"Medicine, Faculty of"@en . "Biochemistry and Molecular Biology, Department of"@en . "DSpace"@en . "UBCV"@en . "Karjala, Geoffrey William George"@en . "2009-12-03T22:22:53Z"@en . "2004"@en . "Master of Science - MSc"@en . "University of British Columbia"@en . "Microtubules are dynamic polymers of the protein tubulin. The dynamic property\r\nof microtubules allows the network to breakdown and form new structures like the\r\nmitotic spindle. Microtubule associated proteins and small molecules can alter the\r\ndynamics. Changing the dynamics by the addition of small molecules can prevent the\r\nproper formation of the mitotic spindle and lead to mitotic arrest. Some of these\r\nantimitotic agents have had clinical success in cancer treatment with the vinca alkaloids\r\ninducing the underpolymerisation and paclitaxel (Taxol\u00E2\u0084\u00A2) promoting the\r\noverpolymerisation of microtubules. In addition, small molecules that alter the\r\nmicrotubule network are gaining use in chemical genetics.\r\n\r\nA phenotypic antimitotic assay has been used to discover three novel compounds\r\nthat interact with the microtubule network. The first two, ceratamine A and B are classic\r\nantimitotic agents that block cells in M-phase and prevent proliferation, probably by\r\ndisrupting the microtubule network. They stimulate the over-polymerization of\r\nmicrotubules in vitro as determined by both microtubule polymerization assays and\r\nelectron microscopy. In vivo, ceratamines induce the formation of tubulin-containing\r\nstructures not previously described. These structures include pillars of tubulin in mitotic\r\ncells and a perinuclear microtubule network during interphase. Ceratamines do not\r\ncompete with paclitaxel for binding to the microtubules, and are very structurally simple.\r\n\r\nThe same phenotypic antimitotic assay was used to discover dimethyl varacin\r\n(DMV). Flow cytometry data indicated that DMV is not an antimitotic agent and does not\r\nblock cells in any particular phase of the cell cycle, though it does inhibit proliferation. It induces strong GF-7 phospho-specific antibody binding (typical only in mitosis) from all\r\nphases of the cell cycle. Western blots and Kinexiis Kinetworks\u00E2\u0084\u00A2 screens showed that a\r\ntwo hours DMV treatment at 5 \u00CE\u00BCM leads to the strong activation of MAPK pathways\r\n(Erkl/2, p38 \u00CE\u00B1 MAPK, and JNK/SAPK) without a significant increase in Cdc2 activity or\r\nglobal phosphorylation. It was also found to strongly inhibit microtubule formation in\r\nvitro and in vivo. Though both DMV and the ceratamines were discovered in the same\r\nantimitotic screen and both target microtubules, they have dramatically different\r\nbiological properties."@en . "https://circle.library.ubc.ca/rest/handle/2429/16296?expand=metadata"@en . "7525479 bytes"@en . "application/pdf"@en . "THE BIOLOGY OF THREE N O V E L N A T U R A L PRODUCT MICROTUBULE INTERACTING AGENTS: CERATAMINE A AND B AND DIMETHYL VARACIN by GEOFFREY WILLIAM GEORGE K A R J A L A B.Sc, The University of British Columbia, 2001 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF T H E REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE i n T H E F A C U L T Y OF G R A D U A T E STUDIES (Department of Biochemistry and Molecular Biology) T H E UNIVERSITY OF BRITISH COLUMBIA December 2004 \u00C2\u00A9 Geoffrey William George Karjala, 2004 Abstract Microtubules are dynamic polymers of the protein tubulin. The dynamic property of microtubules allows the network to breakdown and form new structures like the mitotic spindle. Microtubule associated proteins and small molecules can alter the dynamics. Changing the dynamics by the addition of small molecules can prevent the proper formation of the mitotic spindle and lead to mitotic arrest. Some of these antimitotic agents have had clinical success in cancer treatment with the vinca alkaloids inducing the underpolymerisation and paclitaxel (Taxol\u00E2\u0084\u00A2) promoting the overpolymerisation of microtubules. In addition, small molecules that alter the microtubule network are gaining use in chemical genetics. A phenotypic antimitotic assay has been used to discover three novel compounds that interact with the microtubule network. The first two, ceratamine A and B are classic antimitotic agents that block cells in M-phase and prevent proliferation, probably by disrupting the microtubule network. They stimulate the over-polymerization of microtubules in vitro as determined by both microtubule polymerization assays and electron microscopy. In vivo, ceratamines induce the formation of tubulin-containing structures not previously described. These structures include pillars of tubulin in mitotic cells and a perinuclear microtubule network during interphase. Ceratamines do not compete with paclitaxel for binding to the microtubules, and are very structurally simple. The same phenotypic antimitotic assay was used to discover dimethyl varacin (DMV). Flow cytometry data indicated that D M V is not an antimitotic agent and does not block cells in any particular phase of the cell cycle, though it does inhibit proliferation. It i i induces strong GF-7 phospho-specific antibody binding (typical only in mitosis) from all phases of the cell cycle. Western blots and Kinexiis Kinetworks\u00E2\u0084\u00A2 screens showed that a two hours D M V treatment at 5 uM leads to the strong activation of M A P K pathways (Erkl/2, p38 a MAPK, and JNK/SAPK) without a significant increase in Cdc2 activity or global phosphorylation. It was also found to strongly inhibit microtubule formation in vitro and in vivo. Though both D M V and the ceratamines were discovered in the same antimitotic screen and both target microtubules, they have dramatically different biological properties. in Table of Contents Section Page Abstract ii Table of Contents iv List of Figures vii List of Symbols, Nomenclature, and Abbreviations viii Acknowledgements xiii Chapter I - General Introduction 1 1.1- The Cytoskeleton 1 1.1.1 - Overview 1 1.1.2- Microtubules and Tubulin, Inseparable Partners 1 1.1.3- Microtubules and the Cell Cycle 7 1.1.4 - Microtubule Poisons 12 1.1.5- Microtubules and Microtubule-Associated Proteins 14 1.2- MAP Kinase Signaling 16 1.2.1- Mitogen Activated Protein Kinases (MAPK) 16 1.2.2 - M A P K and the Cytoskeleton 20 1.2.3 - MAP Kinase Phosphatases (MKP) 21 1.3- Goals 22 Chapter II - Materials and Methods 23 2.1-Materials 23 2.2 - Antimitotic Assay 23 2.3 - Antiproliferation Assay 23 i v 2.4 - In Vitro Tubulin/Microtubule Assays 23 2.4.1 - Promotion of Tubulin Polymerization Assay 23 2.4.2 - Inhibition of Tubulin Polymerization Assay 23 2.4.3 - Negative Staining Electron Microscopy of Microtubules 24 2.4.4 - 3H-Paclitaxel Competition Binding to Microtubules 24 2.5 - Flow Cytometry Analysis of the Cell Cycle 25 2.6 - Immunofluorescence Microscopy 25 2.6.1 - P-Tubulin Immunofluorescence Microscopy of D M V Treated Cells 25 2.6.2 - p-Tubulin Confocal Immunofluorescence Microscopy of 25 _ Ceratamine Treated Cells _ 2.6.3 - p-Tubulin Immunofluorescence Microscopy of D M V Treated Cells 26 2.7 - Kinase Assays 26 2.7.1 - Cdc2 Kinase Assays 26 2.7.2 - In-Gel Kinase Assay 26 2.8-Western Blots 26 2.9 - Kinexus Kinetwork\u00E2\u0084\u00A2 Phosphosite Screen 27 2.10- Protein Phosphatase 2A Assay 27 Chapter III - The Ceratamines 29 3.1 - Ceratamine Specific Introduction 29 3.2-Results 31 3.2.1 - Effect of Ceratamine A and B on Cell Cycle Progression 31 3.2.2 - Antimitotic and Antiproliferative Assay 34 3.2.3 - Effect of Ceratamines on Microtubule Organization In Vivo 34 3.2.4 - Effect of Ceratamines on Microtubule Formation In Vitro 39 3.2.5 - Ceratamines do not Compete with Paclitaxel for Binding to ^ Tubulin 3.3 - Discussion 45 Chapter IV - Dimethyl Varacin 52 4.1 - Dimethyl Varacin Specific Introduction 52 v 4.2 - Results 54 4.2.1 - Antimitotic Assay and Antiproliferation 54 4.2.2 - Dimethyl Varacin Inhibits Microtubule Polymerization 54 4.2.3 - Dimethyl Varacin Rapidly Breaks Down the Microtubule Network In Vivo 57 4.2.4 - Dimethyl Varacin Induces Mitotic-Like Phosphorylations from All Phases of the Cell Cycle 57 4.2.5 - Dimethyl Varacin does not Cause Global Activation of Cdc2 60 4.2.6 - In Gel Kinase Assays of Lysate from Dimethyl Varacin Treated Cells 62 4.2.7 - Kinexus Kinetworks\u00E2\u0084\u00A2 Phosphosite Screens 65 4.2.8 - Confirmation Western Blots of Erkl/2, SAPK/JNK, and p38 M A P K Phosphorylation 75 4.2.9 - Protein Phosphatase 2A is not Inhibited by Dimethyl Varacin 78 4.3 - Discussion 79 References 89 v i List of Figures Figure Title Page Figure 1 - Structure of a Microtubule 3 Figure 2 - The Cell Cycle 8 Figure 3 - Microtubule Dynamics and Polymer Mass During the Cell Cycle 10 Figure 4 - The Mitotic Spindle 11 Figure 5 - MAP Kinase Activation and Deactivation Pathways 18 Figure 6 - The Chemical Structure of Paclitaxel, Laulimalide, and Peloruside A 30 Figure 7 - The Chemical Structure of Ceratamine A and B 32 Figure 8 - Flow Cytometry of MCF-7 mp53 Cells Treated with DMSO, Nocodazole, Ceratamine A or Ceratamine B 33 Figure 9 - Antimitotic and Cell Proliferation Assays with Ceratamine A andB 35 Figure 10 - Confocal p-Tubulin Immunofluorescence Images of Ceratamine or Paclitaxel Treated Cells 37 Figure 11 - Cross Section P-Tubulin Staining in Mitotic Cells Figure 12 - In vitro Tubulin Polymerization Assays with Ceratamine A and B 38 40 Figure 13 - Electron Microscopy Photos of Microtubules 41 Figure 14 - 3H-Paclitaxel Binding to Microtubule Assay 44 Figure 15 - The Chemical Structure of Varacin Compounds 53 Figure 16 - Antimitotic ELICA and Cell Proliferation Assays with Various Concentrations of Dimethyl Varacin and Different Cell Lines 55 Figure 17 - Dimethyl Varacin Inhibits Microtubule Formation In Vitro 56 Figure 18 - Dimethyl Varacin Disrupts Microtubules In Vivo 58 Figure 19 - Dimethyl Varacin's Effect on the Cell Cycle 59 Figure 20 - Dimethyl Varacin Does Not Activate Cdc2/cyclin B 61 Figure 21 - In-Gel Kinase Assays 63 Figure 22 - Summary of Data from Kinexus Kinetworks\u00E2\u0084\u00A2 Phosphosite Screens 66-67 Figure 23 - Western Blots of Phospo-Erkl/2, Phospho-JNK, Phospho-p38 MAPK, and Total M A P K from D M V Treated Cell Lysates 76 Figure 24 - Protein Phosphatase 2A Assay 77 vii List of Symbols, Nomenclature, and Abbreviations Abbreviation Name y-TuRC y Tubulin Ring Complex AMP-PK AMP Dependent Protein Kinase ASK1 Apoptosis Signal-Regulating Kinase 1 ATF1 Activating Transcription Factor 1 ATF2 Activating Transcription Factor 2 BSA Bovine Serum Albumin CaMK2 Ca2+/Calmodulin-Dependent Protein Kinase 2 Cdc Cell Division Control CDK Cyclin Dependent Kinase CREB cAMP Responsive Element Binding Protein CTKD C-Terminal Kinase Domain ddGTP dideoxy Guanosine Triphosphate D L K Dual Leucine Zipper-Bearing Kinase DMSO Dimethyl Sulfoxide D M V Dimethyl Varacin DNA Deoxyribonucleic Acid DSP Dual Specificity Phosphatase DTT Dithiothreitol E Exchangeable Site of Tubulin EGFR Epidermal Growth Factor Receptor viii gQ-p^ Ethylene glycol-bis(2-aminoethylether)-N,N,N',N'-tetraacetic acid eIF2Bs Eukaryotic Initiation Factor 2B s Subunit eIF2a Eukaryotic Initiation Factor 2a eIF4 Eukaryotic Initiation Factor 4 eIF4E-BPl Eukaryotic Initiation Factor 4E-Binding Protein 1 ELICA Enzyme Linked Immunocytochemistry Assay ErbB2 Erb2/HER2 Receptor Tyrosine Kinase Erk Extracellular Regulated Kinase FACS Fluorescence-Activated Cell Sorting F A K Focal Adhesion Kinase GDP Guanosine Diphosphate GDP-Tubulin Tubulin with GDP at the Exchangeable Site GRK2 G Protein-Coupled Receptor-Serine Kinase 2 GSK3 Glycogen Synthase Kinase 3 GTP Guanosine Triphosphate GTP-Tubulin Tubulin with GTP at the Exchangeable Site HPLC High Performance Liquid Chromatography IC50 Concentration of 50% Inhibition IGFR Insulin Growth Factor Receptor IKK I kappa B Kinase IRS 1 Insulin Receptor Substrate 1 INK c-Jun N-terminal Kinase ix K-fibre Kinetochore Fibre MAP Microtubule Associated Protein M A P K Mitogen Activated Protein Kinase MAPKAPK-2 MAP Kinase Associated Protein Kinase -2 M A P K K M A P K Kinase M A P K K K M A P K K Kinase MBP Myelin Basic Protein MEF2A Myocyte Ehancer Factor 2A M E K MAP/Erk Kinase M E K K M E K Kinase M K M A P K Activated Protein Kinase M K K MAP Kinase Kinase MKP MAP Kinase Phosphatase M L K Mixed Linage Kinase M N K M A P K Interacting Kinases MNK1 M A P K Interacting Kinase 1 MOPS 3-(N-Morpholino)propanesulfonic acid MSG Monosodium Glutamate MSK M A P K and Stress Activated Protein Kinases MSK1 M A P K and Stress Activated Protein Kinase 1 MSK2 M A P K and Stress Activated Protein Kinase 2 MTOC Microtubule Organizing Centre x mTOR Mammalian Target of Rapamycin J ^ I Y Y 3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazolium bromide N Nonexchangeable Site of Tubulin NF-KB Nuclear Factor KB Op 18 Oncoprotein 18 PAGE Polyacrylamide Gel Electrophoresis PAK p21 Activated Kinase PDK1 Phospho-Inositide Dependent Kinase 1 PIPES Piperazin-l,4-bis(2-ethanesulfonic acid) PKB Protein Kinase B PKC Protein Kinase C PKD Protein Kinase D PKN Protein Kinase N PKR Double Stranded RNA-Dependent Protein Kinase PPI a Protein Phosphatase la PP2A Protein Phosphatase 2A PRK1 Protein Kinase C-Related Protein-Serine Kinase 1 PRK2 Protein Kinase C-Related Protein-Serine Kinase 2 PVDF Polyvinyldifluoride Rb Retinoblastoma RSK Ribosomal S6 Kinase RSK1 Ribosomal S6 Kinase 1 xi RTK Receptor Tyrosine Kinase SAPK Stress Activated Protein Kinase SDS Sodium Dodecyl Sulphate SMAD SMA- and M A D - Related Protein STAT Signal Transducer and Activator of Transcription TAK1 Transforming Growth Factor-beta-Associated Kinase TBS Tris Buffered Saline TBS-T Tris Buffered Saline - 0.1% Tween 20 XKCM1 Xenopus Kinesin Catastrophe Modulator 1 XMAP Xenopus Microtubule Associated Protein Xll Acknowledgements There are many people who I wish to thank. First and foremost is Diane, my wife, whose support during my Master's degree has been immense. I also wish to thank Dr. Roberge for welcoming me into his lab and providing guidance and assistance always with a smile. In addition to Dr. Roberge, the other two members of my supervisory committee, Dr. Andersen and Dr. Numata have been very helpful. The support from the Roberge lab has been wonderful. Along with the lab, there are agencies and people that without whose help the projects I worked on would not have been possible. NSERC provided the money for my stipend and CIHR supported the operation of the Roberge lab. The E7 monoclonal antibody developed by Michael Klymkowsky was obtained from the Developmental Studies Hybridoma Band developed under the auspices of the NICHD and maintained by The University of Iowa, Department of Biological Sciences, Iowa City, IA 52242. Radiolabeled paclitaxel was obtained from the Drug Chemistry and Synthesis Branch, National Cancer Institute. The UBC Bio-imaging facility was also very helpful in obtaining the confocal immunofluorescence images and the electron microscopy images. Andy Johnston at the U B C multi-user flow cytometry facility helped obtain the flow cytometry data. The dimethyl varacin project was greatly aided by two people. First, Lu Yang from the Andersen lab must be mentioned for isolating the compound. Second, Cristina Bigg of the Roberge lab deserves much credit for the project. She is responsible for the data used to prepare Figures 16, 19, 20, and 24. Also, Dr. Pelech and Dr. Zhang of Kinexus were very helpful. xiii The ceratamine project also benefited from the help of others. Emiliano Manzo of the Andersen lab is responsible for isolating the two compounds. Queenie Chan of the Roberge lab also contributed to the project and is responsible for Figures 9, 10, and 11. xiv Chapter I: General Introduction 1.1- The Cytoskeleton 1.1.1 - Overview Cells must exist in a three-dimensional world where external forces are always present and in some cases movement is required. In addition, cells must organize their internal structures, transport vesicles and segregate chromosomes at the appropriate time. In response, cells have evolved a system of filaments to provide a source of structural strength, to partition chromosomes during mitosis, to be \"trails\" for vesicles, to provide motility, and to organize the shape of the cell. These filaments are in fact, three distinct protein polymer classes that together form the cytoskeleton. The first polymer class is actin filaments (microfilaments), which are composed of actin subunits; the second is intermediate filaments composed of intermediate filament proteins; and the third, and the one to be discussed further, is microtubules composed of tubulin subunits. Microtubules partition the organelles in the cell and are responsible for the separation of chromosomes during mitosis. They also form cilia bodies and provide strength to the cell. 1.1.2 - Microtubules and Tubulin, Inseparable Partners Microtubules are polymers of the protein tubulin. The microtubules serve many functions within the cell such as providing strength, shape, and 'traffic' lanes. They also form the mitotic spindle that separates chromosomes during mitosis. To perform these functions tubulin and microtubules must be very adaptable. The innate dynamic instability of microtubules (Desai et al., 1997) allows the microtubules to rearrange and creates this adaptability. 1 Tubulin was initially discovered based on its irreversible binding to the microtubule poison colchicine (Weisenberg et al., 1968). Further work determined that tubulin was a heterodimer of a and p tubulin (Feit et al., 1971) and bound two molecules of guanine nucleotide per mole of tubulin dimer (Weisenberg et al., 1968). One of the guanine nucleotides is GTP bound to a non-exchangeable (N) site. The other guanine nucleotide is far more dynamic, and it is bound at an exchangeable (E) site on the tubulin heterodimer (Nogales et al., 1998). The GTP at the E site is crucial to the correct dynamics of the microtubule. Because tubulin is a heterodimer, it is asymmetrical. As such, the resulting microtubule is asymmetrical with a minus (-) and a plus (+) end. This polarity also plays a critical role in the dynamics of the microtubule as the two ends have different properties (Desai et al., 1997). The microtubule itself is a hollow tube of tubulin typically composed of 13 protofdaments in mammalian cells (Tilney et al., 1973; Evans et al., 1985). A protofilament is a linear polymer of tubulin, and a region of tubulin, the M-loop, is crucial for lateral contacts between protofilaments in the assembly of a microtubule (Li et al., 2002). A schematic of the microtubule structure is shown in Figure 1. It is possible for the number of protofilaments to vary from as few as 8 to as many as 17 (Pierson et al., 1978; Aamodt et al., 1986; Chretien et al., 1992) depending on temperature (Pierson et al., 1979), tubulin source (Aamodt et al., 1986), Microtubule Associated Proteins (MAPs), buffer conditions (Pierson et al., 1978), small molecules (Diaz et al., 1998) and site of nucleation (Evans et al., 1985). While it was initially thought that the number of protofilaments was determined at nucleation, it has now been shown that the number can vary within one microtubule (Chretien et al, 1992). 2 TJLC : - \u00E2\u0080\u00A2 \u00E2\u0080\u00A2 n>~ nut En) Figure 1 - The Microtubule Structure A microtubule is a hollow cylindrical polymer of tubulin. The cylinder is approximately 25 nm across and typically composed of 13 protofilaments in mammalian cells. A protofilament is a linear polymer of tubulin. Figure from Kline-Smith and Walczak (2004). 3 As mentioned previously, microtubules are very dynamic (Desai et al., 1997). They undergo periods of rapid growth where tubulin is added to the plus end before a 'catastrophe' occurs and the microtubule begins to shorten. After a period of shrinkage a 'rescue' can occur and the microtubule will begin to grow again (For an excellent review see Desai and Mitchison 1997). It is thought that these dynamics are in part regulated by the GTPase activity of tubulin. The microtubule is stabilized by a cap of tubulin with GTP at the E site. To this microtubule GTP-tubulin (tubulin with GTP at the E site) may be added. Once bound, the GTP is hydrolyzed to GDP. GDP-tubulin at the plus end of a microtubule, as is thought to occur when the rate of hydrolysis exceeds the rate of polymerization, is not stable and will lead to a catastrophe and shortening. This is thought to occur because the hydrolysis of the GTP to GDP has an allosteric effect on tubulin which introduces a curve into the protofilaments (Howard et al., 1986). This will initiate a period of shrinkage where the protofilaments turn inside out and dissociate from the microtubule (Tran et al., 1997). This occurs until a rescue event, after which the microtubule will enter a period of growth. This is the GTP-cap model first proposed in 1984 (Mitchison et al., 1984). Questions still being examined in this model include the rate of GTP hydrolysis and polymerization, and the amount of GTP-tubulin which must be bound at the plus (+) end of the microtubule to stabilize it. Between periods of growth and shrinkage the microtubule may appear static. However, it may be undergoing a treadmill action (this is also known as a flux state) (Margolis et al., 1998). It is thought this occurs because of the polarity introduced by the asymmetry of tubulin. The two ends have differing affinities for tubulin and consequently the critical concentration will vary (Walker et al., 1988). The critical concentration of 4 tubulin is the concentration above which the microtubule will grow, and below which the microtubule will shrink. Because the critical concentration values vary between ends, it is possible for tubulin to be added to one end, typically the plus, while being lost at the other. This leads to an apparent static microtubule, but it is really dynamic with tubulin dimers being added to one end and lost at the other. This state is typically observed in vitro and is called treadmilling (Margolis et al., 1998). It is believed that treadmilling plays a role in in vivo microtubule dynamics and work is ongoing to demonstrate this (Margolis et al., 1998). The structure of tubulin has been solved allowing a more in-depth explanation of earlier observations. The structure was initially solved by electron crystallography and has since been refined to a 3.5 A resolution (Nogales et al., 1998). a and (3 tubulin are not very similar at the primary level with 40% sequence identity (Nogales et al., 1998). However, at the tertiary level, the two subunits are nearly superimposable (Nogales et al., 1998). The differences are confined to the loops, slight offsets of secondary structures, and amino acid side chains (Nogales et al., 1998). The C-terminus is a region of hypervariability, and it is the main area of difference between the respective isotypes (6 for a, 7 for P) (Luduena 1998). The solution of the tubulin dimer structure led to the identification of three functionally distinct domains. The N-terminal domain is a Rossman fold, and it is where the guanine nucleotide binds. The middle domain is where the majority of microtubule drugs bind, while the relatively short C-terminal domain is where MAPs bind (Nogales et al., 1998). The solution of the structure unequivocally identified the plus end as having the p subunit exposed, while the minus end has the a subunit exposed. It also showed that the N site occurs at the interface between the a and 5 P subunit within one heterodimer while the E site occurs between the P subunit of one dimer and the a subunit of another (Nogales et al., 1998). This, in part, explains why the E site and the hydrolysis of GTP plays such a critical role in the dynamics of the microtubule as the E site is exposed at the plus end of the microtubule. Tubulin is far more complex than described above with multiple isotypes transcribed from different genes, often in a tissue-dependent manner (Luduena 1998). The function for these isotypes is presently under debate (Luduena 1993; Luduena 1998). It does appear that under some circumstances the different isotypes are interchangeable (Bond et al., 1986; Gu et al., 1988) though not in all circumstances (Hoyle et al., 1990; Renthal et al., 1993). More work will be required to determine the roles for each of the isotypes. In addition to the multiple isotypes tubulin undergoes significant post translational modification (Luduena 1998; Westermann et al., 2003). These modifications can include phosphorylation of serine and possibly tyrosine (Eipper 1974; Luduena 1998); acetylation of the epsilon amino group of lysine 40 of several alpha tubulin isotypes (L'Hernault et al., 1985; Piperno et al., 1987; Luduena 1998); tyrosination/detyrosination of the C-terminus of a tubulin (Luduena 1998); the removal of the final two amino acids (tyrosine and glutamate) are removed in a tubulin to form A2-Tubulin (Paturle-Lafanechere et al., 1991; Luduena 1998); polyglutamylation, the addition of multiple glutamate residues to the gamma carboxyl group of a C-terminus glutamate residue, in some isotypes of alpha and beta tubulin (Edde et al., 1990; Mary et al., 1994; Luduena 1998); and polyglycylation, the addition of multiple glycine residues to the gamma carboxyl group of a C-terminus glutamae residues, in a and P tubulin of axenomal microtubules (Rudiger et al., 1995; Luduena 1998). The outcome of these 6 modifications is still being investigated, though acetylation, polyglutamylation and polyglycylation appear to be involved in motility (Luduena 1998; Westermann et al., 2003). In some instances the enzymes responsible for these modifications are still being sought. Further, it is clear that the modifications do not occur to the same degree on both a and p tubulin, with acetylation and tyrosination/detyrosination being observed solely on a and phosphorylation, polyglycylation, and polyglutamylation occurring on various isotypes of both a and P tubulin (Luduena 1998; Westermann et al., 2003). The microtubule network does not occur haphazardly within the cell: the Microtubule Organizing Centre (MTOC) of the cell organizes it. In animal cells, the MTOC is typically the centrosome, which contains two protein structures called centrioles. Remarkably centrioles duplicate in a semi-conservative manner and form two centrosomes such that a bipolar spindle can form during mitosis and each daughter cell will have a centrosome (Wong et al., 2003). A key component of the M T O C is the y tubulin ring complex (y-TuRC) (Zheng et al., 1995), a protein complex that contains y tubulin and serves to nucleate the formation of a microtubule (Zheng et al , 1995). The minus end of the microtubule is capped by the y-TuRC with the positive end extending into the cytoplasm (Zheng et al., 1995). During interphase, the M T O C in animal cells is typically near the nucleus. The MTOCs of the cell separate to form the poles of the mitotic spindle at the onset of mitosis. 1.1.3 - Microtubules and the Cell Cycle The cell cycle consists of four phases: G i , S, G2 and M phase (Figure 2). Collectively G i , S, and G2 form interphase during which the cell grows, replicates its 7 Figure 2 - T h e C e l l C y c l e The cell cycle is composed of four phases: Gi, S, G2, and M. Gi, S, and G2 collectively make up interphase. M phase is further divided into mitosis and cytokinesis. Mitosis can then be separated into five distinct stages: prophase, prometaphase, metaphase, anaphase, and telophase. Figure taken from Karp (1999). 8 genome, and prepares to divide. During M phase the cell undergoes mitosis, itself consisting of 5 phases: prophase, prometaphase, metaphase, anaphase, and telophase. M phase results in the segregation of replicated chromosomes, followed by cytokinesis, the physical division of the cell. The main kinase responsible for entry into mitosis is thought to be the Cdc2/cyclin B complex (Cdc2 is also known as Cdkl) (Nigg 2001). It is generally accepted that this kinase triggers the hallmark events of mitosis such as the condensation of chromosomes and dissolution of the nuclear lamina (Nigg 2001). It may also play a role in the reorganization of the cytoskeleton (Nigg 2001). That said, it is possible to force cells to enter mitosis without Cdc2 phosphotransferase activity indicating more remains to be discovered (Gowdy et al., 1998). The microtubule network, as mentioned, plays a critical role in the cell cycle. During interphase, it is relatively stable with microtubules having a half-life of approximately 200 to 300 seconds (Walker et al., 1988; McNally 1996; Zhai et al., 1996). However, at the start of mitosis, the microtubule network breaks down with a dramatic loss of polymer mass (Zhai et al., 1996) and becomes much more dynamic as it forms the mitotic spindle (for review see (Kline-Smith et al., 2004)). This is reflected by a dramatic decrease in the half-life of microtubules to 15 to 70 seconds (Walker et al., 1988; McNally 1996; Zhai et al., 1996) (Figure 3). This increase in dynamics is thought to help the mitotic spindle find and attach to the chromosomes in a search and capture model. The mitotic spindle contains two types of microtubules (Figure 4) (Kline-Smith et al., 2004). The first are the kinetochore microtubules which bind to the kinetochore of the chromosome. Approximately 25 microtubules will bind to one mammalian kinetochore and associate to form K-fibres (McEwan et al., 1997) that separate the chromosomes 9 G i I S I G 2 I M Cell Cycle Phase Figure 3 - Microtubule Dynamics and Polymer Mass During the Cell Cycle During interphase, microtubules are very stable. However, at the start of M phase the microtubules become much more dynamic. Similarly, at the start of M phase the amount of microtubule polymer rapidly decreases before returning to the interphase level. 10 Astrai . ' M T K-fttwr \u00E2\u0080\u00A2 I ^ B i p o l a r MT r\u00C2\u00ABn*msnrm Chmmnsnms Figure 4 - The Mitotic Spindle The mitotic spindle is composed of kinetochore microtubules, approximately 25 of which associate to form a K-Fiber, and interpolar or overlap microtubules. Kinetochore microtubules are responsible for the separation of the chromosomes during anaphase. Interpolar microtubules separate the centrosome. Astral microtubules localize the centrosome. Figure from (Kline-Smith et al., 2004). 11 during anaphase (McEwan et al., 1997; Kline-Smith et al., 2004). The second are overlap microtubules that are critical for the formation of a bipolar spindle and separation of the centrosomes during mitosis. Overlap microtubules form when the plus end of microtubules that originate from each of the two centrosomes overlap at the midbody of the cell. These overlapped microtubules are cross-linked by a plus-end directed motor that forces the centrosomes apart. Finally, in the mitotic cell there are astral microtubules that orient the location of the spindle pole (Kline-Smith et al., 2004). The dynamics of the microtubules are critical for the proper progression of mitosis (Kline-Smith et al, 2004). If these dynamics are perturbed by small molecule drugs the cells will typically block at the metaphase stage of M-phase with high Cdc2/cyclin B activity. Many chemotypes have been found that block cells in metaphase (Hamel 1996; Jordan et al., 1998). The small molecules that perturb microtubule dynamics and lead to mitotic arrest are termed microtubule poisons or antimitotic agents. While many are known in the laboratory, two families have been brought to the clinic as a very successful chemotherapeutic agents for cancer treatment, the vinca alkaloids and the taxanes (paclitaxel and taxotere) (Rowinsky et al., 2001). 1.1.4- Microtubule Poisons As discussed previously, the correct dynamics of the microtubule network is critical to the proper progression of the cell cycle (Kline-Smith et al., 2004). In fact, if a small molecule perturbs these dynamics, the cell will block in mitosis. This property has been exploited to produce two families of successful anti-cancer drugs: the vinca 12 alkaloids and the taxanes. To understand how these work it is necessary to first examine the effects of microtubule poisons on the microtubule network. There are two large classes of microtubule poisons (Hamel 1996; Jordan et al., 1998). The first, with over 100 members, comprises those that inhibit the polymerization of microtubules and/or produce a breakdown of the microtubule network. The vinca alkaloids are an example of such agents that are used in the clinic for treatment of testicular cancer, Wilms tumour, Acute Lymphocytic Leukemia, lymphoma, rhabdomyosarcoma, soft tissue sarcoma, neuroblastoma, and Non-Small-Cell Lung Cancer (Rowinsky et al., 2001). Nocodazole and colchicine are two examples that are commonly used in the laboratory. The second much smaller group, with approximately 10 members, comprises agents that promote the polymerization of tubulin into microtubules and stabilize existing microtubules. Here, the most well known example are the taxanes (paclitaxel and docetaxel) which are commonly used in the clinic for treatment of Ovarian, Breast, and Non-Small-Cell Lung Cancer as well as Kaposi's sarcoma (Rowinsky et al., 2001), as well as in the laboratory. There exist at least five drug-binding sites on tubulin that can disrupt the microtubule network (Hamel 1996; Altmann 2001). The first is the colchicine-binding site, the second is the vinca domain, the third is the rhizoxin/maytansine, the fourth is the taxoid domain, and the fifth is the sulfhydryl group of assembly critical cysteine residues. The majority of examined microtubule poisons bind at one of these sites. Those binding at the colchicine, vinca, rhizoxin/maytansine or cysteine domains typically inhibit polymerization of tubulin into microtubules, while those binding at the taxoid domain promote the polymerization of microtubules. There is also a peptide binding site on (3 13 tubulin that overlaps with the vinca domain (Mitra et al., 2004). Naturally occurring peptides such as hemiasterlin and doastatin 10 bind in the peptide binding site (Mitra et al., 2004). Other binding sites probably exist as evidenced by laulimalide and peloruside that promote microtubule polymerization but do not compete with paclitaxel for binding to the microtubule (Pryor et al., 2002; Gaitanos et al., 2004). The taxoid site is of particular interest. While still debated, it is thought that the taxoid binding site is only present once the microtubule has formed (Parness et al., 1981; Takoudju et al, 1988). The solution of the microtubule and tubulin structure has allowed for the determination of the location of the taxoid binding site on the interior of the microtubule (Nogales et al., 1998; Li et al., 2002). While diffusion from the ends is possible, this likely would occur slowly and may be blocked by MAPs associated with microtubule ends. Holes have been identified in the wall of the microtubule which might allow rapid diffusion of the taxanes into the interior of the microtubule (Li et al., 2002). While it was initially thought that all microtubule-disrupting agents worked by changing the overall amount of tubulin that was polymerized, a revolution in thinking has begun to occur. It is now thought that these agents work at doses lower than those that affect the mass of polymerized tubulin by decreasing the dynamics, particularly the rescue and catastrophe frequency (Jordan et al., 2004). This decrease in dynamics prevents the formation of the mitotic spindle leading to a mitotic arrest. 1.1.5 - Microtubules and MAPs Microtubules do not exist alone. They have many proteins associated with them that are, in general called MAPs. Some of these proteins, such as members of the kinesin 14 and dynein family, are motors that move along the microtubule and carry cargo with them (Hirokawa et al., 1998; Endow 2003; Vallee et al., 2003). Others, such as XMAP215 (Kinoshita et al., 2002), XMAP230 (Andersen 2000), Tau, MAP4 (Cassimeris 1999; Andersen 2000), MAP2, Opl8/Stathmin (Cassimeris 2002), XKCM1 (Andersen 2000), and katanin (McNally et al., 1993) affect the stability of the microtubule. If microtubules are highways, then kinesin and dynein are the cargo carriers. These motors typically use ATP hydrolysis to power their movement and will move either toward the plus end, as is the case for most kinesin motors, or toward the minus end, in the case of dynein and at least one kinesin (Hirokawa et al., 1998). They carry vesicles and also play a role in the separation of the centrosomes during mitosis (Hirokawa et al., 1998). Motors attach to the overlap microtubules, and each is plus-end directed and will force the centrosomes, where the minus end is attached, to separate. A large number of MAPs, including XMAP215, XMAP230, Tau, MAP4, XKCM1 and katanin, regulate the dynamics of microtubules either in vitro or in vivo. On the one hand, XMAP215, XMAP230, Tau, and MAP4 all increase the length of the microtubules, although by different mechanisms. XMAP215 increases the rate of growth but does not change the rate of catastrophe or rescue (Andersen 2000). XMAP230, Tau and MAP4 stabilize the microtubule and reduce the frequency of catastrophe rather than increase the rate of growth (Andersen 2000). On the other hand, Opl8/stathmin and XKCM1 destabilize the microtubules. Op 18 has been shown to disrupt microtubules in two distinct manners. First, it is possible for Op 18 to sequester tubulin dimers which has the effect of lowering the concentration of tubulin below the critical concentration. It is 15 also possible for Op 18 to actively induce catastrophe (Cassimeris 2002). XKCM1 has only been shown to induce catastrophe (Andersen 2000). Unlike the other MAPs that work at the microtubule ends, katanin activity occurs in the middle of the microtubule where it cleaves microtubules in an ATP-dependent manner (McNally et al., 1993). Katanin has been shown to be active during M phase in Xenopus extracts and is likely involved in the rapid breakdown of the microtubule network and the onset of mitosis (McNally et al., 1998). The MAPs are crucial for the correct dynamics of the microtubule network (Cassimeris 1999). In fact, the in vitro dynamics of purified tubulin are not similar to those observed in vivo. However, correct in vivo dynamics can be reconstituted if XMAP215 and XKCM1 are added at the correct concentration to tubulin in vitro (Kinoshita et al., 2001). MAPs can be controlled via signaling pathways within the cell. They can be phosphorylated which in turn changes their effect on microtubule dynamics (Cassimeris 1999). 1 . 2 - M A P K Signaling 1.2.1 - Mitogen Activated Protein Kinase (MAPK) Protein phosphorylation plays a vital role in the regulation and signaling of the cell. Phosphorylation changes occur in response to a diverse range of stimuli from growth factors to stress. The outcomes of this phosphorylation can be just as varied ranging from proliferation to differentiation. One of the major family of kinases responsible for this signaling is the M A P K family (Roux et al., 2004). This family of kinases has at least five members: Erkl/2, Jnk/SAPK, p38 MAPK, Erk3/4 and Erk5 (Roux et al., 2004). There are 16 examples of MAPK-like families conserved in organisms from yeast to humans. These kinases are grouped as a family because they are activated following a kinase cascade that results in phosphorylations at two critical threonine and tyrosine residues in the consensus sequence T X Y on the M A P K (Roux et al., 2004). This consensus sequence occurs in the activation or T-loop of the MAPK. M A P K Kinase (MAPKK/MKK), also known as a M A P K Kinase (MEK), is responsible for the addition of the activating phosphorylations on M A P K (Roux et al., 2004). M A P K K are dual specificity kinases that add phosphate to both the threonine and tyrosine residue of the activation loop. Similar to the MAPK, the M A P K K (MEK) must be phosphorylated to become activated. The kinase responsible for this phosphorylation is a M A P K Kinase Kinase (MAPKKK). MAPKKKs are typically activated at the plasma membrane by receptor tyrosine kinases (RTKs) or G-proteins (Roux et al., 2004), though there appears to be alternative methods of activation such as microtubule disruption (Stone et al., 2000; Zang et al., 2001). A schematic of MAPK activation is shown in Figure 5. Erkl/2 were the first M A P K to be discovered in mammals, and in fact were originally called MAP2K since it phosphorylated microtubule-associated protein 2 (Ray et al., 1987). Erkl/2 are kinases of approximately 44 and 42 kDa, respectively. The M A P K K for Erkl/2 is MEK1/2, while the M A P K K K for Erkl/2 is the Raf family of kinases (A-Raf, B-Raf, 1-Raf) (Roux et al., 2004). The Raf kinases are activated in response to phorbol esters or growth factors. The exact mechanism of activation is unknown but it does include a series of phosphorylations, the binding of Ras, and recruitment to the plasma membrane (Chong et al., 2003). Erkl/2 has many substrates including membrane proteins, nuclear proteins, cytoskeletal proteins and MAPK-17 M A P K K K v. ) 1 MAPKK \u00E2\u0080\u00A2 \u00E2\u0080\u00A2 MKP Figure 5 - M A P Kinase Activation and Deactivation Pathways Phosphorylation of M A P K K by M A P K K K leads to the activation of M A P K K . In turn, M A P K K phosphorylates M A P K on a threonine and tyrosine residue in the consensus sequence of T X Y . This is the activation loop of the MAPK. In turn, the dual specificity MKP removes these phosphorylations resulting in an inactivation of MAPK. 18 Activated Protein Kinases (MKs) (Roux et al., 2004). These substrates often change transcription either directly or indirectly. Erkl/2 plays a role in activating the ribosomal S6 kinases (RSKs), the M A P K and stress activated protein kinases (MSKs), and the MAPK interacting kinases (MNKs). These activated kinases in turn affect many cellular processes including transcription (Roux et al., 2004). p38 M A P K has four isoforms: a, (3, y and 5. These isoforms are activated in response to stress rather than growth factors and phorbol esters as is the case with Erkl/2. The M A P K K for the p38 family is MKK3/6 while the number of M A P K K K for the p38 family is high and includes MEKKs 1 to 4, MLK2, MLK3, D L K , ASK1, Tpl2, and Tak2. Similar to Erkl/2, p38 M A P K has many substrates including phospholipase A2, Tau, ATF1, ATF2, MEF2A Sap-1, Elk-1, NF-KB, Ets-1 and p53. The MKs MSKs and MNKs are also activated. However, unlike Erkl/2, p38 M A P K do not phosphorylate RSKs (Roux et al., 2004). JNK/SAPK family contains ten isoforms that are activated in response to stress, similar to p38 M A P K (Davis 2000). The MAPKKs for JNK/SAPK include MEK4 and MEK7 (Davis 2000). MEK4 is interesting, as it is known to activate p38 M A P K in vitro (Roux et al., 2004). This provides a possible point of cross talk and a possible explanation for how stress may lead to the activation of both p38 M A P K and JNK/SAPK. An additional potential location for cross-talk is at the M A P K K K level. The known M A P K K K for JNK/SAPK include MEKK1-4, MLK2, MLK3, Tpl-2, D L K , Taol, Tao2, TAK1, ASK1, and ASK2 (Roux et al., 2004). Many of the known MAPKKKs for JNK/SAPK have also been shown to be MAPKKKs for the p38 M A P K pathway. There is one large difference between JNK/SAPK, and p38 M A P K and Erkl/2. There is no 19 known M K that JNK/SAPK activates. However, it does phosphorlyate and activate a number of transcription factors including the most well known, c-Jun (Roux et al, 2004). MAPKs obviously play a large role in signaling. The length of the signal is crucial. It has been shown in the same cell line that the duration of activation can mean the difference between a signal to proliferate and a signal to differentiate (Marshall 1995). 1.2.2 - M A P K and the Cytoskeleton Erkl/2 in mammals was originally identified as a kinase capable of phosphorylating MAP2 (Ray et al., 1987). Despite early investigations, no association with microtubule was observed. However, a paper published in 1995 used multiple Erkl/2 antibodies to observe that upwards of 50% of Erkl/2 in a mammalian cell was associated with the microtubule network (Reszka et al., 1995). In fact, even after stimulation, nearly 50% of the active Erkl/2 remained associated with the microtubule (Reszka et al , 1995). There is a body of evidence that hints that Erkl/2 signaling may be involved in microtubule stability (Reszka et al., 1995).Work has demonstrated a role for MAPK pathways in cell migration (Huang et al., 2004) which requires a rearrangement of the cytoskeleton. Recent work has hinted that it might be necessary to think of microtubules in a different manner. It has been shown that signaling can result in microtubule rearrangement. However, it has also been repeatedly shown that disruption of the dynamics of the microtubule can lead to the activation of the Erkl/2 and other M A P K pathways (Stone et al., 2000; Zang et al., 2001; Boldt et al., 2002). This activation 20 appears to be cell type specific, and the reason is not known. Nevertheless, it may be necessary to reconsider the microtubules not only as a target of signals but also as a 'receptor' capable of initiating a signaling response. In addition to being regulated by M A P K signaling, there is evidence that the microtubule network of the cell is controlled during the cell cycle. The evidence suggests that this control is via phosphorylation of regulatory proteins (Cassimeris 1999). Initially, it was found that active Cdc2/cyclin B could alter the lengths of microtubules in Xenopus egg extracts. Further, MAP4 phosphorylation deficient mutants block entry into mitosis and there are at least four required phosphorylations on Op 18 for mitosis to occur (Cassimeris 1999). 1.2.3 - MAP Kinase Phosphatase (MKP) As discussed, MAP Kinases must be phosphorylated at their T-loop in order to be activated (Pearson et al., 2001). In particular, there must be both a threonine and a tyrosine phosphorylated at the T X Y consensus sequence. To be deactivated, either the threonine or the tyrosine, or both residues must be dephosphorylated. As the length of the activation can have profound consequences it is probable that dephosphorylation is critical. In vitro it has been shown that a protein-serine/threonine (Alessi et al., 1995) or tyrosine (Pettiford et al., 2000) phosphatase can remove a phosphate from these residues in M A P K leading to its inactivation. However, it has been discovered that M A P K dephosphorylation is typically performed by dual specificity phosphatases (DSP) that dephosphorylate both threonine and tyrosine (Theodosiou et al., 2002; Farooq et al., 2004). These DSPs have come to be known as MAP Kinase Phosphatases (MKPs). There are now 10 known MKPs which have differing M A P K specificity (Theodosiou et al., 21 2002). The majority of the activity assays have been conducted in vitro. More work will be required to determine which substrates are physiologically relevant to which MKPs. However, some interesting results have been obtained. It has been shown that the binding of Erkl/2 results in the activation of MKP-3 (Camps et al., 1998). Substrate binding at a location distinct from the catalytic residue leading to activation may be critical for the regulation of M A P K signaling. There has also been work on the catalytic residue of MKPs. It has been found that the sulfhydryl of a cysteine residue is the catalytic residue (Denu et al., 1995). 1.3 - Goals The goal of this thesis has been the characterization of the biology of three new microtubule poisons. The first two, ceratamine A and B, are structurally very related. They were discovered in an antimitotic screen and their effect on microtubules in vivo, microtubules in vitro, and the cell cycle has been examined. The third molecule is dimethyl varacin (DMV). D M V was also discovered in an antimitotic screen. Work with D M V has concentrated on its effect on microtubules, both in vivo and in vitro as well as its effect on the cell cycle, M A P K signaling, and cellular phosphorylations. 22 Chapter II: Materials and Methods 2.1 - Materials: Unless otherwise noted, all materials were purchased from Sigma-Aldrich Chemical Company, St. Louis, Missouri. 2.2 - Antimitotic Assay - The phenotypic antimitotic ELICA was performed as described previously (Roberge et al., 2000). 2.3 - Antiproliferation Assay - The antiproliferation assay was performed as described previously (Anderson et al., 1997). 2.4 - In Vitro Tubulin Polymerization Assays 2.4.1 - Promotion of Tubulin Polymerization: On ice, 35 uL of buffer A (80mM PIPES pH 6.8, 1 mM MgCl 2 , 1 mM EGTA) were added to each well of a half-area 96-well plate, followed by 5 uL of 10 mg/mL bovine brain tubulin (TL238, Cytoskeleton Inc. Denver, Colorado) in buffer A containing 10% glycerol, 5 uL of a lOx solution of ceratamine A or B in buffer A, and 5 uL of 10 mM GTP in buffer A. The optical density at 340 nm was followed at 32\u00C2\u00B0C over time. 2.4.2 - Inhibition of Tubulin Polymerization: This assay was performed in a similar manner to the promotion of polymerization assay but with the following differences. On ice, 25 uL of buffer A plus 10% glycerol were added to each well of a half-area 96-well plate, followed by 15 uL of 10 mg/mL tubulin in buffer A containing 10% glycerol, 5 uL of lOx ceratamine A or B solution in buffer A containing 10% glycerol and 5 uL of 10 23 mM GTP in buffer A plus 10% glycerol. The optical density at 340 nm was followed at 37\u00C2\u00B0C over time. 2.4.3 - Negative Staining Electron Microscopy: The promotion of tubulin polymerization assay was performed as described above. Ten uL of the reaction (10 pg of protein) were removed and placed on a Formar coated copper grid and left for 1 minute at room temperature. The fluid was wicked off by Whatman filter paper and 10 uL of 2% uranyl acetate solution were added and immediately removed by Whatman filter paper wicking. The sample was analyzed by electron microscopy using a Hitachi H7600 Transmission Electron Microscope. (University of British Columbia, Bioimaging Facility). 2.4.4 - 3H-Paclitaxel Competition Binding Assay: The paclitaxel binding assay was conducted similar to that described previously (Pryor et al., 2002). Briefly, solutions of 10 uM H-paclitaxel (NCI) with a specific activity of 1600 dpm/pmol, various concentrations of trial compound [20 uM eletherobin, 100 uM or 250 uM ceratamine A], DMSO (20%o), and monosodium glutamate (0.75M) were prepared. A solution of microtubules was prepared by incubating tubulin (2.5 uM; 0.25 mg/mL), 2',3'-dideoxyguanosine 5'-triphosphate (ddGTP; 25 uM) and monosodium glutamate (0.75 M). These solutions were incubated for 30 minutes at 37\u00C2\u00B0C. Twenty five pL of drug solution were added to 100 uL of microtubule mixture for a final volume of 125 uL, and the mixture was incubated for 30 minutes at 37\u00C2\u00B0C. Following incubation, 50 uL of reaction mixture were removed, centrifuged for 20 minutes at 13,000 rpm and the supernatant was counted in a liquid scintillation counter to determine the amount of unbound, or free, H-paclitaxel. Also, 50 uL of uncentrifuged reaction mixture were 24 counted to determine the total amount of H-paclitaxel. The statistical significance of the results was determined by a Student's t test. 2.5 - Flow Cytometry Analaysis of the Cell Cycle: Analaysis of mitosis-specific markers was conducted as described (Rundle et al., 2001) using a FACS Calibur and GF-7 primary antibody, with the following minor modification: RNase was from Qiagen and used at a final concentration of 100 ug/mL. GF-7 primary antibody was raised using paired helical filaments from Alzheimer brains. While several bands are detected by Western blot (data not shown), the epitopes are typically only observed during mitosis. Data were analyzed using MDI 2.8. 2.6 - Immunofluorescence Microscopy 2.6.1 - f-Tuhulin Immunofluorescence Microscopy of DMV Treated Cells: 1 mL of a 1/10 dilution of 80% confluent Swiss-3T3 cells was seeded onto glass coverslips in 12-well microplates and incubated 48 hours to allow the cells to attach. The cells were treated with 10 uM D M V or 0.1% DMSO for either 30 or 120 minutes. The microtubules were visualized using the E7 monoclonal P-tubulin antibody (Hybridoma Development Bank, University of Iowa) as described in Anderson et al (Anderson et al., 1997). DNA was not stained. 2.6.2 - f-Tubulin Confocal Immunofluorescence Microscopy of Ceratamine Treated Cells: MCF-7 cells cultured on glass coverslips were treated without or with ceratamine A and microtutubles were immunostained using the E7 monoclonal p-tubulin antibody. A 25 stack of 40 optical sections each 0.35 pm was collected using a Bio-Rad Radiance Plus confocal microscope and analysed using ImageJ software (NIH). 2.6.3 - P-Tubulin Immunofluorescence Microscopy of Ceratamine Treated Cells: MCF-7 cells cultured on glass coverslips were treated without or with ceratamine and microtutubles were immunostained using the E7 monoclonal P-tubulin antibody (Anderson et al., 1997). DNA was stained using Hoechst 33248. 2.7 - Kinase Assays 2.7.1 - Cdc2 Kinase Assay: MCF-7 mp53 cells were plated in 100 mm dishes and grown to approximately 80% confluence. The cells were treated with 300 nM nocodazole, 0.1 % DMSO, and 10 uM D M V for 5 hours before harvesting by trypsinization. The assay was then carried out as described previously (Gowdy et al , 1998). 2.7.2 - In Gel Kinase Assay: Following treatment by either D M V at 1, 5, or 10 uM; 300 nM nocodazole; or 0.1 % DMSO for 2 hours, an in gel kinase assay was carried out as described elsewhere (Gowdy et al , 1998). The following changes were made. The lysis buffer contained neither DNase I nor RNase A. Further, the gel contained either no polymerised substrate to detect possible autophosphorylation, or 0.2 mg/mL myelin basic protein (MBP) from Invitrogen. Finally, the gel was visualized by autoradiography. 2.8 - Western Blots: Ten ug of protein from a whole cell lysate were separated on a 10% SDS-PAGE. A semi-dry transfer (Biorad; Transblot SD) was used to transfer to a PVDF membrane in Towbin buffer. The membrane was washed in dt^O prior to being blocked with a solution of 5% non-fat milk in TBS containing 0.1% Tween-20 (5%-TBS-T). The 26 membrane was incubated with primary antibody [Phospho-p38 M A P K T180/Y182; Cell Signaling Technology; Phospho-p44/42 MAP Kinase (T202/Y204), New England Bioloabs; p44/42 MAPK, Stressgen; Phospho-JNK/SAPK, Cell Signalling Technology] at the manufacturers' recommended dilutions in 5%-TBS-T for Phospho-p44/42 MAPK, Total Erkl/2 (p44/42 M A P K antibody), and phospho-JNK/SAPK or TBS containing 5% bovine serum albumin (BSA) and 0.1% Tween-20 for phospho-p38 MAPK. The blots were incubated overnight at 4\u00C2\u00B0C. The membranes were washed in TBS-T prior to being incubated in the appropriate secondary antibody at the manufacturers' recommended dilution in 5%-TBS-T. The membrane was washed extensively before being developed using Enhanced Chemiluminesence (Pierce, SuperSignal West Pico). 2.9 - Kinexus Kinetworks\u00E2\u0084\u00A2 Screens: Cells at approximately 80% confluency in three 100 mm tissue culture plates were treated with stimuli (0.05% DMSO or 5 uM DMV) for 2 hours. Cells were collected, pooled and lysed as described in the Kinexus Bioinformatics Corp. (Vancouver, Canada) sample submission form. For each Kinetworks\u00E2\u0084\u00A2 phospho-site screen (KPSS 1.3, KPSS 2.0, KPSS 3.0, KPSS 4.0), 500 uL of lysate at 1.0 mg/mL in loading buffer were submitted to Kinexus Co. 2.10 - Protein Phosphatase 2A Assay - Approximately 0.125 U (1 unit releases 1 nmol of phosphate per minute from 15 uM phosphorylase A at 30\u00C2\u00B0C) of PP2A from Upstate Biotechnology were added to 10 uL of enzyme dilution buffer (20 mM MOPS, pH 7.5, 0.15 M NaCl, 60 mM p-mercaptoethanol, 1 mM MgCl 2 , 1 mM DTT, 10% glycerol, 0.1 mg/mL BSA). To this solution, 10 uL of assay buffer (10% w/v glycerol, 0.1 mg/mL BSA) containing 25 uM or 2.5 uM D M V were added for a final volume of 20 uL. This 27 was left on ice for 5 minutes. Previously, 5 uL of peptide solution (0.75 mg/mL RRAT-Phosphate-VA-amine in phosphate free water) were added to 25 uL of assay buffer in a half-area 96 well plate and incubated at 37\u00C2\u00B0C for 30 minutes. The enzyme solution containing the inhibitors was then added to the peptide solution for a final volume of 50 uL. This was incubated for 45 minutes at 37\u00C2\u00B0C. Fifty uL of Malachite Green (1.05% w/v ammonium molybdate, 0.034% malachite green, 1 M HCI) was added and the absorbance measured at 595 nm. Controls were also performed with no enzyme present or no substrate present. 28 Chapter III: Identification of Ceratamines A and B as Microtubule-Stabilizing Agents With Unusual Biological Properties 3.1 - Specific Introduction: The number of microtubule-stabilizing chemotypes is on the order of 10 with the most well known example being the taxanes, including paclitaxel (Figure 6A) and its derivative taxotere (He et al., 2001). Other examples include the eleuthesides (Long et al., 1998) , laulimalides (Mooberry et al., 1999), GS-164 (Shintani et al., 1997), epothilones (Bollag et al., 1995), discodermolides (ter Haar et al., 1996), sarcodictyins (Hamel et al., 1999) , peloruside (Hood et al., 2002) and dictyostatin (Isbrucker et al , 2003). Discodermolide, eleutherosides, sarcodictyins and epothilones compete with paclitaxel for binding at the taxoid binding site (Bollag et al., 1995; Kowalski et al., 1997; Hamel et al., 1999), but laulimalide (Figure 6B) and peloruside A (Figure 6C) have been unequivocally shown to bind at a site distinct from paclitaxel on the microtubule (Pryor et al., 2002; Gaitanos et al., 2004). Significantly, all of these microtubule-stabilizing agents except GS-164 have come from natural products and are very structurally complex (Figure 6). This is exemplified by paclitaxel which was isolated from the Pacific Yew tree (Taxus breviofolia) and has 11 chiral centers. This structural complexity led to large obstacles in the development of paclitaxel as a chemotherapeutic agent (Cragg et al., 1993). It is still made semi-synthetically though a total synthesis has been reported (Kingston 2000). As such, structurally simple antimitotic agents would be attractive lead 29 A C Figure 6 - The Chemical Structures of Paclitaxel (A), Laulimalide (B), and Peloruside A (C) Paclitaxel, laulimalide and peloruside A promote the polymerization of microtubules and microtubule bundling, but laulimalide and peloruside A cannot compete with paclitaxel for binding to microtubules. 30 compounds as they should be relatively simple to supply. There is also a pressing need to develop new antimitotic agents as cells can develop resistance to paclitaxel. Resistance may take the form of different tubulin isotypes (Verdier-Pinard et al., 2003), point mutations in tubulin (Verdier-Pinard et al., 2003), and expression of drug efflux pumps (Yusufet al., 2003). A cell-based phenotypic assay for antimitotic agents was used to screen extracts from natural sources. An extract from the marine sponge Pseudoceratina sp. collected in Papua New Guinea showed activity in the assay. The active compounds were isolated by chromatographic steps using the assay to guide purification, and they were identified as ceratamines A and B (Figure 7) by analysis of their mass spectrometry and nuclear magnetic resonance data (Manzo et al., 2003). 3.2-RESULTS 3.2.1 - Effect of Ceratamines A and B on Cell Cycle Progression: The effect of ceratamines A and B on cell cycle progression was investigated using a dual labeling flow cytometry procedure that can identify the Gi , S, G 2 and M cell populations. Untreated cells showed a cell cycle profile typical of asynchronously proliferating cells, with large proportions of Gi, S and G2/M (Figure 8A). Exposure of cells to 21 uM ceratamine A or 22 uM ceratamine B for 16 h caused an almost complete disappearance of the Gi and S peaks with a corresponding large increase in the G2/M peak (Figure 8A). Staining with the GF-7 antibody, a phospho-specific antibody that is typically recognized only at mitosis, to distinguish between G 2 and M showed the arrest to be at M and not G 2 : ceratamine A treatment caused 74% of the cells to arrest at M and 31 Figure 7 - The Chemical Structure of Ceratamine A and B 32 DMSO Ul J8 Nocodazole \u00C2\u00A7 m Ceratamine A Ceratamine B G1 ^ ^ 2 / M DMA (PI) DNA (PI) I i Ul : 1 J i DNA (PI) L i DNA (PI) \u00C2\u00A7 U-O B M G2 GI S DNA (PI) 5 1 < LL o DMA (PI) I5 LL. * o DNA (PI) DNA (PI) Figure 8 - Flow Cytometry of MCF-7 mp53 Cells Treated with DMSO, Nocodazole, Ceratamine A or Ceratamine B (A) DNA profile of cells stained with propidium iodide. (B) GF-7 versus DNA dual labeling experiment. This permits the differentiation of G2 and M phase cells. The colours represent the number of events. As one moves through the spectrum from blue to red, the number of events increases. 33 ceratamine B caused 81% M arrest, (Figure 8B). The extent of M phase accumulation was very similar to that observed with the microtubule-targeting agent nocodazole (Figure 8B). 3.2.2 - Antimitotic and antiproliferative activity of ceratamines. To examine the concentration dependence of M arrest by ceratamines, cells were treated with different concentrations of compound and the proportion of cells in M was determined by staining the DNA and counting the number of cells with condensed mitotic chromosomes in the microscope. As shown in Figure 9A, M arrest was concentration-dependent, with half-maximal arrest at 12 uM for ceratamine A and 21 uM for ceratamine B. The IC50 results do not completely agree with those obtained with flow cytometry. The IC50 for ceratamine B is very similar to the concentration used in the flow cytometry experiment, where >80% of cells were blocked in mitosis. The reason for this discrepancy is unclear, but possible explanations may be differing batches. Also, IC50 values are approximations from the curve in Figure 9A. As expected, the compounds also inhibited cell proliferation, with IC50 of 2 uM for ceratamine A and 3 uM for ceratamine B (Figure 9B). 3.2.3 - Effect of ceratamines on microtubule organization in vivo. We next examined the effect of ceratamines on the morphology of interphase and mitotic microtubules using immunofluorescence microscopy with a P-tubulin antibody and counterstaining of DNA. In ceratamine-treated cells arrested in M , the chromosomes were condensed but failed to align at the cell equator (Figure 10B). Examination of microtubules showed that instead of containing a bipolar mitotic spindle as in control 34 Ceratamine (uM) Ceratamine (uM) Figure 9 - Antimitotic and Cell Proliferation Assays with Ceratamine A and B Antimitotic (A) and Cell Proliferation Assays (B) with various concentrations of either ceratamine A (open circles) or ceratamine B (closed circles). The percent of mitotic cells was determined by mitotic spreads after 16 hours with various concentrations of ceratamine. Cell proliferation was determined by MTT assays. Both assays used MCF-7 mp53 cells. Antimitotic assay was conducted once though the results are supported by flow cytometry. Cell Proliferation Assays were conducted 3 times; error bars represent the standard deveiation. 35 mitotic cells (Figure 10A), cells arrested in M by ceratamine treatment had no microtubules. Instead, multiple foci of P-tubulin staining were observed (Figure 10B and Figure 11). The number of foci varied from cell to cell from three to as high as about thirty. Not all foci had tubulin staining of equal intensity nearly all cells containing a few that were clearly brighter than the rest. Examination of the structure of these foci by three-dimensional confocal microscopy showed that they consisted of pillar-like structures that extend vertically from the glass surface upon which the cells were attached (Figure 1 IB and Figure 1 IC). The intensely staining foci spanned the entire thickness of the cell whereas less intensely staining foci had a similar structure but did not span the entire thickness of the cell. Focus formation occurred rapidly. Incubation with 21 uM ceratamine A for 30 min was sufficient to induce multiple centers of intense microtubule staining. Interestingly, some cells with apparently normal microtubules were observed but they were all at stages later than metaphase, indicating that ceratamines might only affect microtubules at early stages of mitosis. The focal tubulin staining persisted for more than ten hours in the presence of ceratamines. Longer incubation times caused a decrease in the number of tubulin foci, such that cells often had either two or four regions of heavy tubulin staining. The effect of ceratamines on mitotic microtubules was markedly different from that of the microtubule polymerizing agent paclitaxel, which caused the appearance of dense, \"bushy\" mitotic spindles (Figure 10C). Ceratamine A treatment also changed the organization of microtubules in some interphase cells. In these cells, ceratamine A caused the appearance of a pronounced perinuclear microtubule network with a distinct depletion of peripheral microtubules 36 Tubulin DNA No drug Ceratamine A Paclitaxel Figure 10 - Confocal P-Tubulin Immunofluorescence Images of Ceratamine or Paclitaxel Treated Cells Images of P-tubulin staining and DNA (Hoescht 33248 staining) were collected of mitotic cells treated with (A) DMSO, (B) ceratamine A, or (C) paclitaxel. Images were collected of P-tubulin staining in interphase cells treated with (D) DMSO or (E) ceratamine A. 37 Figure 11 - Cross Section p-Tubulin Staining in Mitotic Cells P-Tubulin staining in mitotic cells treated with (A) DMSO or (B,C) ceratamine A were collected by confocal microscopy. Image J software was used to reconstitute cross-sectional images. 38 (Figure 10E). In addition, the microtubule network of ceratamine-treated cells appears disorganized compared to that of control cells (Figure 10D). 3.2.4 - Effect of ceratamines on microtubule formation in vitro Most antimitotic agents interfere with microtubule function by interacting directly with them and causing their depolymerization or their overpolymerization. To examine the mechanism of action of ceratamines, we examined the effect of these compounds on the polymerization of microtubules from pure tubulin in vitro. We first determined whether ceratamines inhibit microtubule polymerization. As shown in Figure 12A, 10 uM nocodazole prevented the assembly of microtubules but ceratamines showed no significant inhibition of tubulin polymerization at any concentration tested (up 110 uM). We next examined whether ceratamines can promote microtubule polymerization. As shown in Figure 12B, 10 uM paclitaxel stimulated microtubule polymerization, as expected. Ceratamine A (100 uM) also promoted microtubule polymerization (Figure 12B) but to a lesser extent than paclitaxel. To determine whether the increase in absorbance measured in this assay was indeed due to the formation of microtubules, a sample was removed and analyzed by negative staining transmission electron microscopy. As shown in Figure 13A and Figure 13B, ceratamine A caused the formation of microtubules, similar to those observed for the paclitaxel treated sample (Figure 13C). No microtubules were observed in samples treated with DMSO, the drug solvent (not shown). 39 A 0.3 0.2 0.1 0.0 0 5 10 15 Time (min) B 0.12 0.10 0.08 J 0.06 \u00C2\u00A7 0.04 0.02 0.00 -0.02 Figure 12 - In vitro Tubulin Polymerization Assays with Ceratamine A and B (A) Inhibition of tubulin polymerization with DMSO (open squares), nocodazole (open triangles), ceratamine A (open circles) and ceratamine B (closed circles). The final tubulin concentration was 3 mg/mL; the positive control was nocodazole. (B) Promotion of tubulin polymerization assay with DMSO (open squares), paclitaxel (closed triangles), ceratamine A (open circles), and ceratamine B (closed circles). The final concentration of tubulin was 1 mg/mL; the positive control was paclitaxel. Similar results for both (A) and (B) were obtained in a second experiment. 5 10 15 Time (min) 40 Ceratamine A - 30,000x Paclitaxel - 120,000x Figure 13 - Electron Microscopy Images of Microtubules A portion of the products from the promotion of tubulin polymerization assay was examined by negative staining electron microscopy. Paclitaxel was a positive control with clear microtubules present. The bar represents 500 nm. 41 3.2.5 - Ceratamines do not compete with Paclitaxel for binding to Tubulin: To probe the molecular mechanism by which the ceratamines promote the polymerization of tubulin into microtubules, a competition assay with radiolabeled paclitaxel was conducted. This was done to determine if the ceratamines bind to the taxoid binding site of tubulin. In the assay, microtubules are initially formed by incubation with ddGTP and a high concentration of monosodium glutamate (MSG). These conditions result in the almost complete polymerization of available tubulin. A solution containing the\" radiolabeled paclitaxel and the compound in question, in this case ceratamine A, is then added to the microtubule solution. After 30 minutes to come to equilibrium a portion of the total reaction is counted in a scintillation counter. A second portion is centrifuged. This causes the microtubules and the small molecules bound to them to pellet at the bottom of the tube. The supernatant, containing unbound small molecules, is counted in a liquid scintillation counter. With both counts, it is possible to obtain a ratio of free to total paclitaxel. If the compound in question competes with paclitaxel binding, the amount of free paclitaxel will increase in the presence of the compound. This will be reflected by an increase in the ratio of free to total paclitaxel. Using this assay, developed by Hamel (Kowalski et al., 1997), it was discovered that ceratamine A does not compete with paclitaxel for binding to microtubules at concentrations up to 100 uM (Figure 14). As can be seen in Figure 14 when no drug is present in the binding assay, the ratio of free to total paclitaxel is 0.22\u00C2\u00B10.04. The ratio with 4 uM eleutherobin, a known competitor of paclitaxel binding (Hamel et al., 1999), is 0.40\u00C2\u00B10.06. However, with 20 uM ceratamine A the ratio is 0.19\u00C2\u00B10.02 and is not significantly different from that found 42 in the absence of drug. Even when the amount of ceratamine A is increased to 50 uM, a 25-fold excess over the amount of tubulin and paclitaxel, the ratio is 0.26\u00C2\u00B10.03, again a non-significant difference. This is clear evidence that ceratamine A is unable to compete with paclitaxel for binding to the microtubules at these concentrations. 43 o LU Figure 14 - 3H-Paclitaxel Binding to Microtubule Assay Microtubules were preformed and then incubated with no drug, eleutherobin, or one of two concentrations of ceratamine A. Each reaction contained 2 uM 3H-paclitaxel with a specific activity of 1600 dpm/pmol. The ratio of free to total paclitaxel was then determined by scintillation counting over multiple experiments. The bar represents the mean of 3 independent experiments with the error bars representing standard deviation. Eleutherobin is a positive control. The change between eleutherobin and DMSO is statistically significant. However, the changes between DMSO and either ceratamine treatement are not statistically significant. 44 3.3 - Discussion This work describes the characterization of the cellular effects of two novel chemically simple antimitotic agents, ceratamine A and ceratamine B. These compounds block the cell cycle in M phase and inhibit cellular proliferation. Ceratamine A disrupts the microtubule network in vivo in a manner that, to our knowledge, has not been previously described. In vitro, ceratamine A and ceratamine B promote the polymerization of tubulin similar to paclitaxel, but ceratamine A does not compete with paclitaxel for binding to microtubules. Ceratamine A and ceratamine B were discovered in a phenotypic antimitotic ELICA screen of natural products (Manzo et al., 2003). Not surprisingly, mitotic spreads and flow cytometry confirmed that ceratamine A and ceratamine B blocked the cells in M phase. This is similar to what has been reported with paclitaxel (Roberge et al., 2000). Significantly, the ceratamines had IC50 values of 12 and 21 uM, respectively, for antimitotic properties. These values are three orders of magnitude higher than the 3 nM reported for paclitaxel (Roberge et al., 2000). Ceratamine A and ceratamine B were also found to inhibit cellular proliferation in a dose-dependent manner. Similar to the case with their antimitotic property, ceratamine A has a lower IC50 value than ceratamine B (2 uM versus 3 uM). However, both ceratamine A and B have IC50 values significantly higher than the 5 nM reported for paclitaxel (Anderson et al., 1997). It must be mentioned that there is a 5- to 7-fold increase in the IC50 values of antimitotic activity versus antiproliferative activity for the ceratamines. While this was not observed by Anderson et al. (1997) for paclitaxel it was observed for vincristine where the reported antimitotic IC50 was 15 nM and the reported antiproliferative IC50 was 2 nM. This is a 7.5-fold 45 increase. The M T T assay used for the antiproliferative assay measures a slow down in the cell cycle, while the antimitotic assay measures a complete halt. It may be that at lower concentrations of antimitotic agents, the cells are not becoming blocked in mitosis, but the cell cycle is slowed resulting in an inhibition of proliferation. Although they are not very potent, the ceratamines may be useful lead compounds for drug development. Combinatorial chemistry using the structure of ceratamines as a starting point may lead to the development of more potent analogues. From these studies, there is evidence that changing a hydrogen to a methyl group can decrease the IC50 values of both the antimitotic and antiproliferation assay. Whether the experiments are conducted on more potent analogues or the ceratamines, only animal studies can address their potential for clinical use. While it had been found that the ceratamines block cells in mitosis and inhibit cellular proliferation, their mechanism of action was unknown. Since many antimitotic agents disrupt the microtubule network of the cell, this was investigated. Confocal immunofluoresence microscopy revealed that ceratamine A treatment of cells resulted in a disruption of the microtubule network. Only ceratamine A was examined as it was the more potent of the two, and the mechanism of action for ceratamine A and B is almost certainly the same given their close structural similarity. In interphase, not all of the cells were affected by ceratamine A treatment. However, those that were affected had intense perinuclear tubulin staining. Similar perinuclear staining has been described in response to aphidicolin treatment for 72 hours (Tanaka et al., 1998). What is more striking is the absence of microtubules in mitotic cells. Instead, pillars of tubulin are present. These structures are in sharp contrast to the bushy microtubule bundles observed following 46 treatment with paclitaxel. This hints that the ceratamines affect microtubules in a different manner than paclitaxel. The pillars are of particular interest. While it is clear that these structures are composed of tubulin, it is not clear whether they are groupings of microtubules or aggregates of tubulin. Electron microscopy could provide an answer to this question. It may be possible to visualize a cross section of the pillars and identify microtubules versus protein aggregates. These pillars raise several interesting questions: is their formation nucleated or not? If there is a site of nucleation, is it a single protein such as membrane bound tubulin that has been reported (Luduena 1998), or a protein complex? Why does the number of foci decrease with time? Do the pillars aggregate or do they breakdown and then reform on pre-existing ones? Are proteins other than tubulin found in the pillars and if so what are the roles of these proteins? With clear in vivo data that microtubules are affected by ceratamine A treatment, work shifted to an in vitro assay in an attempt to characterize the nature of this effect. As the vast majority of chemotypes that affect microtubule dynamics inhibit their formation, this was initially investigated. The ceratamines were found to not inhibit the polymerization of tubulin to microtubules in vitro. This is in direct contrast to nocodazole, which was used as a control and which effectively inhibited microtubule formation. This result clearly demonstrated that ceratamine1 A and ceratamine B did not function as antimitotic agents through inhibition of microtubule polymerization. Next, the ceratamines were investigated for their ability to promote the polymerization of microtubules. In the in vitro assay designed to examine the ability of a compound to stimulate the polymerization of microtubules, the ceratamines were found 47 to be positive. Their ability to promote the overpolymerization was similar to that observed with paclitaxel, but they were much less potent. This is in line with earlier antimitotic and antiproliferative results which demonstrated that paclitaxel was more potent than ceratamine A and ceratamine B. Also consistent with earlier results was the relative potency of ceratamine A and B. At the same concentration, ceratamine A induced more microtubule polymerization than ceratamine B. This provides a possible explanation for why ceratamine A is more potent in the in vivo assays than ceratamine B. Ceratamine A may have a more dramatic effect on the microtubule network of cells. The microtubule polymerization assay measures the increase in light scattering as microtubules form. However, the assay does not differentiate between light scattering due to microtubule formation, precipitation of the drug or formation of protein aggregates. To determine whether ceratamine A was indeed promoting the formation of microtubules, the reaction mixture was analyzed by electron microscopy to visualize the products. It was found that similar to paclitaxel, ceratamine A was indeed promoting the formation of microtubules. This provided clear evidence that ceratamine A could promote the polymerization of microtubules. At this point in the investigation there was some conflicting data. From the in vitro work it was clear that the ceratamines promoted the polymerization of microtubules.. At this level they are similar to paclitaxel despite being less potent. However, from in vivo immunofluorescence microscopy it was evident that ceratamine affected microtubules in a manner distinct from paclitaxel. Paclitaxel treatment results in the formation of microtubule bundles in interphase and \"bushy\" spindles at mitosis. This was never observed with ceratamine treatment. Instead, the dramatic pillars of tubulin staining 48 were observed in mitotic cells, and perinuclear networks were observed in some interphase cells. We thus hypothesized that while both ceratamine A and paclitaxel promote microtubule polymerization, they might bind at different sites in the microtubule. This was examined by a paclitaxel binding competition assay. The paclitaxel binding assay demonstrated that ceratamine A could not compete with paclitaxel for binding to the microtubule polymer. There were three possible explanations for this. First, the concentration of ceratamine A may be insufficient to bind to the microtubule. This is unlikely as the highest concentration tested, 50 uM, was sufficient to induce polymerization of microtubules in vitro (data not shown). The second explanation is that the binding sites overlap but the binding affinity of ceratamine A is too low to compete with paclitaxel. This is a possibility that cannot be discounted. There was a 25-fold excess of ceratamine A over paclitaxel and still no competition was observed, but it is true, from earlier experiments, that ceratamine A was much less potent than paclitaxel at inducing microtubule polymerization. The third explanation is that ceratamine A binds to a region of the microtubule outside the taxoid binding site and does not disrupt it. This seems the most likely because it explains the results of the paclitaxel binding assay and provides an explanation for the in vivo results. While more evidence is required, intuitively it follows that binding to a site other than the taxoid binding site might produce different microtubule morphologies. Work remains to determine which of the possible three explanations for the competition assay results is correct. The most direct experiment to determine where the ceratamines bind on microtubules is to solve the structure of tubulin/microtubules with the ceratamines bound. Other methods that are theoretically possible include isothermal 49 titration calorimetry and HPLC. Both of these methods depend on prohibitively large quantities of both protein and drug. Another possible mechanism to resolve this question involves using a radiolabeled ceratamine molecule. If paclitaxel can compete with labeled ceratamine, then this would be evidence of binding to the same site. While this experiment does not require prohibitive amounts of either ceratamine or protein, it obviously requires that the labeled ceratamine be synthesized. This may not prove to be trivial. If the ceratamines do bind at a region on the microtubule distinct from the taxoid binding site, it places them in a very small group. To date, there are only two other molecules known to promote the stabilization of microtubules and bind outside the taxoid binding site. One is laulimalide (Pryor et al., 2002) and the other is peloruside (Gaitanos et al., 2004). In contrast to the ceratamines, laulimalide treatment of cells produces the characteristic microtubule bundling typical of paclitaxel treatment (Mooberry et al., 1999). On this basis alone, it would appear that the ceratamines bind to a site unique not only from paclitaxel, but laulimalide as well. The work to characterize the biology of the ceratamines has demonstrated that they are small molecules with traditional antimitotic and antiproliferative properties but unique cellular effects. The ceratamines affect the microtubule network both in interphase and mitosis, producing very unique tubulin staining, and promote the polymerization of microtubules in vitro. A simple model for the ceratamines mechanism would be that the drug binds to either tubulin or microtubules and inhibits the correct dynamics of the mitotic spindle. This leads to cells blocked in mitosis that in turn inhibits their 50 proliferation. Further, this model would include the ceratamines binding to a portion of the microtubule separate from the taxoid binding site. The goal of our antimitotic screening program is to discover novel antimitotic agents with potential for the treatment of cancer. The ceratamines' potential to become clinically relevant is unclear at this point. They are significantly less potent than paclitaxel, and hypothesizing from their structure their water solubility may be limited. These two factors may combine to prevent appropriate activity within either animal models or humans. These questions will have to be answered in animal studies with the ceratamines. Despite questions about their solubility and potency, the ceratamines do appear to have one strong asset that may be very useful for further work. They are structurally very simple, especially when compared to the taxanes. For a quick measure, this structural simplicity is reflected in the number of chiral centers present in the molecules. The ceratamines do not have any chiral centers, whereas the paclitaxel has eleven. Each one of these chiral centers creates significant problems when attempting to synthesize the drug. It is hoped that the absence of these chiral centers would allow the ceratamines to have none of the supply problems that so plagued the development of paclitaxel (Cragg et al., 1993). Even if not developed further, the ceratamines represent what is one of the simplest chemical structures of a microtubule-stabilizing agent yet described. 51 Chapter IV - The Biology of Dimethyl Varacin 4.1 - Dimethyl Varacin Specific Introduction Varacin, a benzopentathiepin or cyclic polysulfur, was initially isolated from the marine sponge Lissoclinum vareau and found to be a cytotoxic and antifungal agent (Davidson et al., 1991). Further work with varacin has shown that it can cleave DNA in vitro which may be the cause of its cytotoxic properties, though in vivo work is required to verify this observation (Chatterji et al., 1998). Other work has concentrated on the polysulfur ring of varacin. In silico calculations have shown that the polysulfur ring may open in a thiol dependent manner to release a triatomic sulfur molecule and a thiolate anion of the remaining varacin (Greer 2001). While this has not been demonstrated for varacin, some thiol reactive agents are known to affect the microtubule network of the cell (Jordan et al , 1998). In particular, they usually inhibit the polymerization of tubulin in vitro and cause a breakdown in the microtubule network in vivo. Dimethyl varacin was isolated in our phenotypic antimitotic assay. Its structure was elucidated by Lu Yang using X-ray crystallography and is shown in Figure 15 A. The structure of varacin is shown in Figure 15B. 52 Figure 15 - The Chemical Structure of Varacin Compounds The structure of Dimethyl Varacin (A) and Varacin (B) is shown. 53 4.2 - RESULTS 4.2.1 - Antimitotic and antiproliferative activity The dose dependence of DMV's activity in the antimitotic assay was examined. It was found that D M V had an IC 5 0 value of 2 uM in HCT-116 p53 +/+ cells and 2 uM in HCT-116 p53 -/- cells (Figure 16A). The presence or absence of p53 did not affect the antimitotic properties of D M V . The dose dependence of DMV's activity in the antimitotic assay with MCF-7 mp53 cells was also examined with an IC50 of 2 uM, similar to HCT-116 cells. It was next examined if D M V could prevent the proliferation of MCF-7 mp53 cells. D M V does inhibit proliferation with an IC50 value of approximately 50 nM (Figure 16B). The effect of D M V on the cellular proliferation of M D A 231 cells was examined and found to be very similar to that observed with MCF-7 mp53 cells (Figure 16B). 4.2.2 - Dimethyl Varacin inhibits microtubule polymerization: The effect of D M V on the polymerization of tubulin into microtubules was examined using an in vitro tubulin polymerization assay (Figure 17). It was found that, similar to 10 uM nocodazole, D M V at 50 uM was able to inhibit the polymerization of tubulin into microtubules. A sample with no drug (1% DMSO) did polymerize. 54 A 0.1 1 10 DMV Cone (uM) B DMV Cone (uM) Figure 16 - Antimitotic E L I C A and Cell Proliferation Assays with Various Concentrations of Dimethyl Varacin and Different Cell Lines (A) Antimitotic ELICA assays were performed with MCF-7 mp53 (closed diamonds), HCT-116 p53 +/+ (open triangles) and HCT-116 p53 -/- (closed triangles) at various doses of DMV. (B) Cell Proliferation Assays were performed with MCF-7 mp53 (closed triangles) and MDA-231 (open inverted triangles) to examine DMV's effect of cellular proliferation. In both assays the error bars represent standard deviation of 3 measurements. 55 0.3 Time (min) Figure 17 - Dimethyl Varacin Inhibits Microtubule Formation In Vitro Inhibition of Tubulin Polymerization Assay with DMSO (open squares), nocodazole (open triangles), or D M V (closed circles). The assay was performed with a final tubulin concentration of 3 mg/mL. 56 4.2.3 \u00E2\u0080\u0094 Dimethyl Varacin rapidly breaks down the microtubule network in vivo: Indirect immunofluorescence microscopy was used to analyze the microtubule network of cells treated with 10 uM D M V (Figure 18). It was found that D M V disrupts the microtubule network in the majority of cells within 30 minutes. Analysis of cells treated with 0.1% DMSO at 30 and 120 minutes revealed the presence of a defined microtubule network with a clear MTOC. However, in cells treated with 10 uM D M V for 30 and 120 minutes the microtubule network was clearly disrupted. The overall staining was far less intense and is distributed in the cytoplasm in an irregular manner. There did not appear to be any significant microtubule network present with no obvious MTOC. These results indicated that microtubules are disrupted in vivo by D M V treatment. 4.2.4 - Dimethyl Varacin induces mitotic-like phosphorylations from all phases of the cell cycle: Dimethyl varacin was discovered in a phenotypic antimitotic assay in which mitotic cells are identified indirectly, via the detection of a phosphorylation site on nucleolin (Roberge et al., 2000). To further examine the effect of D M V on the cell cycle, flow cytometry was utilized. Flow cyotometry by staining DNA with propidium iodide allows the detection of cells in Gi , S, and G2/M phase based on differing amounts of DNA. Cells treated with 1 % DMSO showed a typical DNA profile for asynchronously cycling cells (Figure 19A). Treatment of cells with 300 nM nocodazole for 24 hours resulted in a strong accumulation of cells in G2/M phase (Figure 19A). This type of profile is typical for antimitotic agents. Treatment with either 5 or 25 uM D M V for 5 hours did not produce a mitotic arrest (Figure 19A). Instead, the D N A profile resembled 57 DMSO 1 0 u M D M V Figure 18 - Dimethyl Varacin Disrupts Microtubules In Vivo Indirect immunofluorescence of ^-tubulin in Swiss 3T3 fibroblasts treated with DMSO DMV. 58 DMSO Nocodazole 5uM DMV 25uM DMV Figure 19 - Dimethyl Varacin's Effect on the Cell Cycle Flow cytometry of MCF-7 mp53 cells treated with DMSO, nocodazole for 24 hours, 5 uM D M V or 25 uM DMV for 5 hours. (A) DNA profile of cells stained with propidium iodide. (B) GF-7 versus DNA content. The darkness of the spot represents the number of events. that of DMSO for asynchronously cycling cells. To gain further information, it is possible to use an antibody with a phospho-epitope typically present only during M phase to distinguish between cells in G2 and M phase. Cells examined in this manner after exposure to 1 % DMSO revealed the expected pattern for asynchronously cycling cells with the majority of cells in interphase (Figure 19B) and a small proportion in mitosis. Cells treated with 300 nM nocodazole for 24 hours resulted in a profile characteristic of antimitotic agents with the majority of cells blocked in mitosis (Figure 19B). However, cells treated with 5 uM D M V for 5 hours yielded very startling results (Figure 19B): the phosphorylations that are typically only observed during mitosis are induced in all phases of the cell cycle including G] and S phase. Treatment with 25 uM D M V resulted in stronger induction of these phosphorylations (Figure 19B). Thus, D M V is not a traditional antimitotic agent, and the antimitotic activity describe in Figure 16 is not due to mitotic arrest but to an increased phosphorylation of proteins in all phases of the cell cycle. 4.2.5 - Dimethyl Varacin does not cause global activation of Cdc2: The primary kinase responsible for mitosis is thought to be Cdc2 (Nigg 2001). It was hypothesized that D M V induces the premature formation of mitotic phosphorylations by causing the overactivation of Cdc2 phoshpotransferase activity. A five hour treatment of MCF-7 mp53 cells with 10 uM D M V led only to an approximate 70% increase in Cdc2 phosphotransferase activity (Figure 20). However, a five hour treatment with nocodazole, an agent known to block cells in mitosis leading to an activation of Cdc2, led to an approximate 300% increase in Cdc2 phosphotransferase activity (Figure 20). This 60 Figure 20 - Dimethyl Varacin Does Not Activate Cdc2/cyclin B Cdc2 Kinase Assays with DMSO, DMV, or nocodazole. The assay was conducted by an immunoprecipitation of Cdc2/cyclin B from a cell lysate following incubation with the indicated drugs. The assay was only conducted once, but the findings are supported by results from Kinexus Kinetwork\u00E2\u0084\u00A2 screens. 61 result indicates that activation of Cdc2 kinase is not responsible for the extensive mitotic-like phosphorylations observed following D M V treatment. 4.2.6 - In gel kinase assays: We wished to further investigate the potential kinase activity changes following D M V treatment. To this end, an in-gel kinase assay was conducted either with or not myelin basic protein (MBP), a general kinase substrate, polymerized directly into the polyacrylamide gel. The gel is treated to renature kinases and the gel is incubated with [y-3 2 P]ATP and active kinases can be detected by the resulting radioactive bands. The identities of the kinases in question are not needed making this technique an excellent starting place. However, one must be mindful that heteromeric kinases, such as Cdks, and non-renaturable kinases will not be detected (Gowdy et al., 1998). A number of kinases were activated by treatment with D M V (Figure 21 A). Significantly, there was a large increase in overall kinase activity following treatment with 5 uM D M V compared with 1 uM DMV. Kinases of approximately 65, 52, and 44 kDa were strongly activated. The kinases of 65 and 52 kDa were not activated by nocodazole treatment, whereas the kinase at 44 kDa was activated by nocodazole. The 44 kDa kinase in cells treated with nocodazole might be activated by nocodazole induced disruption of the microtubules. Alternatively, if it is typically activated at mitosis increased activity may be observed because nocodazole is blocking the cells in mitosis. The similarities and differences between nocodazole and D M V indicate that while both depolymerize microtubules, the downstream effects are not necessarily the same. Significantly, the concentration at which D M V strongly activates the three kinases (5 62 A B * 2 5 A o =J. i _ N ^ LT> 2 O \"O co g > > > 2 o 2 2 2 Q Z Q Q Q 178 \u00E2\u0080\u0094-114 \u00E2\u0080\u0094 82 \u00E2\u0080\u0094 life-\u00E2\u0080\u00A2 61 \u00E2\u0080\u0094 47 \u00E2\u0080\u0094 36 \u00E2\u0080\u0094 \u00E2\u0080\u00A2mmm* 25 \u00E2\u0080\u0094 Figure 21 - In Gel Kinase Assays Ten \ig of whole cell lysate from cells treated in the indicated manner were separated on a 10% SDS-PAGE containing either (A) 0.2 mg/mL MBP or (B) not. A kinase assay using y-32P ATP was conducted directly in the gel. The ladder represents the distance traveled by prestained protein markers (kDa). 63 uM) is the same concentration at which D M V was observed to be active in the antimitotic assay and to induce microtubule depolymerization in vivo. It was possible that the active kinases were not phosphorylating MBP but were autophosphorylating. To examine this, an in-gel kinase assay was performed with no polymerized substrate (Figure 2IB). This would most likely examine autophosphorylation; however, it cannot be eliminated that the substrate is a different protein with a similar molecular weight to the kinase. There were similarities and differences observed between the assays performed with MBP present or not. Similar to the case with MBP polymerized into the gel, a band appeared at 44 kDa following both D M V treatment and nocodazole. It is therefore likely that a kinase of approximately this molecular mass was capable of autophosphorylation. However, the kinases at approximately 62 and 55 kDa that were very active with MBP present, did not appear when MBP was omitted. This would indicate that these kinases were incapable of autophosphorylation. Further, a kinase at approximately 34 kDa appeared to be inhibited or degraded following D M V treatment. This was not observed with MBP present. One possible explanation is the presence of a kinase of similar molecular mass capable of phosphorylating MBP but not autophosphorylation. When MBP was present in the gel, the activity of this second kinase masked any changes in the activity of the 34 kDa kinase. However, when there was no MBP present in the gel, no signal from the second kinase was observed allowing one to observe changes in the activity of the 34 kDa kinase. A second explanation is that there was only one kinase capable of both autophosphorylation and phosphorylating MBP at 34 kDa. However, D M V treatment may inhibit its ability to autophosphorylate but not to phosphorylate MBP. 64 4.2.7 - Kinexus Kinetwork \u00E2\u0084\u00A2 phospho-site screens: We had data from both flow cytometry and in-gel kinase assays that a number of kinases were becoming activated post-DMV treatment. However, we had no simple and efficient method of determining the identity of the kinases. To survey the widest possible number of possible substrates in the shortest time, four Kinexus Kinetwork\u00E2\u0084\u00A2 phospho-site screens, 1.3, 2.0, 3.0, and 4.0 were conducted. Kinexus Kinetwork\u00E2\u0084\u00A2 screens are conducted by multi-immunoblotting (Pelech 2004). In this technique 500 pg of protein from a cell lysate are separated in one broad lane on an SDS-PAGE. The proteins are transferred to a membrane such as nitrocellulose. A plastic manifold with approximately 20 different channels is then placed over the membrane effectively creating 20 equivalent membranes suitable for Western blotting. By conducting thorough testing beforehand to avoid cross-reactivity issues, a mixture approximately 3 antibodies may be added to each channel. The blot is then incubated with appropriate secondary antibodies and developed with E C L (Pelech 2004). This technique allows Kinexus to probe with approximately 35 primary phospho-specific antibodies per protein sample (Pelech 2004). The four screens conducted allowed us to determine changes in the phosphorylation status of 80 proteins at 111 independent sites within a three week turnaround time. This presented a significant time and cost savings over conducting the similar number of individual Western blots. Phosphorylations were detected at 72 of 111 sites (65%). Twenty of the sites investigated had changes between 100%, (or two-fold) and infinity; phosphorylation was not detected at 39 sites (35%). The results of this experiment are summarized in Figure 22. There are numerous observations emerging from these data: 65 Protein Phosphorylation (com) im m$ mm soao mm t2cco I\u00C2\u00ABOCO Accv: 11\u00C2\u00AB S72 M1P-PKTI72 COKi YlS COK I TW1& COKI TfC? 5S!\u00C2\u00A9?;S2a7 { j j j ^ j fg MKK0S2C7 MLK312? 7i?S2St MSKU2p?SS37fi MSKU2p858376 451 255 mTOftS2\u00C2\u00ABH$| -12 tm\ S89S \u00E2\u0080\u00A2 Figure 22 - Summary of Data from Kinexus Kinetworks\u00E2\u0084\u00A2 Phosphosite Screens (continued on next page) The combined results from the Kinetworks\u00E2\u0084\u00A2 KPSS 1.3, 2.0, 3.0, and 4.0 screen analyses of the control (light bars) and those from the drug-treated (dark bars) cells were averaged and then graphed. The percentage difference between the drug-treated and control cells is indicated at the end of the bars. Only those target phosphoproteins that were detected with these screens are shown. 66 Protein Phospfsosylation (cpm) m am im PAKpMSUl P A K p S O S U I PDKi $241 PKBaS4J3 PKCa$\u00C2\u00A357 PKCS TSOS PKC8 1538 4 0 l PKft 1451 PP1aT32D -g^ PfiK1T77S PRX2 IBIS RE/I pSi S 2 \u00C2\u00BB \u00C2\u00ABsflp$9$2\u00C2\u00A79 -52-RST3SS R8SC12 86 S7S0 R&S807ISS11 -1 RBT82I Rik! $380 R s i l T573 8\u00C2\u00A3Kp70T\u00C2\u00ABVH24 SSK o7C im $5Kp?3 Sncl 546 Y23iMY240> Stel pGGY23iuY240 Src Y "Thesis/Dissertation"@en . "2005-05"@en . "10.14288/1.0091880"@en . "eng"@en . "Biochemistry and Molecular Biology"@en . "Vancouver : University of British Columbia Library"@en . "University of British Columbia"@en . "For non-commercial purposes only, such as research, private study and education. Additional conditions apply, see Terms of Use https://open.library.ubc.ca/terms_of_use."@en . "Graduate"@en . "The biology of three novel natural product microtubule interacting agents : ceratamine A and B and dimethyl varacin"@en . "Text"@en . "http://hdl.handle.net/2429/16296"@en .