranin may play similar roles as both are calmodulin-binding proteins whose affinity for calmodulin decreases when the protein is phosphorylated. Thus, phosphorylation of these substrates by PKC may increase free calmodulin levels during LTP, potentially feeding into the CaMKII cascade. These increases in phosphorylation of both of these substrates is maintained for 1-2 hours, an independent line of evidence for the existence of a persistently active PKC in LTP. Interestingly, as one substrate is largely pre- and one is largely postsynaptic, these data also suggest that PKC is persistently activated both pre- and postsynaptically.
NMDAR As described above, the NMDAR is a primary molecule involved in LTP, acting as a coincidence detector to sense presynaptic stimulation (as indicated by synaptic glutamate) and postsynaptic depolarization. The NMDARs are a heteromultimeric channel, which is made up by NR1 and NR2 subunits. PKC has been shown to phosphorylate the NR1, 2A and 2B subunits of the NMDAR.77,78 Furthermore, PKC activation increases NMDAR channel open probability and NMDA-elicited currents in spinal cord neurons,79,80 but decreases the response to NMDA in hippocampal CA1 neurons.81 These conflicting results may be due to tissue specificity or other effects of phorbol esters in the hippocampus (i.e., presynaptic), as studies on recombinant NMDARs consistently demonstrated that PKC activation increases NMDAR currents in Xenopus oocytes.82-84 Perhaps not surprisingly, while NR2A and NR2B subunits are phosphorylated by PKC and are strongly implicated in synaptic plasticity,9,85,86 NMDARs containing the NR2C subunit, which have limited distribution in the brain, are not potentiated by PKC.87 Interestingly, the typical serine/threonine phosphorylation by PKC of the main subunits does not appear to be involved in the functional PKC modulation of the channel as mutant subunits lacking all the known sites of PKC phosphorylation still exhibit PKC potentiation.88 Modulation of the current by PKC appears to occur through increasing NMDA channel opening rate and cycling of new NMDARs to the cell surface, possibly through phosphorylation of other associated proteins.89 Another mechanism, tyrosine phosphorylation may also be involved, as PKC activation has also been shown to induce tyrosine phosphorylation of NR2A and NR2B90 and tyrosine phosphorylation of the NMDAR appears to contribute to the maintenance of LTP.91 A recent study, however, shows that activation of PKC directly phosphorylates two sites in the NR2B C terminus, leading to enhanced currents through NMDAR
120
Protein Kinase C
channels.92 Indeed, PKC phosphorylation of the NMDAR increases current through relief of the Mg2+ block.80 In addition, activation of several G-protein coupled receptors increases NMDAR currents most likely through PKC activation. These receptors include: mGluRs,84,93-95 µ opioid receptors82 and muscarinic receptors.96 The upregulation of the NMDARs by PKC activation provides a possible explanation for PKC’s role in regulating the induction of LTP. It is also important to note that phosphorylation may not simply regulate receptor function, but may play a role in its binding to various cytoskeletal proteins.
AMPAR α-amino-3-hydroxy-5-methylisoxazole-4-propionic acid (AMPA) receptors mediate the majority of fast synaptic transmission throughout the nervous system.97 Four homologous subunits of GluR1-GluR4 combine in a pentameric complex to form AMPARs. These subunits are also phosphorylated by PKC in cultured hippocampal neurons,45,98,99 and are phosphorylated at Ser831 by PKC in the hippocampal slice preparation.100 Phosphorylation of the GluR1 increases the receptor’s conductance, potentially providing a direct route for PKC to enhance synaptic efficacy during LTP,99 as the current through the AMPAR is the major component of the fast excitatory postsynaptic current. Another AMPAR subunit, GluR2, is phosphorylated at the C-terminus by protein kinase C at serine 880, and phosphorylation at that site reduces the ability of GluR2 to bind with glutamate receptor interacting protein (GRIP), a PDZ-domain containing protein that links AMPARs to cytoskeletal proteins and other neuronal proteins.101,102 Interestingly, PKC stimulation of neurons results in rapid internalization of GluR2,101 which appears to underly cerebellar LTD (see below for a more detailed discussion).
Metabotropic Glutamate Receptors While the AMPARs and NMDARs are ligand-gated glutamate receptors and mediated fast synaptic transmission, the metabotropic glutamate receptors (mGluRs) are coupled to G-proteins and mediate slower, modulatory functions of neurons. Three groups of the mGluRs exist, and group I mGluRs (mGluR1 and mGluR5) activates PKC through Phospholipase C (PLC). All groups (1-3) are also modulated by PKC. Metabotropic glutamate receptors (mGluRs) have been implicated in various forms of plasticity, including LTP14,17,108 (but see refs. 109,110). The mechanisms of mGluR involvement in plasticity are still under investigation, however, the group I mGluRs (mGluR1 and mGluR5) provide a source of intracellular Ca2+, as their activation induces release of Ca2+ from intracellular stores.111 Phosphorylation of mGluR5 appears to mediate the oscillatory nature of mGluR5-induced Ca2+ elevations,112 as opposed to mGluR1, which exhibits a single spike of intracellular Ca2+. These oscillations could have important implications for plasticity as CaMKII autophosphorylation is critically dependent on individual Ca2+ spikes and can decode the frequency of Ca2+ oscillations into specific cellular processes.113 Presynaptic mGluRs (group II and group III) are inhibited by PKC activation,115 an effect that may be caused by blocking the coupling of the mGluR to GTP-binding proteins.116,117 Furthermore, activation of adenosine receptors reduces mGluR activity at the Shaffer collateral-CA1 synapse by a PKC-dependent mechanism.117 As activation of presynaptic receptors tends to act to reduce release of GABA and glutamate from presynaptic terminals throughout the nervous system,118-121 the effect of inhibiting these receptors may actually increase transmitter release from the presynaptic terminal. In addition, PKC activation is involved in the desensitization of group I (PI-coupled) mGluRs and this desensitization is blocked by PKC inhibitors.122-124 This appears to be an effect of direct phosphorylation of mGluRs by PKC,125 involving multiple phosphorylation sites.126 The group I receptors perform various functions in hippocampal cells, most of which are excitatory.127 Thus the rapid desensitization of these receptors could limit excitotoxic actions of glutamate, as well as dynamically regulating synaptic plasticity.
Protein Kinase C: A Molecular Information Storage Device in Neurons
121
Protein Kinases Interestingly, it appears that protein kinases themselves are substrates during LTP. As described above, PKC is autophosphorylated during LTP.55 When PKC autophosphorylates, the Km for calcium decreases and a slight increase in autonomous activity is observed.128 Autonomous activation of PKC occurs within 2 min of LTP induction and lasts for up to 2 h. However, autophosphorylation may not underly the increased autonomous activation of PKC observed during LTP, but rather by oxidative modification as described above.66 Recent evidence suggests that PKC autophosphorylation may contribute to protecting PKC from downregulation or may control its subcellular localization by regulating cytoskeletal interactions.129-132 In addition, as described above, the extracellular signal-regulated kinase (ERK)/MAPK cascade is involved in triggering LTP (for detailed review see ref. 25). PKC activation of ERK in neurons has been described,133,134 an effect that has been described in a wide variety of cell types.135 Thus one potential target of PKC in LTP induction is the ERK/MAPK cascade, although investigations of the role of this system in LTP are still in the very early stages.
Cerebellar Long-Term Depression While we have focused on LTP in the hippocampus for most of our discussion of roles for PKC in synaptic plasticity, we would be remiss if we did not briefly address an elegant series of studies implicating a role for PKC in LTD, in the cerebellum particularly. Parallel fiber long-term depression in the cerebellum is a persistent, input-specific decrease in the efficacy of synaptic transmission between the parallel fibers and Purkinje cells136 that has been suggested to underly several forms of cerebellar motor learning.137,138 It is induced by low-frequency coactivation of climbing fibers and parallel fiber inputs and induction requires mGluR and AMPAR activation, postsynaptic Ca2+ influx through voltage-gated Ca2+ channels (reviewed by ref. 139). The Ca2+ subsequently activates PKC, as PKC is necessary for induction of cerebellar LTD140-147 and PKC activators mimic LTD.148 As mentioned above, a series of studies have detailed the specific role of PKC in cerebellar LTD in great detail. AMPAR internalization appears to underly LTD, as interference with clathrin endocytosis blocks the induction of cerebellar LTD, and induction of AMPAR internalization produces an LTD that occludes stimulus-induced LTD.149 PKC phosphorylates the GluR2/3 AMPA receptor at serine 880 (serine 885 in GluR3). This phosphorylation decreases the binding of the GluR to GRIP/AMPAR binding protein (ABP),101,102,150,151 a PDZ domain containing protein, which serves as an adaptor to cross-link AMPAR to other neuronal proteins.152 In addition, PKC activation is associated with a translocation of the phosphorylated GluR2 and protein interacting with C kinase (PICK)-1, another PDZ containing protein that interacts with AMPARs, but not GRIP immunoreactivity from dendritic shafts to spines in cultured hippocampal cells.101 Thus, LTD requires PKC phosphorylation of the GluR2/3 receptors, which regulates intereactions with several PDZ domain-containing proteins.153
PKC in Learning and Memory As discussed above, synaptic plasticity may constitute the cellular mechanism by which the brain encodes cognitive alteration and memory formation. The necessity for PKC in the molecular mechanism of synaptic plasticity suggests the possibility of a prominent role for PKC in learning and memory processes. Correlative studies involving PKC and mammalian learning and memory rely on three basic investigative approaches. These approaches include: 1) measuring changes in neuronal PKC activity, cellular distribution and substrate phosphorylation with learning, 2) determining the consequences of pharmacological inhibition of PKC on memory formation, 3) measuring behavioral changes following genetic deletion of specific PKC isotypes. Each of these approaches has shed light on the role of PKC in memory formation in various spatial and associative learning paradigms, and at present there is compelling evidence of a necessity for PKC for normal learning and memory formation.
122
Protein Kinase C
Although many behavioral tests exist which are used to evaluate learning and memory, we will discuss only the watermaze paradigm used as a spatial learning test, and eyeblink conditioning, fear-conditioned learning and one-trial step-down avoidance tasks used to test associative learning. Each of these paradigms is used to evaluate hippocampus-dependent learning and memory. The Morris water maze consists of a round tank (pool) of water in which animals are trained to find an invisible submerged escape platform using distal visual cues for spatial orientation. Learning is assessed by the latency to find the platform following several days of training. Memory of the training is determined by removing the platform and counting the number of times the animal swims across the area where the platform was located during training, and the amount of time spent in the target quadrant of the tank during the platform search. During classical associative learning, an animal is taught to associate a neutral conditioned stimulus (CS) with an aversive unconditioned stimulus (US). Classical conditioning to the eyeblink response in rabbits uses the association of a neutral stimulus such as tone or light with a nociceptive stimulus, such as an airpuff delivered to the eye or a periorbital shock. Re-presentation of the CS results in an eyeblink conditioned response (CR) in anticipation of the US. In classical fear conditioned learning, animals are placed in a novel environment (context) and exposed to the pairing of a tone (CS) with a mild foot shock (US). Fear learning is assessed by the amount of freezing behavior exhibited by the animal following re-presentation to the context or to the CS in the presence of a novel context. During one-trial step-down avoidance, animals are placed on an elevated platform and given a mild foot shock when the animal steps off the platform onto the grid below. Memory is then assessed by the latency to step off the platform following training. These behavioral tests are being used both as learning paradigms to check for increased PKC activity and as assays for the generation of learning and memory deficits when PKC is inactivated pharmacologically or genetically.
Changes in Activity, Cellular Distribution and Phosphorylation of Substrates A clear relationship exists between PKC activity, cellular distribution, and performance in a number of hippocampus-dependent spatial and associative learning tasks. Spatial learning using hole board and radial arm mazes results in an increase in PKC activity and subcellular translocation from the cytosolic fraction to the membrane fraction following training in rabbit. 154 Spatial memory assessed with the watermaze task reveals an increase in PKC membrane-associated activity in rat, which is comparative to the increase in performance. Atkins et al observed an increase in hippocampal PKC activation in contextual fear conditioning as well.155 Evidence suggests that the calcium-dependent forms of PKC are involved, specifically the gamma isotype of PKC, which shows an increase in membrane-associated concentration related to enhanced spatial learning performance. These results suggest an explanation for observations of age-related learning deficits in the watermaze task.156 It is postulated that the decline in neuronal plasticity and cognitive function that occurs in advanced age may result in part from altered PKC phosphorylation of specific proteins. Aged rats show a reduction in hippocampal translocation of PKC from the cytosolic to membrane fraction coupled with a reduced PKC substrate phosphorylation (B-50/ Gap-43).157,158 In addition, this also appears to be the case in strain specific differences in spatial learning abilities. The C57BL/6Ibg (C57) mouse strain has higher neuronal PKC activity than the DBA/2Ibg (DBA) strain, and subsequently exhibits greater spatial memory as assessed by the water maze.159 The trace eyeblink, step down avoidance and fear conditioning learning paradigms are all hippocampus-dependent forms of associative learning used to evaluate changes in PKC. As with certain spatial learning tasks mentioned above, PKC activity is increased in the hippocampus and cerebellum following classical conditioning of the eyeblink response in rabbit160
Protein Kinase C: A Molecular Information Storage Device in Neurons
123
Although no changes in PKC-γ subcellular localization has been seen in associative learning, changes in PKC-γ immunoreactivity, but importantly not in total activity, were observed following trace eyeblink conditioning in rabbit.161 Administration of the nootropic agent, oxiracetam, in mice prior to training showed improved learning ability in the watermaze and associative fear conditioning tasks.162 This improvement was correlated to an increase in membrane-bound PKC. Taken together, these data suggest that the changes in neuronal PKC activity and subsequent substrate phosphorylation and translocation from cytosol to membrane, particularly the PKC-γ isotype, are correlated to spatial and associative learning events.
Pharmacological Inhibition of PKC Given these observations of a broad functional relationship of PKC in learning and memory, it is important to assess the behavioral effects of blocking PKC actions. Pharmacological inhibition of PKC has the advantages of allowing targeting of a specific brain region, identifying temporal involvement of kinase activity and conferring the ability to either block or activate PKC at key times in the memory process. The use of pharmacological inhibition of PKC to test the role of this enzyme in the hippocampus-dependent, one-step inhibitory avoidance task has proven valuable in defining a role for PKC in the formation of short-term (STM) and long-term memories (LTM). Intrahippocampal injection of the PKC inhibitors, staurosporin and CGP41231, into area CA1 within the first two hours of training was shown to disrupt memory formation, but did not affect memory when injections were given 180 minutes following training.163 These studies suggest a critical temporal necessity for PKC activation in area CA1 of the hippocampus following conditioned learning and suggest the dissociation in the mechanisms required for STM versus LTM for this learning paradigm. Furthermore, the use of one-trial step-down avoidance task with intrahippocampal injection of a specific PKC-isotype inhibitor (Gö 6976) suggests that this type hippocampus-dependent memory formation is reliant on PKC-βI.164-166 Injection of the PKC inhibitor, H7, directly into the amygdala prior to training attenuates fear-conditioned learning in a subregion specific manner with an affect of infusion into the basolateral amygdala (BLA), but not the central nucleus of the amygdala in rats.167 In addition, H7 injection into the cerebellum can disrupt the acquisition, but not the retention, of classical eyeblink conditioning in rabbits,168 again suggesting a brain region-specific PKC dependence in learning and memory.
Isotype Specific PKC Knockout While a role for the PKC enzyme family is broadly established for various forms of mammalian learning, very little is known concerning the contributions of the specific subtypes of PKC to learning and memory processes. One of the most successful approaches to address this dilemma is testing for learning and memory alterations in mouse models deficient for specific PKC isotypes. There are recognized caveats with this line of research, including the unknown affect on fetal development in the absence of a PKC isotype and the likelihood that all PKC isotypes may serve together to fine-tune neuronal signaling. However, this method of investigation also alleviates potential damage that may occur to brain regions of interest from a penetrating injection or canulae used for delivery of pharmacological agents. Regardless, this strategy has helped to better define the possible roles of the gamma, beta and delta isotypes of PKC in learning and memory. Mice deficient for PKC-γ exhibit modified LTP, but only mild spatial and associative learning deficits.169 Importantly, PKC-γ deficient mice are impaired in the contextual component but not in the auditory component of fear conditioned learning. This supports the hypothesis that the PKC-γ isotype is involved in hippocampus-dependent learning and memory mechanisms, however amygdala-dependent processes, appear to be unaffected.169 Protein phosphorylation of neurogranin/RC3 (Ng), a prominent PKC substrate in the brain involved in synaptic plasticity, is reduced in the PKC-γ knockout.170 This inability to increase RC3 phosphorylation may contribute to the spatial and associative learning deficits as well as the impaired
124
Protein Kinase C Figure 5. Learning and memor y in PKC-β-deficient mice. A) Freezing behavior on the day of training for PKC-β-deficient or littermate control mice. Control mice displayed significantly higher freezing in response to the shock (p