Isolation and characterization of nitrogenous metabolites from the ascidians Cystodytes sp., Lissoclinum patella, and Botryllus sp.

This dissertation describes the isolation, characterization, and biological evaluation of several nitrogenous metabolites from ascidians. The first chapter reviews marine natural products that have shown promising biological activities or have been instrumental in enhancing our understanding of certain cellular processes. Since the research described is largely based on bioassay guidance to fractionate extracts of marine ascidians into bioactive secondary metabolites, the second chapter covers general approaches used to find bioactive natural products. The subsequent chapters detail the isolation and characterization of several new nitrogenous metabolites using a variety of spectroscopic techniques. The active extract of a Fijian Cystodytes sp. ascidian was fractionated to yield a series of pyridoacridine alkaloids with topoisomerase II (topoII) activity. The structures of these alkaloids were elucidated primarily using nuclear magnetic resonance (NMR) spectroscopic techniques. The pyridoacridine alkaloids were cytotoxic towards the cultured mammalian cell line HCT116; acting by intercalating into deoxyribonucleic acid (DNA) and inhibiting the catalytic activity of topoII. Chapter 3 details the structure elucidation of these alkaloids and the mechanistic and biological experiments undertaken to support this conclusion. A new cyclic peptide, patellamide E, was isolated from the ascidian Lissoclinum patella collected at Pulau Salu, Singapore. Its structure was determined by NMR spectroscopy, and its absolute configuration by acid hydrolysis and analysis of the derivatized constituent amino acids by high performance liquid chromatography (HPLC). The first part of the fourth chapter describes the structure elucidation of patellamide E. The second part of Chapter 4 describes the new cytotoxic cyclic peptides, tawicyclamides A and B, which were isolated from the ascidian Lissoclinum patella collected in the Philippines. The structures of these peptides were determined by NMR spectroscopy, oxidation studies, and tandem mass spectrometry (MS/MS). Their absolute configurations were again determined by HPLC analysis of derivatized constituent amino acids obtained from acid hydrolysis. X-ray crystallography confirmed the structure of tawicyclamide B and showed that the compound assumes an unusual three-dimensional conformation, which was facilitated by a cis-valine-proline amide bond and stabilized by an intramolecular hydrogen bond. Tawicyclamides A and B represent a new family of cyclic octapeptides, possessing thiazole and thiazoline amino acids but lacking the oxazoline ring characteristic of previously reported cyclic peptides from L. patella. NMR data supported the hypothesis that isomerization of the valine-proline amide bond occurred on oxidation of the thiazoline ring to a thiazole. Molecular modeling studies, undertaken to establish the solution structures of tawicyclamide B and dehydrotawicyclamide B, supported the isomerization conjecture. The new bromotyrosine derivatives, botryllamides A-D, which were isolated from a Philippine Botryllus sp. ascidian are discussed in Chapter 5. The structures of these compounds were deduced from NMR and mass spectrometric data.

ISOLATION AND CHARACTERIZATION OF NITROGENOUS METABOLITES FROM THE ASCIDIANS CYSTODYTES SP., LISSOCLINUM PATELLA, AND BOTRYLLUS SP. by Leonard A. McDonald A dissertation submitted to the faculty of The University of Utah in partial fulfillment of the requirements for the degree of Doctor of Philosophy Department of Medicinal Chemistry The University of Utah August 1994 Copyright © Leonard A. McDonald 1994 All Rights Reserved THE UNIVERSITY OF UTAH GRADUATE SCHOOL SUPERVISORY COMMITTEE APPROVAL of a dissertation submitted by Leonard A. McDonald This dissertation has been read by each member of the following supervisory committee and by majority vote has been found to be satisfactory. Chair: Chris M. Ireland Arthur D. Broom arrell R. DaVIS William W. Epst~ THB UNIVBRSITY OF UTAH GRADUATB SCHOOL FIN AL READING APPROV AL To the Graduate Council of the University of Utah: I have read the dissertation of Leonard A. McDonald in its final form and have found that (1) its fonnat, citations and bibliographic style are consistent and acceptable; (2) its illustrative materials including figures, tables, and charts are in place; and (3) the final manuscript is satisfactory to the supervisory committee and is re~y d;~on to The Gnd~_te_C1-=c::;;;..h__OO_ _j_)__ ______ _ Dale Chris M. Ireland Chair, Supervisory Committee Approved for the Major Department ~~ Arthur D. Broom Chair/Dean Approved for the Graduate Council Ann W. Hart Dean of The Graduate School ABSTRACT This dissertation describes the isolation, characterization, and biological evaluation of several nitrogenous metabolites from ascidians. The first chapter reviews marine natural products that have shown promising biological activities or have been instrumental in enhancing our understanding of certain cellular processes. Since the research described is largely based on bioassay guidance to fractionate extracts of marine ascidians into bioactive secondary metabolites, the second chapter covers general approaches used to find bioactive natural products. The subsequent chapters detail the isolation and characterization of several new nitrogenous metabolites using a variety of spectroscopic techniques. The active extract of a Fijian Cystodytes sp. ascidian was fractionated to yield a series of pyridoacridine alkaloids with topoisomerase II (topoII) activity. The structures of these alkaloids were elucidated primarily using nuclear magnetic resonance (NMR) spectroscopic techniques. The pyridoacridine alkaloids were cytotoxic towards the cultured mammalian cell line HCTl16; acting by intercalating into deoxyribonucleic acid (DNA) and inhibiting the catalytic activity of topoII. Chapter 3 details the structure elucidation of these alkaloids and the mechanistic and biological experiments undertaken to support this conclusion. A new cyclic peptide, patellamide E, was isolated from the ascidian Lissoclinum patella collected at Pulau Salu, Singapore. Its structure was determined by NMR spectroscopy, and its absolute configuration by acid hydrolysis and analysis of the derivatized constituent amino acids by high performance liquid chromatography (HPLC). The first part of the fourth chapter describes the structure elucidation of patellamide E. The second part of Chapter 4 describes the new cytotoxic cyclic peptides, tawicyclamides A and B, which were isolated from the ascidian Lissoclinum patella collected in the Philippines. The structures of these peptides were determined by NMR spectroscopy, oxidation studies, and tandem mass spectrometry (MSIMS). Their absolute configurations were again determined by HPLC analysis of derivatized constituent amino acids obtained from acid hydrolysis. X -ray crystallography confirmed the structure of tawicyclamide B and showed that the compound assumes an unusual three-dimensional conformation, which was facilitated by a cis-valine-proline amide bond and stabilized by an intramolecular hydrogen bond. Tawicyclamides A and B represent a new family of cyclic octapeptides, possessing thiazole and thiazoline amino acids but lacking the oxazoline ring characteristic of previously reported cyclic peptides from L. patella. NMR data supported the hypothesis that isomerization of the valine-proline amide bond occurred on oxidation of the thiazoline ring to a thiazole. Molecular modeling studies, undertaken to establish the solution structures of tawicyclamide Band dehydrotawicyclamide B, supported the isomerization conjecture. The new bromotyrosine derivatives, botryllamides A-D, which were isolated from a Philippine Botryllus sp. asci dian are discussed in Chapter 5. The structures of these compounds were deduced from NMR and mass spectrometric data. v iJ; For -{f Mollie ,'- To my family for all their love, support, and understanding TABLE OF CONTENTS ABSTRACT ....................................................................................... iv LIST OF FIGURES .............................................................................. ix LIST OF TABLES ................................................................................ xi LIST OF ABBREVIATIONS .................................................................... xii ACKNOWLEDGMENTS ........................................................................ xv 1. INTRODUCTION AND BACKGROUND ........................................... 1 Rationale for Studying Marine Organisms ............................................ 1 Marine Natural Products Chemistry: Evolution from "Phytochemical" to Biomedical Discipline ............................................................... 2 Summary .................................................................................. 20 2 . STRATEGIES FOR DISCOVERING POTENTIAL ANTICANCER AGENTS: BIOASSAY GUIDED FRACTIONATION ............................. 22 In Vitro Assays ........................................................................... 23 Mechanistic Assays ...................................................................... 26 In Vivo Assays ........................................................................... 29 3. THE CHEMISTRY AND BIOLOGY OF THE ASCIDIAN CYSTODYTES SP ....................................................................... 30 Chemistry ................................................ " ................................. 31 Biology .................................................................................... 42 Mechanistic Studies ...................................................................... 43 Summary .................................................................................. 45 Review of Pyridoacridines .............................................................. 48 4. THE CHEMISTRY OF THE ASCIDIAN LISSOCLINUM PATELLA .......... 49 L. patella from Singapore ............................................................... 49 L. patella from the Philippines .......................................................... 55 Review of Cyclic Peptides .............................................................. 80 5. THE CHEMISTRY OF THE ASCIDIAN BOTRYLLUS SP ...................... 84 Isolation of the Botryllamides ............................ , .............................. 84 Structure Determination of Botryllamide D ............................................ 84 Structure Determination of Botryllamides B, C, and A .............................. 91 Biology .................................................................................... 91 Review of Linear Peptide Alkaloids from Ascidians ................................. 94 6 . EXPERIMENTAL ............................................. : ......................... 95 Chemicals, Reagents, and Organisms ................................................. 95 General Experimental Procedures ...................................................... 95 The Chemistry of Cystodytes sp ....................................................... 96 The Chemistry of Lissoclinum patella (Singapore) ................................... 99 The Chemistry of Lissoclinum patella (Philippines) ................................. 100 The Chemistry of Botryllus sp .......................................................... 104 APPENDICES A. NMR SPECTRA OF COMPOUNDS FROM CYSTODYTES SP ............ 107 B. NMR SPECTRA OF COMPOUNDS FROM LISSOCLINUM PATELLA ............................................................................ 120 C. CIDSPECTRAOFTAWICYCLAMIDEA ...................................... 141 D. NMR SPECTRA OF COMPOUNDS FROM BOTRYLLUS SP .............. 155 E. ISOLATION FLOW DIAGRAMS ................................................ 170 REFERENCES .................................................................................... 175 viii LIST OF FIGURES 2. 1 . Chemistry involved in the biochemical induction assay ............................. 27 3.1. Region of a 300 ms ROESY spectrum of dehydrokuanoniamine B showing the H3-H4 crosspeak .................................................................... 34 3 .2. Dehydrokuanoniamine B with HMBC correlations used in assigning the quaternary carbon atoms ................................................................. 34 3.3. Structure of shermilamine C with HMBC correlations used in assigning the quaternary carbon atoms .................................... ; ............................ 37 3.4. HMBC correlations used in assigning the quaternary carbon atoms in cystodytin J ............................................................................... 39 3.5. Fluorescence spectra of ethidium bromide in PBS solution with various concentrations of diplamine ............................................................. 40 3.6. Normalized fluorescence for ethidium bromide in PBS solution with calf thymus DNA and increasing concentrations of pyridoacridines .................... 41 3.7. Fluorescence spectra of eilatin in PBS solution with various concentrations of calf thymus DNA ...................................................................... 42 3.8. Agarose gel showing inhibitory effects of etoposide and diplamine on topoll catalyzed decatenation of kDNA ........................................................ 44 4.1. Proton and COSY derived amino acid spin networks for patellamide E ........... 53 4.2. Partial structures and HMBC correlations for patellamide E ........................ 54 4.3. Nickel peroxide oxidation of tawicyclamide A ........................................ 57 4.4. Proton and COSY derived amino acid spin networks for tawicyclamide A ....... 58 4.5. HMBC correlations for tawicyclamide A .............................................. 59 4.6. Linear acy lium ion of tawicyclamide A ................................................ 62 4.7. High mass region (mlz 825-370) from the cm spectrum of the (M+H)+ ion of tawicyclamide A ....................................................................... 63 4.8. Low mass region (mlz 625-80) from the CID spectrum of the (M+H)+ ion of tawicyclamide A ....................................................................... 64 4.9. CID spectrum and structure of the m/z 682 ion of tawicyclamide A ............... 65 4.10. Partial structure of tawicyclamide B showing five bond correlation between leucine and thiazoline .................................................................... 69 4. 11 . X-ray model of tawicyclamide B ....................................................... 70 4. 12. Stereo drawing of the solution structure of tawicyclamide B ....................... 71 4. 13. Region of a 300 ms ROESY spectrum of tawicyclamide B showing the HI0-HI5 crosspeak ...................................................................... 73 4.14. Region of a 300 ms ROESY spectrum of tawicyclamide B showing the H3- H27 crosspeak ............................................................................ 74 4.15. Stereo drawing of the solution structure of dehydrotawicyclamide B .............. 76 4.16. Region of a 300 ms ROESY spectrum of dehydrotawicyclamide B showing the H 13-HIS crosspeak ................................................................. 78 4.17. Region of a 300 ms ROESY spectrum of dehydrotawicyclamide B showing the NH1-H10 crosspeak ................................................................. 79 5. 1 . EI mass spectrum and fragmentation of botryllamide D ............................. 86 5.2. Select HMBC correlations for Botryllamide D ........................................ 89 5.3. Botryllamide D ROESY correlations and slice through 400 ms ROESY spectrum ................................................................................... 89 5.4. Slices from 400 ms ROESY spectra used to establish the configuration about the C5-C6 olefinic bond in botryllamides B, C, and A ....................... 93 x LIST OF TABLES 2.1 . Human Cell Lines used in Natural Products Screening .............................. 23 2.2. Multidrug Resistant Cell Lines used in Evaluating Pure Compounds .............. 24 2.3. CHO Cell Lines used in Screening and Mechanistic Studies ........................ 25 3.1. NMR Assignments for the TFA Salt of Dehydrokuanoniarnine B ................. 33 3.2. NMR Assignnlents for the TFA Salt of Shermilarnine C ............................ 36 3.3. NMR Assignments for Cystodytin J ................................................... 38 3.4. Cytotoxicity, Differential Toxicity, Topoisomerase Inhibition and Intercalation of the Pyridoacridines Alkaloids ........................................ 41 4. 1 . NMR Assignments for Patellamide E .................................................. 51 4.2. NMR Assignments for Tawicyclamide A ............................................. 60 4.3. NMR Assignments for Dehydrotawicyclamide A .................................... 61 4.4. NMR Assignments for Tawicyclamide B ............................................. 67 4.5. NMR Assignments for Dehydrotawicyclamide B .................................... 68 5.1. NMR Assignments for Botryllamide D ................................................ 87 5.2. Proton NMR Assignments for Botryllamides D, B, C, and A ...................... 92 LIST OF ABBREVIATIONS ABNR Adopted Basis-set Newton-Raphson BCNU I ,3-bis( chloroethyl)-l-nitrosourea BIA biochemical induction assay BNG 6-bromo-2-naphthyl-~-D-galactopyranoside cAMP cyclic adenosine monophosphate CRO Chinese hamster ovary cm collision induced dissociation COLOC correlation by long range coupling COSY correlated spectroscopy CT calf thymus oc DEPT DMSO DNA ds ED50 EI FAB FDAA RCT RIV RMBC differential cytotoxicity distortionless enhancement by polarization transfer dimethyl sulfoxide deoxyribonucleic acid double stranded effective dose 50% electron impact fast atom bombardment I-fluoro-2,4-dinitrophenyl-5-L-alanineamide human colon tumor hunlan immunodeficiency virus heteronuclear multiple bond correlation HMQC HPLC HR IC50 ID50 INEPT IR kDNA LR MD MDR MS MSIMS MTT heteronuclear multiple quantum coherence high performance liquid chromatography high resolution inhibitory concentration 50% inhibitory dose 500/0 insensitive nuclei enhancement by polarization transfer infrared kinetoplast deoxyribonucleic acid low resolution molecular dynamics multidrug resistance mass spectrometry tandem mass spectrometry 3-[4,5-dimethylthiazoy-2-yl]-2,5-diphenyltetrazolium bromide NCDDG National Cooperative Drug Discovery Group NCI National Cancer Institute NMR nuclear magnetic resonance nOe nuclear Overhauser enhancement NOESY nuclear Overhauser enhancement spectroscopy PB S phosphate buffered saline PKC protein kinase C PS-DQF phase sensitive double quantum filtered RN A ribonucleic acid ROESY rotating frame Overhauser enhancement spectroscopy SDS sodium dodecylsulfate s s single stranded TIC treated/control xiii TF A trifluoroacetic acid/trifluoroacetate TLC thin-layer chromatography TOCSY total correlation spectroscopy topoI topoisomerase I topoII topoisomerase II TPA 12-0-tetradecanoylphorbol-13-acetate IN ultraviolet XIV ACKNOWLEDGMENTS I would like to thank Prof. Chris M. Ireland for his inspiration, guidance, support, and most importantly, for allowing me the freedom to grow as a scientist. His foresight and trust enabled me to pursue my dreams and facilitated my development as a natural products scientist. I am indebted to Prof. Darrell R. Davis for his help and advise, especially for posing questions such as "Have you thought about: .. ?" when I am desperately attempting to solve an NMR problem. I am equally indebted to Mr. Jay Olsen for countless hours of help and prayers to the NMR gods. I would like to thank Prof. Louis R. Barrows for the fruitful collaboration on the pyridoacridines and for his encouragement. I am grateful to Dr. Mark P. Foster for his help, advice, collaboration, and especially for his friendship. I would like to thank Drs. Dennis R. Phillips, Elliot Rachlin, and John Peltier for acquiring my mass spectral data. I an1 thankful to all of the Medicinal Chemistry faculty for always taking the time to answer questions or give useful advice, and to my colleagues in the Ireland Research Group (past and present) for their help, friendship and understanding. Thanks Swersey, I didn't let them give me any crap. I am grateful to Bristol-Myers Squibb and Lederle Laboratories for providing biological test results. The National Institutes of Health (NIH) provided financial support through a Minority Graduate Fellowship for which I am very grateful. The research described in this dissertation was supported in part by NIH grants CA 36622 and CA 50750, awarded to Prof. Chris M. Ireland. My deepest gratitude goes to my family (present and future) for their love, support, and understanding throughout the years. Thanks Mom, I love you too. You taught me the value of hard work and persistence, which were crucial to my success in graduate school. Uncle B., Aunt G., and Anthony, I appreciate all of the help that you and the rest of the family have given me over the years. Finally, a very special thanks to Mollie for your love, friendship, support, understanding, patience, trips to the desert, etc. I could not have done it without you. You are largely responsible for my success in Utah. XVI CHAPTER 1 INTRODUCTION AND BACKGROUND Rationale for Studying Marine Organisms Oceans occupy more than 70% of the earth's surface and, in terms of volume, over 95 % of the earth's biosphere. I More than 95 % of the animal species on earth are invertebrates and marine organisms make up the majority of invertebrate phyla. 2 The commercial availability of the self-contained underwater breathing apparatus (SCUBA) has made the top layer of the marine environment (-50 m) more accessible to scientists while the depths at which marine organisms could be collected have been greatly extended by submersibles such as the Johnson-Sea-Link II.3 Marine natural products have contributed significantly to our understanding of important molecular processes such as tumor promotion4 and phosphorylation­dephosphorylation in malignancy.5 Marine organisms have provided a variety ofbioactive compounds with therapeutic potential.6 Many of these bioactive marine natural products have novel and varied molecular structures that are without terrestrial precedence,7-14 and may therefore serve as pharmacophores, which can be chemically elaborated to yield more specific and efficacious therapeutic agents. Chemical modifications of pharmacophores can widen their therapeutic window by either increasing their efficacy or decreasing their toxicity. Ecology also provides a rationale for focusing on marine organisms. 15 Because the sessile, soft bodied, shell-less marine organisms have no physical. means of protection, it has been proposed that some of these organisms use chemicals (secondary metabolites) in defense against predation. 16,17 Further arguments have been made suggesting that marine invertebrates have developed in a fairly stable environment and can therefore divert a large amount of their biosynthetic resources to developing secondary metabolites for survival, IS Some scientists have looked at the competitive interplay among various marine organisms in their natural environment and have inferred biomedical potential to the compounds used in this interaction. I? The biomedical potential of compounds derived from marine animals has inspired the research undertaken and presented in the upcoming chapters. Marine Natural Products Chemistry: Evolution from "Phytochemical" to Biomedical Discipline Research focusing on marine natural products has undergone an evolution over the 2 past three decades from a "phytochemical" based to a biomedically oriented discipline. The phytochemical approach, following the same strategy applied to plants, generally involves the isolation and characterization of compounds with little consideration of biological activity. The biomedical importance of marine natural products has been recognized and mechanism based approaches are increasingly being used to guide the isolation of novel compounds. The evolution came about because of the fruitful interaction between scientists from various disciplines such as chemistry, biochemistry, pharmacology, and molecular biology and this interdisciplinary approach has led to increasingly sophisticated methods for targeting, isolating, and characterizing novel marine natural products with biomedical potential. The advancements in the field of marine natural products chemistry are at the same time both the result of and the impetus for increasing sophistication in allied fields such as pharmacology and molecular biology. In fact, the novel structures and wide ranging pharmacological activities of marine natural products have helped to fuel this evolution and have kept scientists focused on marine organisms as a source of potential pharmaceuticals. Some early biomedical efforts in marine natural products chemistry focused on secondary metabolites such as toxins that posed significant threats to humans. 19 Examples include the effort to isolate and characterize saxitoxin (1), the agent responsible for 3 paralytic shellfish poisoning and tetrodotoxin (2), which is responsible for the fatalities caused by the puffer fish. 2o Studies using these toxins as molecular probes have subsequently made significant contributions to our understanding of the sodium channel. 21 Much research today still focuses on compounds and organisms that are human poisons, and additional interesting compounds, mostly derived from dinoflagellates or microalgae, have been uncovered and are also used as molecular probes.22 H2Ny O O· 0 HN H HN +H2~O~ •.. >== NH2+ +H2~N N-HN HN HO CH20H OH OH OH 1 2 Most early researchers, however, looked primarily at the novel chemistry that could be found by thoroughly examining new organisms, a method paralleling the phytochemical approach used in studying plants. Although the phytochemical method tended to focus on compounds that were abundant and easy to isolate, it has fortuitously led to the discovery of some biologically active secondary metabolites, as well as compounds that served as molecular prototypes in the development of biomedically useful agents. Examples of such agents include the clinically useful antiviral/anticancer compounds vidarabine (ara-A) (3);23 and cytarabine (ara-C) (4),24 which were inspired by the novel structures of arabinosyl nucleosides from sponges,25,26 and the insecticide cartap HCl (5), which was inspired by nereistoxin (6) from the annelid Lumbriconereis heteropoda.27 Regardless of the initial impetus for isolating a compound, intensive pharmacological investigations often follow isolation to establish its molecular mechanism of action. The mechanism in turn forms the basis for developing additional assays capable of detecting other similarly acting compounds. Bioactivity results fronl interaction between a compound and macromolecular 4 tN~H2N ~ .J-N-- 5I HO~ o N HO HO~ HO HO HO 3 4 'N--( SCONH2 -Hel / SCONH2 5 receptor(s). Pharmacological studies using "toxins" often lead to the conclusion that these compounds are too toxic for therapeutic use. Though therapeutically limited, these bioactive substances may still be used as probes to provide information about a specific receptor, and knowledge gained in such receptor studies may be used in designing compounds that are highly specific to that receptor. The use of compounds as probes can ultimately advance the biomedical sciences. In the case where mechanistic studies have yielded sufficient information about specific receptors involved in aberrant processes, this knowledge can also be instrumental in designing compounds to be used in therapy. The use of compounds as probes to better understand biological processes at the molecular level fills the literature and has played significant roles in the advancement of the field of marine natural products chemistry. The vastness of the marine environment, the abundance of its resources, and the potential for yielding bioactive compounds continue to attract rese~chers to the field of marine natural products chemistry. The need and desire to find more therapeutically useful compounds have spurred many of the advances in this field. Other factors contributing to the progression towards biomedically oriented research in marine natural products chemistry include fundamental advances in physics, chemistry, biochemistry, pharmacology, and molecular biology. These advances have led to the development of new bioassays that exploit a variety of organisms, cell types, cellular processes, or n10lecular targets to detect bioactive compounds. Some marine natural products have played significant roles in defining new receptor targets as noted earlier. The desire to achieve the commercial successes enjoyed by microbial natural products has led scientists to use increasingly sophisticated bioassay guidance to increase the probability of finding a therapeutically useful compound. Discovering Biomedically Significant Marine Natural Products: An Assay Based Approach Research groups looking for biomedically significant natural products from marine organisms often use a disease oriented approach, with focus on one or more therapeutic areas. Examples in the area of anticancer drug discovery will be used to chronicle the developments in natural products chemistry. Since the results from bioassays are the main determinant of which compounds will be isolated, this section has been organized in terms of the assays of increasing sophistication and specificity. A clear evolution in drug screening strategies from "compound-oriented" (or chemistry driven) to "disease-oriented" (or biology driven) can be seen as paralleling the revolutionary changes in direction undertaken by the National Cancer Institute (NCI).28 In reviewing compounds that are significant in the history of marine natural products chemistry, emphasis will be placed on those derived mainly, but not exclusively, from marine invertebrate animals. Many compounds show activities in multiple assays, and these will be discussed in the appropriate sections according to where they have made the most significant contributions. 5 6 No Assays: The Search for New Chemistry The arabinosyl nucleosides spongothymidine (7) and spongouridine (8) from the Caribbean sponge Cryptotethia crypta were of interest because they contained the rare sugar arabinose.25,26 These arabinosyl nucleosides inspired the syntheses of ara-C (4),24 a widely used antileukemic chemotherapeutic agent29 and ara-A (3),23 a clinically effective antiviral drug.30,31 The arabinosyl nucleosides have been credite~ with propelling the field of marine natural products chemistry towards biomedical significance. The discovery of 7 and 8 are examples where isolation was empirical and based solely on thin layer chromatographic (TLC) detection of a compound. This method, using no biological or mechanistic assay for directing isolation, has serendipitously led to important compounds in several therapeutic areas. o JrCH3 HN I O~N HO~ HO HO 7 8 Biological Assays Since anticancer agents are expected to act in vivo, some biological measure of effectiveness is necessary when targeting these compounds. Many of the approaches to finding potential anticancer agents use some biological assay as the endpoint; however, the majority of them are in vitro assays. Many compounds are capable of exerting multiple biological activities and some may be detected by in vitro assays while others produce responses only in in vivo assays. In some cases in vitro or in vivo activity is a good predictor of anticancer activity whereas in other instances this is not true. 7 Antinlicrobial Assays The loose correlation between in vitro antimicrobial and in vivo antitumor activities has been used to justify the use of antimicrobial assays for targeting potential anticancer compounds. A multitude of marine natural products have been isolated using antimicrobial screens because testing against bacterial strains is simple and inexpensive.32 However, since none of these compounds have progressed to the clinic as anticancer agent (or an antibiotic), it appears that antimicrobial activity is not predictive of anticancer activity. Cytostatic or Growth Inhibitory Assays The failure of antimicrobial assays to correlate with anticancer activity led some researchers to tum to the more specific cytostatic assays, which were expected to be better predictors of anticancer activity. The fertilized sea urchin egg assay or the fertilized starfish egg assay were used to detect compounds that can act as DNA, RNA, or protein synthesis inhibitors, or as antimitotic agents.33 One of the most potent compounds in the cytostatic assay, c'alyculin A (9), was isolated from the marine sponge Discodennia calyx.34 The compound inhibited the first division of fertilized starfish eggs at 0.001 JlglmL and also caused fragmentation of the DNA in fertilized sea urchin eggs.33 Calyculin A, also a potent tumor promoter, was reported to be a more potent inhibitor of protein phosphatases 1 and 2A than okadaic acid (see below) and thus, has contributed to our understanding of tumor promotion and phosphorylation-dephosphorylation in cell growth.35,36 Calyculin A was also active in vivo in the P388 and Erlich tumor models (TIC = 144 and 245, respectively). Ichthyotoxicity Assays Some compounds that are isolated as a result of their toxicity to fish appear to act by inhibiting microfilament organization. Ichthyotoxicity assays may be considered in vivo assays and appear to be a good predictors of anticancer activity. Stypoldione (10), an ortho-quinone isolated from the Caribbean brown alga OH 0 H3CO~N~°!J N OH H N H C .... 'CH 3 3 h OCH3 9 OH Stypopodium zonaie,37,38 proved to be a potent inhibitor of cell division in the fertilized sea urchin egg assay (EDso = 1.1 J.lg/mL). Stypoldione was initially identified as one of 8 the ichthyotoxic constituents of the alga and was subsequently shown to inhibit microtubule polymerization by binding to tubulin.39 Compound 10 is thought to poison sulfhydryl­dependent proteins by adding to their sulfhydryl groups.40 HO 10 HN) }-s o 11 Latrunculin A (11), a 2-thiazolidinone containing macrocyclic diterpene initially isolated from the sponge Latrunculia magnijica,41 showed in vivo activity in the A549 subcutaneous-implanted lung tumor xenograft model in mice (TIC of 146% at 0.3 9 mglkg).42 The compound was initially targeted because of its toxicity to fish. Latrunculin A binds to actin in a 1: 1 molar ratio,43 is able to alter cell shape, disrupt microfilament organization without affecting microtubules in cultured neuroblastoma and fibroblast mouse cells,44,45 and inhibits several microfilament nlediated processes.46 Other In vivo Assays In an effort to inlprove the correlation between bioassays and antitumor activity, researchers established various in vivo tumor models, using experimental animals as the initial screening step. For example, until 1985 the NCI used a leukemia mouse model (L1210 or P388) for initial prescreening.28 Although in vivo assay methods are expensive and time consuming,47 there have been some significant successes, particularly in finding compounds that are active against leukemias and lymphomas. The outstanding in vivo activity associated with extracts of the Caribbean ascidian Ecteinascidea turbinata was ultimately traced to the ecteinascidins.48-51 These compounds exhibited impressive cytotoxic and antitumor activities. The most potent, ecteinascidin 729 (12) was extremely cytotoxic (ID50 ofO.0004Ilg/rnL) towards L12IO leukemia cells52 and showed TIC of 214% at 3.81lglkg against P388 murine leukemia and 2460/0 at 10 Ilglkg against B 16 melanoma. 50 These compounds are undoubtedly excellent candidates for anticancer agents. The initial claim that the bryozoan Bugula neritina nlay contain anticancer constituents53 was strengthened by the isolation of bryostatin I (13) from this organism. 54 This macrocyclic lactone was quite cytotoxic towards P388 cells in vitro (ED50 of 0.89 Ilg/mL) and exhibited excellent activity in the P388 lymphocytic leukemia model in mice (TIC of 152-196 at 10-70 Ilglkg).54 Bryostatin 1 has shown protein kinase C (PKC) activating and antitumor promoting activities,55 and is now undergoing clinical trials as an anticancer agent. The crude extract of a Mycale sp. sponge showed 830/0 life extension of P388 HO 12 OH o o ~O 13 leukemia bearing mice relative to the control group. This in vivo activity was ultimately traced to the tricyclic amide mycalamide A (14), which showed activity in a variety of 10 o tumor models (TIC = 233,175, and 156% against M5076, B16, ~nd P388, respectively at the optimal dosages). 56 OH . OH ~ H ?~ N~OMe 0'-.../0 14 Cytotoxicity Assays Cytotoxicity assays using established tumor cells in culture have been used for many years to identify novel compounds with antitumor potential. There are many compounds that are classified as cytotoxic and these have been the subject of many recent reviews. 12- 14,57 Compounds that have shown exceptional cytotoxicity and have subsequently demonstrated in vivo activity or have been utilized to probe specific receptors will be recounted here. Didemnin B (15), a depsipeptide isolated from the colonial ascidian Trididemnum solidum, was initially targeted by shipboard assays screening for a variety of activities including cytotoxicity.58-60 Didemnin B showed promising in vivo antileukemic and antitumor activities60 and has become the first marine natural product to be evaluated in clinical trials in the United States as an anticancer agent.61 -64 Didemnin B also shows promising antiviraI58,60,65 and immunosuppressive activities.66,67 o )". r JL .. ,( ~ '0 o o o.~r'NH 0 HN N""'~ H ~ ooo~ I~ OMe 15 11 The cytotoxic peptide dolastatin 10 (16) obtained, by bioassay-directed fractionation, from the opisthobranch nlollusc Dolabella auricularia exhibited exceptional in vivo activity in the PS leukemia (TIC of 169-202% at 1-4 Jlg/kg) and B16 melanoma (TIC of 142-238% at 1.44-11.1 Jlg/kg) models.68 Dolastatin 10 was cytotoxic towards L1210 murine leukemia cells with an ICso value of 0.3-0.9 nM and resulted in mitotic arrest of the leukemic cells.69,70 Mechanistic studies showed that dolastatin 10 inhibited tubulin polymerization and microtubule assembly. 69 The conlpound also reportedly inhibits tubulin dependent GTP hydrolysis and binds noncompetitively to the vincristine binding site on tubulin, thus adding to the small number of tubulin inhibitors binding at this site.?l 16 H N 12 The pyridoacridine alkaloid, dercitin (17), from a Bahamian Dercitus sp. sponge,72 showed in vitro (P388 1Cso: 0.05 Jlg/mL) and in vivo activity in the P388 mouse leukemia model (TIC of 170 at 5 mg/kg). Additionally, the compound was shown to disrupt DNA and RNA synthesis and intercalate into DNA.73 Dercitin is among a large group of cytotoxic pyridoacridine alkaloids (Chapter 3) thought to exert their cytotoxicity by intercalating into DNA and inhibiting the function of nucleic acid specific enzymes.74 17 HO 18 Okadaic acid (18), initially isolated from the marine sponges Halichondria okadai and H. melanodocia,75 was toxic towards P388 and L1210 cells with EDso's of 1.7 x 10-3, and 1.7 x 10-2 Jlg/mL, respectively. Okadaic acid was subsequently isolated from the marine dinoflagellate Prorocentrum lima.76 Although the cytotoxicity of this metabolite 13 initially attracted attention, this compound, by virtue of its tumor promoting and phosphatase inhibitory activities, has made vast contributions to our understanding of cell regulation (see below). Selective Cytotoxicity The approaches using single cell lines or in vivo tumor models (e.g., murine leukemias and lymphomas) have yielded several potential antileukemic agents, but the majority of the compounds found using these methods are also cytotoxic towards normal cells. The past decade has seen a tremendous emphasis placed on assays capable of detecting compounds that are selectively cytotoxic towards certain solid tumors.28 These efforts resulted in batteries of human tumor cell lines such as NCr s panel of over 60 tumor cell lines, representing more than 10 different cancer types. Compounds demonstrating selectivity in this assay are tested against the sensitive human tumor xenografts in nude mice. It is also expected that this approach will give rise to patterns of cytotoxicity that could correlate with specific mechanism(s) of action.28,47 - H H H H H HO,; H H H H HO,;. '. "..C ( HO I "*. a 19 The polyether macrolide halichondrin B (19), initially isolated from the marine sponge Halichondria okadai, showed remarkable in vivo activity in the P388 lymphocytic leukemia (TIC of 323% at 10 Jlglkg) and B16 melanoma (TIC of 244% at 5 Jlglkg) 14 nlodels.77 Renewed interest in this compound stemmed from analysis of its differential cytotoxicity pattern in the NCI panel of cell lines which predicted that 19 would be an antimitotic agent. 78 Confirmation of this prediction came when halichondrin B was shown to inhibit glutamate-induced tubulin polymerization. Halichondriri B caused accumulation of L12l0 cells arrested in mitosis at the cytotoxic concentration of 0.3 nM.78 Halichondrin B thus confirms the importance of the differential cytotoxicity assay as a predictor of potential anticancer agents. The antimitotic agent, curacin A (20), from the cyanobacterium Lyngbya majuscula, was initially detected because of its toxicity to brine shrimp and antiproliferative activity. Curacin A, when tested in the NCI panel of tumor lines, revealed a cytotoxicity pattern consistent with an antimitotic agent and was subsequently shown to be a colchicine class tubulin inhibitor.79 20 Cytotoxicity Assays with Molecular Target Defined (DNA Damage Repair Assays) Toxicity towards cells in which the molecular target is known can be considered as a demonstration of selective cytotoxicity. Enhanced toxicity towards DNA repair deficient Chinese hamster ovary (CHO) cells, relative to the repair proficient variants, is an example of differential or selective cytotoxicity. Compounds showing activity in some molecular-target -defined assays are believed to interact with nucleic acids or may have topoisomerase (topo) inhibiting activity. Examples of marine natural products that are active in the differential cytotoxicity assay include the makaluvamines and wakayin. The makaluvamines, from the Fijian sponge Zyzzya sp., showed enhanced toxicity 15 towards DNA double strand break repair deficient xrs-6 CHO cells relative to the repair competent BRI strain and are (topoisomerase II) topoll inhibitors. 80,81 For exanlple, makaluvamine F (21) showed sixfold enhanced toxicity toward xrs-6 relative to BRI and an IC90 of 25 JlM in the topoll decatenation inhibition assay.80 Wakayin (22), a pyrroloiminoquinone alkaloid, isolated from a Clavalina sp. ascidian,82 showed 9.8-fold enhanced toxicity towards xrs-6 relative to BR 1 CHO lines and inhibited topoll catalyzed decatenation of kDNA at 250 JlM. 83 o 21 H N Br OH 22 Another cell line used to detect compounds acting by a specific mechanism is the E. coli BR513 strain that is specifically engineered to detect DNA damaging compounds in the biochemical induction assay (BIA). Eilatin (23) is the only reported example of a marine natural product active in this assay (Chapter III). Eilatin causes induction of~- galactosidase at 2 J.1g/disk in the BIA.74 23 16 Mechanism-Based Assays Mechanism-based approaches are increasingly being employed in the search for novel, potential chemotherapeutic agents. Very specific biochemical assays are used to find cytotoxic compounds that work by mechanisms that are similar to those of established drugs. Many compounds have affinity for certain cellular receptors in mammalian cells and most of them ultimately exert their activity via interaction with one or more of these receptors. In fact, it is believed that all secondary metabolites evolved under selective pressure from some receptor. 84 It is anticipated that greater efficacy and cell type selectivity may be achieved by employing mechanism-based approaches for discovering compounds that target receptors or enzymes involved in the pathogenesis of neoplastic diseases. The use of these assays stems from advances in the understanding of tumor biology and biochemistry made over the past decades. The assays are inherently more specific than those using whole cells or animals. Although this section focuses on the use of mechanism-based assays, it also includes compounds discovered using more traditional approaches but later shown to have some effect at a specific receptor. Studies using these compounds have either demonstrated the importance of the receptor or have helped to initially identify it. Topoisomerase Inhibition Assays Recently, much attention has been focused on topoisomerase inhibitors since many clinically useful drugs have been shown to act by inhibiting topoII, an enzyme crucial to cell proliferation. Marine natural products that inhibit topoll include the makaluvamines,80,81 pyridoacridines (Chapter 3),74,85 and wakayin.83 There are many topoll inhibitors known, but there are relatively few topoisomerase I (topoI) inhibitors.86 Since topoI was shown to be the target of the antitumor plant alkaloid camptothecin and its derivatives,87 intensive efforts are also underway to find new inhibitors of this enzyme. Antimitotic Agents (Microtubule Assembly-Tubulin Polymerization Inhibition Assays) 17 Antimitotic agents that bind to tubulin and interfere with microtubule assembly often result in mitotic arrest. 88 Some of the more promising antitumor agents acting by this mechanism include dolastatin 10 (16)68 and halichondrin B (19)77 discussed earlier. Nucleotide Biosynthesis Inhibition Assays Marine natural products that affect de novo nucleotide biosynthesis or their incorporation into DNA are not common. However, the previously discussed arabinonucleosides (7-8), credited with propelling the field of marine natural products chemistry towards prominence in the 1950s and synthetic analogs, act by this mechanism.31 Signal Transduction Inhibition Assays Macromolecules involved in signal transduction during cell growth provide excellent targets for inhibition. One of the most intensively studied molecules in the signal transduction pathway, protein kinase C (PKC), has become an important target in the effort to stem malignant transformation. Mechanism-based assays designed to detect compounds that perturb signal transduction pathways are being employed in marine natural products chemistry. For example, latrunculin A (11) showed activity in an EL-4:IL-2 cell adhesion assay that correlates with PKC agonism or antagonism. Latrunculin A reduces EL-4.IL-2 cell adhesion presumably by antagonizing the activity of PKC and showed enhanced cytotoxicity towards the A549 non-small cell lung carcinoma line, relative to the P388 leukemic line.42 Some pyridoacridine alkaloids (e.g., eilatin (23)) from a Eudistoma sp. ascidian were recently reported to exhibit growth regulatory properties (inhibition of proliferation and induction of differentiation and reverse transformation of transformed cells), presumably by acting on the cyclic adenosine monophosphate (cAMP) signaling system.89 DN A Interactive Agents 18 Compounds that interact with DNA are among the most useful anticancer compounds in the clinic. These compounds interact with DNA in a variety of ways including intercalation, alkylation, and noncovalent binding. Many marine natural products are screened for their ability to intercalate into DNA as part of routine evaluation of their mode of action. Members of the pyridoacridine class (e.g., dercitin (17) and eilatin (23))73,74,90 and the makaluvamines80 are examples of marine natural product intercalating agents. Protein Phosphatases and Tumor Promotion Protein phosphatases have been shown to be important in malignant transfom1ations, and apparently act as tumor suppressors in vivo by maintaining the proper level of protein phosphorylation.s Although okadaic acid (18), one of the toxins responsible for diarrhetic shellfish poisoning,22 has not shown much potential for clinical utility, it has proven to be a valuable probe in defining the role of phosphorylation-dephosphorylation in malignant transformations, and as a probe for specific protein phosphatases:91 The phorbol ester TPA92 (24) promotes tumors in animals exposed to carcinogens by activating PKC, whereas okadaic acid, a non-TPA type tumor promoter, acts by a mechanism not involving activation of PKC.4 The toxin 18 hinders dephosphorylation by inhibiting protein 19 phosphatase activity.93 Although not a TPA-type tumor promoter, okadaic acid, by inhibiting protein dephosphorylation, causes accumulation of the same phosphorylated proteins believed to be involved in tumor promotion.94 Okadaic acid is routinely used as a probe for classifying serine-threonine protein phosphatases as either type 1, 2A, or 2B, based on the concentration of compound required for inhibition of phosphatase activity.91 Other compounds such as motuporin (25),95 microstatin-LR (26),96 calyculin A (9),35 and palytoxin (27)22 have also shown phosphatase inhibitory activity. H .0. ~ N HYiHO N N N~ o 0 25 26 20 OH OH HO, o 0 = OH HO ". HO~N~N~"'" H H O=H O=H OH -OH - OH 27 Summary Although marine derived natural products have been intensively studied by chemists over the past 35 years, their biomedical potential was not fully recognized until recently. The early "phytochemical" approach to discovering biomedical marine natural products have yielded several important leads; however, it has been largely supplanted by a more targeted approach which relies heavily on the use of increasingly sophisticated biological and mechanism-based assays. The success of the biological assay guided approach is evidenced by the increased number of potentially useful marine natural products discovered through the use of such assays.7-14 Although the use of mechanism-based assays for compound screening is on the rise, these types of assays are, at present, unlikely to replace the more traditional biological assays such as in vitro cytotoxicity and in vivo tumor models, which remain the primary tools for the discovery of potential antineoplastic compounds. This is because mechanism-based screens are, by design, expected to detect 21 only a very narrow range of compounds, while biological assays are broad based screens. There are several advantages to using whole cell or animal assays over mechanism­based assays. First, a compound showing activity in a cell based assay has exerted its effects by presumably binding to some receptor. Since the desired effect has been obtained in the form of activity, it is not necessary to isolate or characterize that receptor. Second, activity in such assays tends to indicate that the compound is capable of crossing the cell membrane. Third, since the cell contains a multitude of receptors against which a con1pound can potentially act, it is not necessary to use a bank of different enzymes for screening. It is therefore important to use a combination of approaches in the search for novel, potentially useful therapeutic agents from marine organisms. Although there has been a tremendous increase in our understanding of the etiology of cancer from a molecular perspective during the past decade, effective treatment for many solid tumors remains elusive. Thus, many research groups have active programs directed towards finding useful anticancer chemotherapeutic agents. A variety of approaches using a multitude of assay methods are undertaken. The efforts of the Ireland Research Group are directed toward the goal of finding potential anticancer agents and the approaches used are described in the next chapter. The projects described in this dissertation are aimed at finding active or novel compounds with anticancer potential, or those that are potentially useful as molecular probes. CHAPTER 2 STRATEGIES FOR DISCOVERING POTENTIAL ANTICANCER AGENTS: BIOASSAY GUIDED FRACTIONATION The goals of my research are to identify and characterize new biologically active conlpounds from marine organisms; and in some instances, to establish the mechanism of action of these compounds. The research strategy begins with the choice of organism. In the Ireland Research Group, there is a bias towards the phylum Chordata (subphylum: Urochordata (= Tunicata), class: Ascidiacea (= ascidians or sea squirts)) consisting of invertebrate animals that are exclusively marine and predominantly sessile. The strategy for discovering active compounds involves the use of a variety of biological assays ("bioassays") to initially identify and subsequently guide the fractionation of active extracts. Bioassay guidance ultimately leads to the component(s) of the crude extract that initially gave rise to the activity. This research is oriented towards the isolation of potential anticancer agents; consequently, the assays used are designed to detect compounds that kill or retard the growth of tumor cells. These assays range from relatively simple antimicrobial assays to assays for cytotoxicity, selective cytotoxicity, in vivo activity, and enzyme inhibition. By its very nature, the research conducted in the Ireland Research Group is interdisciplinary and therefore involves several collaborators. This chapter discusses the various approaches used in our group for identifying potential anticancer agents. 23 In Vitro Assays Antimicrobial Assays Antimicrobial assays determine if a crude extract or pure compound has antibiotic activity against a given organism. Organisms such as the yeasts Candida albicans and Saccharomyces cerevisiae and the bacteria Bacillus subtilis and Pseudomonas aeruginosa have been used in our laboratory. Antimicrobial activity against a specific strain of Escherichia coli used in the BIA may also be assessed in addition to induction (see below). These assays are performed using the disk diffusion method. Due to poor correlation between antimicrobial and antitumor activities, these assays are considered of limited utility. However, when a clear correlation exists between antimicrobial and in vivo activity, the antimicrobial assay may be used to guide fractionation. Cytotoxicity Assays The ability of a compound to inhibit the replication of various cell lines is evaluated using the MTT assay.97-99 This work is done in the laboratory of Prof. Louis R. Barrows of the Department of Pharmacology and Toxicology. Several human tumor cell lines derived from various tissues and one leukemic line are used (Table 2.1). Selective cytotoxicity may be assessed using this panel of cell lines. Table 2.1 Human Cell Lines used in Natural Products Screening Cell Line HCTl16 A498 MCF-7 PC-3 SK-MEL-1 A549 HL-60 Type Colon cancer Kidney carcinoma Breast adenocarcinoma Prostate adenocarcinoma Skin (malignant melanoma) Lung adenocarcinoma Blood (promyelocytic leukemia) 24 Via collaboration with Prof. Graydon Harker, a number of mitoxantrone resistant human cell lines are also available to evaluate pure compounds that have shown selective cytotoxicity or enhanced toxicity towards xrs-6 CHO mutants discussed below. Table 2.2 contains a description of these multidrug resistant (MDR) cell lines. Cell Lines Sensitive to DNA Damaging Agents Our interest on anticancer compounds has led to specifically focusing those agents that interact with DNA. A mechanism based screening protocol to discover cytotoxic compounds that act by mediating topoisomerase activity or DNA strand breaks has been employed. These assays are often used as secondary screens to test compounds or extracts showing antimicrobial or cytotoxic activity. CHO Differential Cytotoxicity Assays Single strand (ss) break repair-deficient EM9 CHO cells are sensitive to topoI inhibitors (e.g., camptothecin). This cell line lacks a functional XRCCl gene and is deficient in its ability to repair DNA single strand breaks. 100 Double strand (ds) break repair-deficient xrs-6 CHO cell lines are sensitive to compounds that cause ds breaks either directly or by inhibiting topoll. By stabilizing cleavable complexes and inhibiting the Cell Line (origin) MXl (from CEM) Table 2.2 Multidrug Resistant Cell Lines used in Evaluating Pure Compounds Genetic Alteration Characteristics Reduced levels of nuclear topon and Mitoxantrone resistant and absence of the P isoform. reduced sensitivity to topon poisons. MX2 Reduced levels of topon due to Mitoxantrone resistant and reduced sensitivity to topoH poisons. (from HL-60) absence of the P isoform and attenuated topoH activity. LK2/10 (from LK2) Increased p-glycoprotein levels. Mitoxantrone resistant. General MDR resistant. 25 rejoining reaction catalyzed by the enzyme, topoII poisons result in ds breaks. 101,102 The UV20 CHO line is deficient in its ability to repair bulky DNA adducts or crosslinks. Since many clinically useful compounds danlage DNA or interfere with its replication, compounds isolated in our laboratory are tested for their ability to exert enhanced toxicity towards the ss-break repair deficient EM9, the excision repair deficient UV20, or the ds-break repair-deficient xrs-6103 CHO mutant line relative to the repair­proficient BRl 100 cell line. Toxicity is measured using a modification of the MTT tetrazolium salt colorimetric assay used to determine HCT cytotoxicity.97-99 This work is also conducted in the laboratory of Prof. Barrows and a number of drugs acting by known mechanisms have been used to validate this assay. 83 Table 2.3 summarizes the genetic alterations and characteristics of the CHO cell lines used in screening crude extracts or to evaluate pure DNA active compounds. Escherichia coli responds to various adverse conditions by inducing SOS functions. 105 Agents responsible for this induction include genotoxic agents, mutagens or compounds that inhibit DNA replication. The mechanism that appears to mediate the induction response involves the activation of the recA protease by binding to single strand DNA (or DNA-degradation products). The recA protease then cleaves a repressor molecule, the lexA product, which in tum causes the cell to make more recA protein and other enzymes involved in DNA repair. 106 Cell Line EM9 xrs-6 UV20 BRl Table 2.3 CHO Cell Lines used in Screening and Mechanistic Studies. Genetic Alteration Reduced XRCCl gene expression. DNA ds break repair deficient. Lacking DNA excision repair gene. Elevated levels of 06-alkylguanine­DNA- alkyltransferase. Characteristics Sensitive to topol poisons Sensitive to topoll poisons and agents that cause ds-DNA breaks. Sensitive to DNA alkylation (bulky adducts and DNA crosslinks). DNA repair proficient (BCNU104 resistant) 26 Biochemical Induction Assay (BIA) The repressors of lytic growth of some bacteriophages, such as A phage, are also cleaved by the recA protease, thereby inducing lytic growth under adverse conditions such as assault to the host's DNA. The BIA uses a A-lacZ lysogenic E. coli strain. E. coli that are lysogenic for a A-lacZ fusion phage produce p-galactosidase, a product of the lacZ gene, on induction of the prophage by DNA damaging agents. 107 Since many useful anticancer agents act by interacting with DNA, the BIA is expected to detect such agents when used to screen crude extracts. The BIA is carried out by Leonard A. McDonald. Induced levels of p-galactosidase are a direct response to SOS-induced A gene expression. Quantification of p-galactosidase is accomplished colorimetric ally by detecting the cleavage product of the 6-bromo-2-naphthyl-p-D-galactopyranoside (28) substrate using fast blue RR salt (4-benzoylamino-2,5-dimethoxybenzenediazonium chloride hemi[zinc chloride] salt, (29)).108 p-Galactosidase is essential in lactose metabolism. The normal function of p-galactosidase is to metabolize the disaccharide lactose to galactose (30) and glucose. E. coli uses galactose as its sole carbon source under normal conditions. The cleavage of the substrate 28 is a p-galactosidase catalyzed hydrolysis of the p-glycosidic bond (Figure 2.1). Coupling between the bromonaphthol (31) cleavage product and the diazonium salt, 29, produces the highly colored azo dye (32), which may be quantified. Aromatic rings that couple to diazonium salts must contain good electron releasing "activating" groups such as -OH, -NH2, -NHR or -NR2 since diazoniums are only weakly electrophillic. The substitution generally occurs ortho to the activating group as in Figure 2.1. Mechanistic Assays Topoisomerase Assays The topoisomerase enzymes are common targets for many compounds that ultimately affect DNA replication. 109-1 12 DNA topological transformations such as catenation- 27 H~CH2~I ~-CQ~ ~-Galactosidase OH 0 0 ~ Br OH 28 Br OCH3 < }-tNH) }-N=N 32 OCH3 HO Figure 2.1. Chemistry involved in the biochemical induction assay. decatenation and unwinding are controlled by topoisomerases. 113 The three-dimensional structure of DNA regulates many of its functions within the cell. Topoisomerases are ultimately involved in regulating the tertiary structure of DNA and are instrumental in maintaining chromosome structure and separating daughter chromatids during mitosis. 113,114 TopoII acts by transiently breaking both strands of double helix DNA, thereby allowing the topological transformations that are necessary for DNA replication. 114 The high levels of topoII that exist in proliferating cells versus the lower levels present in nonreplicating cells result in some topoII inhibitors preferentially killing cancer cells, 112 presumably because of the greater need for cell replication machinery in the former. Drugs that interact with nucleic acids are therefore among the most useful cancer chemotherapeutic agents. 1 IS Intercalating compounds such as acridines, ellipticines, anthracyclines, actinomycins, and anthracenediones are known to induce topoII-mediated double strand breaks in DNA.I09-112 Many intercalators are thought to stabilize the covalent enzyme-DNA "cleavable" complex by interfering with the DNA rejoining step catal yzed by topoII. 116,117 The anticancer activity and mode of action of the camptothecin analog topotecan, a topoI inhibiting drug, have increased interest in designing assays ~o find other topol pOisons.87 Topol is thought to be involved in gene transcription and DNA replication in normal cells. The enzyme acts by breaking and religating only a single DNA strand. The topoisomerase assays are carried out in the laboratory of Prof. Barrows. Decatenation Inhibition and DNA Relaxation Assays 28 The decatenation inhibition assay measures a compound's ability to inhibit the topoII­catalyzed release of linear and circular DNA monomers from high molecular weight kDNA.118 The released DNA monomers can be resolved and quantitated on a 1 % agarose gel while the fully catenated kDNA is too large to penetrate into the gel. Enzyme inhibition is measured by monitoring the disappearance of the monomer-length DNA bands as a function of increasing compound concentration. A DNA relaxation assay is used to detect topol poisons. This assay measures the ability of a compound to inhibit the topoI catalyzed relaxation (unwinding) of supercoiled pBR322 DNA.l19 The decatenation inhibition and DNA relaxation assays are carried out in the laboratory of Prof. Barrows. Cleavable Complex Assay The K+-SDS precipitation assay (performed in Prof. Barrow's laboratory) is used to assess the involvement of cleavable complex formation in topoll inhibition. The assay is used to measure covalent DNA-protein complexes in whole cells by measuring precipitable radioactivity. Intercalation Assay The intercalation assay is designed to provide information about the mechanism of action of a compound. Compounds that bind DNA often do so by intercalation. The ethidium bromide displacement assay shows the ability of a compound to intercalate into DNA. When intercalated into DNA, ethidium bromide exhibits dramatically enhanced 29 fluorescence along with a shift of its emission maximum.120 When a competing intercalator displaces ethidium bromide from DNA, the fluorescence of ethidium bromide decreases. The fluorescence of ethidium bromide determined in the presence of DNA as a function of compound concentration provides a measure of intercalative ability of the compound. Inhibition of Macromolecular Synthesis In an effort to establish molecular mechanism of action, compounds that are good intercalators or topoH inhibitors are often tested for their effects on the rates of DN A, RNA, and protein synthesis. These tests are done in the laboratory of Prof. Barrows. The inhibition of incorporation of the radiolabelled precursors [3H]leucine (protein), [3H]uridine (RNA), and [3H]thymidine (DNA) into HeT or BRI cells is used to assess the effects of these compounds on macromolecular synthesis. In Vivo Assays In vitro active compounds are tested for antileukemic or antitumor activities in the P388 leukemia or ovarian carcinoma (Ovcar3) tumor models at Lederle Laboratories (Pearl River, NY). CHAPTER 3 THE CHEMISTRY AND BIOLOGY OF THE ASCIDIAN CYSTODYTES SPa This chapter describes the isolation and characterization of a series of pyridoacridine alkaloids from a Fijian Cystodytes Spa (Order: Aplousobranchia, Family: Polycitoridae (colonial)) ascidian: dehydrokuanoniamine B (33), shermilamine C (34), cystodytin J (35), and the known compounds cystodytin A 121 (36), kuanoniamine D122 (37), shermilamine B123 (38), and eilatin90 (23). These compounds and the previously reported diplamine124 (39) were evaluated in order to establish their mechanism of action. Initial interest in this organism stemmed from activity noted in the biochemical induction assay (BIA)108 and from the observed cytotoxicity of the crude extract against the human colon tumor cell line HCTl16. Pyridoacridine alkaloids derived from ascidians and sponges are reported to exhibit a variety of interesting properties including DNA intercalation,?3,125 topoll inhibition,85 calcium release,121 anti HIV,126 and in vivo antitumor?3 activities. The pyridoacridines discussed herein exhibited varying degrees of cytotoxicity towards HCT cells and were subsequently shown to inhibit the function of topoll and to intercalate into DNA. The ability of the pyridoacridines to inhibit topoll correlated with their ability to intercalate and inhibit the growth of HCT cells. Their capacity to bind DNA, inhibit topoII, disrupt DNA and RNA synthesis, and inhibit replication of HCT cells is consistent with a cytotoxicity mechanisD1 involving intercalation and prevention of topoll from binding to its DNA substrate. This proposed mechanism for pyridoacridine-induced inhibition of HCT cell replication is exa~ned in this chapter. 33 R= 37 R= u ~ 16 ~18 17 2) 0 l~ 6 u 34 R= ~ 17 ~19 18 21 0 38 R= ~~ 0y HN o 39 Chemistry 14 RHN 35 R= 36 R= Isolation of the Pyridoacridine Alkaloids 0 0 l~ 16 ~U Shipboard screening using the BIA 108 showed that the MeOH extract from a purple fleshy Cystodytes sp. ascidian 127,128 was capable of interacting with DNA. lOS Subsequent bioassay guided fractionation of the MeOH extract of ~he frozen animal led to the isolation of a bright yellow alkaloid which was solely responsible for the BIA activity of the crude extract. This compound was spectroscopically identical to eilatin (23), a dibenzotetraazaperylene pyridoacridine previously isolated from a Red Sea Eudistoma sp. ascidian.9o Further examination of the extract led to the isolation of several additional 31 pyridoacridines (23, 33-38) which were inactive in the BIA, but showed significant cytotoxicity toward HCT cells in vitro. 32 The numbering scheme for compounds 33, 34, and 35 parallels those published for compounds 37, 38, and 36 respectively. Structure Elucidation of Dehydrokuanoniamine B Compound 33 was obtained as an orange amorphous solid. High resolution F AB mass measurement provided the molecular formula C23H21N40S. Due to its limited solubility in several deuterated solvents, 33 was converted to its TFA salt for subsequent NMR studies. The proton, carbon, and COSY spectra of 33 are shown in Appendix A. The extensive conjugation and heteroaromatic nature of 33 was evident from its UV spectrum and blue shift upon addition of acid. Analysis of the NMR data (Table 3.1) revealed four isolated spin systems, two exchangeable signals and one downfield aromatic singlet. The signals at 8.27 (H4, d, J = 8.2 Hz), 7.27 (H5, ddd, J = 8.2, 5.3, 2.1 Hz), and 7.72 ppm (H6 & H7, m) in the 1 H NMR spectrum (DMSO-d6) were assigned to a 1,2- disubstituted benzene ring. Another spin system consisting of signals at 8.46 (H2, d, J = 6.5 Hz) and 7.87 ppm (H3, d, J = 6.5 Hz) was indicative of a trisubstituted pyridine ring. A very strong nuclear Overhauser enhancement (nOe) correlation between 7.87 (H3) and 8.27 ppm (H4) in the ROESy129,130 spectrum of 33 facilitated as.signment of these proton signals to adjacent rings (Figure 3.1). The proton detected heteronuclear multiple bond correlation (HMBC) 131 experiment showed correlations between the quaternary carbon signal at 114.13 ppm (C3b) and the exchangeable proton signal at 11.71 ppm (H8) and also to the aromatic signals at 7.27 ppm (H5) and 7.87 ppm (H3) (Figure 3.2). Additional HMBC correlations between the 118.25 ppm (CI2c) quaternary carbon and H8 as well as H3 were indicative of a pyridoacridine. Signals at 3.31 (HI4, dt, J = 5.3 7, 3 Hz) and 3.12 ppm (HI3, t, J = 7.3 Hz) indicated a pair of coupled methylene groups, whereas those at 5.63 (HI7, septet, J = 1.1 Hz), 2.15 33 Table 3.1 NMRa Assignments for the TFA Salt of Dehydrokuanoniamine Bin DMSO-d6 Atom no. B l3C B IH (mult., J (Hz)) HMBCb correlations 2 142.72 8.46 (d, 6.5) C3, C3a, C12b 3 107.57 7.87 (d, 6.5) C2, C3b, C 12c 3a 148.50 3b 114.13 4 125.53 8.27 (d, 8.2) C3a, C6, C7a 5 122.98 7.27 (ddd, 8.2, 5.3, 2.1) C3b,C6,C7 6 135.30 7.72 (m) C4, C7a 7 117.51 7.72 (m) C3b,C5 7a 140.48 8 11.71 (bs) C3b, C12c 8a 132.80c 9 108.44 9a 143.15 11 153.83 9.43 (s) G9a ,d C12a 12a 132.74c 12b 131.80 12c 118.25 13 30.98 3.12 (t, 7.3) C8a, C9, C9a, C14 14 36.19 3.31 (td, 7.3, 5.3) C9, C16 15 8.35 (t, 5.3) C14, C16 16 167.84 17 118.08 5.63 (septet, 1.1) C16, C18, C19, C20 18 150.30 19 19.50 2.15 (d, 1.1) C17, C18, C20 20 26.84 1.79 (d, 1.1) C17, C18, C19 a Proton and carbon data were acquired at 500 and 125 MHz, respectively. b The HMBC experiment was optimized to observe n JCH couplings of 8.5 Hz. c Interchangeable assignments. d Seen only when the HMBC experiment was optimized to observe 5 Hz n JCH couplings. 7.2 7.4 7.6 7.8 8.0 8.2 8.4 @ ~ • • 4 H5-H4 H5-H6 .. 0 <iiilC 4- CIt - 0° H3-H2 H3-H4 <is) ~ Occ> CD ~ 8.6 8.4 8.2 8.0 7.8 7.6 7.4 7.2 7.0 F1 (ppm) Figure 3.1. Region of a 300 nlS ROES Y spectrum of dehydrokuanoniamine B showing the H3-H4 crosspeak. Figure 3.2. Dehydrokuanoniamine B with HMBC correlations used in assigning the quaternary carbon atoms. 34 35 (HI9, d, 1 = 1.1 Hz), and 1.79 ppm (H20, d, 1 = 1.1 Hz) suggested an isobutenyl group. In addition to the downfield shift of C20, 121 the presence of a significant ROESY correlation between HI7 and H20 established that they were cis relative to each other. These data, plus a singlet at 9.43 ppm (HI 1), suggested that the compound belonged to the kuanoniamine class122,125 and was in fact dehydrokuanoniamine B. Assignments of C9a and C 12a were based on HMBC correlations to H 11 and chemical shift comparisons with literature values for these carbons.122 The regiochemistry of the nitrogen and the sulfur of the thiazole ring and the correct assignments of the adjacent carbons have been firmly established for the kuanoniamine class.72,122,125 Carroll and Scheuer used the benzothiazole model to show that electron delocalization and conjugation across the Hll-C11-N12-C12a bonds facilitated a large long range proton-carbon coupling between the highlighted atoms (3lC12a-Hll = 13.8 Hz) (see structure 33 for numbering scheme). The long range proton-carbon correlation across the Hll-C11-S10-C9a bonds is less (3lC9a-HII < 5 Hz).122 Although only the C12a-H11 correlation is observed when the HMBC experiment was optimized for observing lllCH of 8.5 Hz, both the C12a-H11 and C9a-H11 correlations were observed when the experiment was optimized to observe lllCH of 5 Hz. Structure Elucidation of Shermilamine C The NMR data for shermilamine C (34), contained in Table 3.2, were almost identical to those reported for shermilamine B (38),123,132 with the greatest differences occurring in the regions of the spectrum corresponding to the side chain. The proton, carbon, and COSY spectra of 34 are shown in Appendix A. Thorough examination of these data revealed that 34 possessed the same isobutenyl side chain found in 33. The HMBC correlations instrumental in assigning the quaternary carbon atoms of shermilamine C are shown in Figure 3.3. 36 Table 3.2 NMRa Assignments for the TFA Salt of Shemlilamine C in DMSO-d6 Atom no. o 13C o IH (mult., J (Hz)) HMBCb correlations 2 146.06 8.35 (d, 5.9) C3, C3a, C13b 3 105.92 7.45 (d, 5.9) C2, C3b, C13c 3a 144.48 3b 114.74 4 124.78 8.04 (d, 7.9) C3a, C6, C7a 5 122.22 7.11 (dd, 7.9, 7.9) C3b, C7 6 133.78 7.55 (dd, 7.9, 7.9) C4, C7a 7 116.97 7.51 (d, 7.9) C3b,C5 7a 140.09 8 11.37 (bs) 8a 131.45 9 111.46 9a 127.97 11 29.49 3.58 (s) C9a, C12 12 164.27 13 9.64 (bs) C9a, C11 13a 117.86c 13b 130.56 13c 117.14 14 27.97 2.89 (t, 7.2) C8a, C9, C9a, C15 15 36.61 3.08 (m) C17 16 8.56 (t, 5.7) C15, C17 17 168.31 18 117.92c 5.70 (t, 1.1) C17, C20, C21 19 150.81 20 19.60 2.22 (s) C18, C19, C21 21 26.95 1.84 (s) C18, C19, C20 a Proton and carbon data were acquired at 500 and 125 MHz, respectively. b The HMBC experiment was optimized to observe nJCH couplings of 8.5 Hz. c Interchangeable assignments. o Figure 3.3. Structure of shermilamine C with HMBC correlations used in assigning the quaternary carbon atoms. Structure Elucidation of Cystodytin J The iminoquinone nature of 35 was initially revealed by its behavior during F AB mass spectrometry. FAB mass spectrometry in a reducing matrix such as glycerol produced an ion at mlz 320 corresponding to the reduced molecular ion ((M+2)+H)+. A similar experiment in 3-nitrobenzyl alcohol, a less reducing matrix, produced ions 37 predominantly at mlz 318 (M+H)+ but also some at mlz 320. This behavior has been reported for the iminoquinone cystodytins by Kobayashi et al. 121 HR FAB MS provided the molecular formula C19H18N302 for the reduced form of 35. The proton, carbon and COSY spectra of 35 are shown in Appendix A. The l3C spectrum showed 18 distinct resonances. The HMBC experiment later revealed that C 1 and C2 were degenerate, both resonating at 131.87 ppm. Examination of the NMR data (Table 3.3) revealed the presence of a 1,2-disubstituted aromatic ring, a polarized pi bond in a six membered aromatic ring, an isolated olefinic proton, two coupled methylene groups -one of which was further coupled to an exchangeable proton, and a methyl singlet. A strong NOESY correlation between H4 and H5 placed the HI-H2-H3-H4 spin system of the 1,2-disubstituted aronlatic ring adjacent to the H5-H6 polarized pi bond. 38 Table 3.3 NMRa Assignments for Cystodytin J in CDCl3 Atom no. () l3C () IH (mult., J (Hz)) HMBCb correlations 1 131.87 8.29 (d,8.1) C3, C4a, Cl1a 2 131.87 7.94 (ddd, 8.1,7.1, 1.4) C4, Cl1a 3 129.83 7.83 (ddd, 8.1, 7.1, 1.4) ClfC2, C4a 4 122.84 8.42 (d, 8.1) C2, C4b, Clla 4a 121. 78 4b 136.92 5 118.99 8.18 (d, 5.5) C4a, CI0b 6 149.76 8.94 (d, 5.5) C4b, C5, C7a 7a 146.48 8 183.31 9 132.82 6.85 (s) C7a, CI0a, C12 10 152.17 lOa 150.33 lOb 117.84 Ila 145.32 12 31.72 3.25 (t, 6.4) C9, CI0, CI0a, C13 13 39.28 3.79 (dt, 5.9, 6.4) CI0, C12, CI5 14 6.59 (br) C13, C15 15 170.43 16 23.30 2.02 (s) C15- a Proton and carbon data were acquired at 500 and 125 MHz, respectively. b The HMBC experiment was optimized to observe nJCH couplings of 8.5 Hz. 39 HMBC correlations (C4a-H3, C4b-H4, C4a-H5, and C4b-H6) assigned the intervening quaternary carbons C4a and C4b. The isolated olefinic proton H9 was determined to be a part of the iminoquinone ring and was positioned relative to the methylene side chain via HMBC correlations as indicated in Table 3.3 and Figure 3.4. HMBC correlations between H12 and CI0 established the connection between the methylene side chain and the iminoquinone ring. The spectral data for 35 were virtually identical to those reported for 36. 121,133 The acetyl side chain by which 35 differed from other cystodytins (e.g., 36) was evident from the three proton singlet at 2.02 ppm (H 16). Both H 16 and H 13 showed HMBC correlations to the C15 carbonyl group confirming the acetyl side chain. Figure 3.4 summarizes the HMBC correlations relevant in establishing the structure of 35 and assigning the quaternary carbon atoms. Structures of Known Pyridoacridines The additional pyridoacridines 36, 37, 38, and 23 showed spectroscopic properties matching those reported in the literature.90,121-123 Pyridoacridine 39, isolated from a Diplosoma sp. ascidian, was previously reported by Charyulu et. al. 124 o Figure 3.4. HMBC correlations used in assigning the quaternary carbon atoms in cystodytin J. 40 DNA Intercalation Studies An ethidium bromide displacement assay was used to evaluate the relative ability of the pyridoacridines to interact with DNA.134 Figure 3.5 shows a series of fluorescence emission spectra for ethidium bromide measured in the presence of calf thymus (CT) DNA and various concentrations of 39. It is apparent that upon increasing the concentration of 39, there was a decrease in the fluorescence of ethidium bronride due to its displacement from DNA. For example, 69 JiM 39 was sufficient to displace nearly all the ethidium bromide (2.5 JiM) from DNA, bringing its fluorescence to a level approaching that of the non-intercalated molecule. Figure 3.6 and Table 3.4 show the relative ability of the pyridoacridines to intercalate into DNA. The ethidium bromide displacement constants (K, Table 3.4) are measures of intercalation. Compounds 35 and 39 were the most efficient intercalators and were also the most cytotoxic of the series as shown in Table 3.4. Equally important is that the less active compounds did not intercalate well into DNA. For example, 34 was the poorest intercalator and the least cytotoxic pyridoacridine, followed by 38 which was the second to last in terms of cytotoxicity and intercalative ability. 14000 12000 Q) 10000 (,) c:: Q) 8000 (,) (f) Q~) 6000 0 ::J u.. 4000 2000 0 550 580 610 640 670 700 Wavelength (nm) Figure 3.5. Fluorescence spectra of ethidium bromide in PBS solution (2.5 JiM) with 50 Jig CT DNA and 0.0 (a), 3.4 (b), 6.8 (c), 13.7 (d), 27.4 (e), and 68.5 (f) JlM diplamine (39). Excitation at 530 nm. (]) () c (]) () (IJ ~ o 6 ;:) u. ~ 4 -8---------0 204-~~~~~~~~~~~ o 10 20 30 40 50 60 Concentration (Jlg/ml) Figure 3.6. Normalized fluorescence for ethidium bromide (1.27 JlM) in PBS solution with calf thynlus DNA and increasing concentrations of pyridoacridines 33 (---Ll---), 34 (-.&-) , 35(-0-), 37 (---.---), 38 (---0---), 23 (-X-), and 39 (---e---). Excitation at 530 nm, emission at 600 nm. The fluorescence in the presence of DNA and absence of compound represents 100%. Table 3.4 Cytotoxicity, Differential Toxicity, Topoisomerase Inhibition and Intercalation of the Pyridoacridines Alkaloids and Control Conlpounds. RCT xrs-6 Inhibition Inhibition Topoll Inhibition Compound IC50, JlM IC50, JlM DC ratioa IC90,b flM 33 8.3 80.0 1 115 34 16.3 8.1 1 138 35 1.6 135.6 1 8.4 37 7.8 88.9 2 127 38 13.8 14.9 1 118 2 3 5.3 NJ)d NJ)d NJ)d 39 <1.4 71.2 1 9.2 Etoposide 2.5 .14 7 68 m-AMSA 6.3 0.24 4 33 Mitoxantrone NJ)d .001 9 1.1 a Differential Cytotoxicity (DC) ratio = BR 1 IC50/ xrs-6 IC50. 41 Intercalation K,c JlM >100 >100 54 62 >100 >100 21 N1)d N1)d N1)d b Concentration at which 90% of monomer-length kDNA production is apparently inhibited. C Concentration of compound required to reduce ethidium bromide fluorescence to 50% of control (see results and discussion). d ND = Not determined. 42 Eilatin (23) is fluorescent and produces an emission peak at 514 nm when irradiated at 520 nm. Figure 3.7 shows the increase in the fluorescence intensity of 23 at the emission maximum upon addition of CT DNA. For example, addition of 50 f..lg CT DNA caused a 484% increase in the fluorescence of a 2.8 f..lM solution of 23. The t1uorescence of 23 continued to increase as more DNA was added, similar to the intercalator ethidium bromide. 120 The increase in fluorescence intensity upon the addition of DNA provided additional evidence for intercalative binding. Biology RCT Cytotoxicity The cytotoxicity assay provided a relative measure of the pyridoacridines toxic potencies against RCT cells in vitro. The results in Table 3.4 show that all compounds are cytotoxic to varying degrees. Compounds 35 and 39 are the most potent; both inhibiting RCT replication with IC50 values of less than 2 f..lM. Q) 0 c Q) 0 en Q) ~ 0 ::::l u. 120000 100000 80000 60000 40000 20000 0~~0ir~~~~~~~i*TI0;~r! 490 500 510 520 530 540 550 Wavelength (nm) Figure 3.7. Fluorescence spectra of eilatin in PBS solution (2.8 f..lM) with 50 (a), 250 (b), and 450 (c) f..lg calf thymus DNA. Excitation at 520 nm. 43 Biochemical Induction Eilatin (23) was active in the BIA at greater than 2 ~g/disk, suggesting that it was capable of inducing an SOS response in E. coli cells. 108 The renlaining compounds were inactive in this assay. Conlpound 23 is aI, lO-phenanthroline derivative and has two sets of nitrogen atoms capable of chelating Ni(II) ions.9o 1,lO-Phenanthrolines with the nitrogen atoms correctly situated for bidentate coordination are excellent ligands for chelating metal ions. 135 1,lO-Phenanthroline metal complexes have been reported to cleave ds-DNA in an oxygen dependent reaction yielding products that are inhibitors of DNA polymerase 1.136 It has been shown that metal complexes of 1,1 O-phenanthroline bind to DNA by intercalation. 137 It is therefore not surprising that 23 showed activity in the BIA since the compound also intercalates into DNA (Figures 3.6 and 3.7). However, these data suggest that intercalation alone is not sufficient for the induction of SOS response since other pyridoacridines which are inactive in the BIA are better DNA intercalators than 23 (Figure 3.6). Differential Cytotoxicity Enhanced toxicity towards the DNA ds-break repair-deficient CHO xrs-6 cell line versus the repair-competent BRlline indicates "cleavable complex" mediated cytotoxicity; BRl/xrs-6 ICSO ratios greater than 3 are considered significant in this assay.80,81.83 None of the pyridoacridines show significant BRlIxrs-6 differential (Table 3.4), suggesting that no cleavable complex formation has occurred and that the compounds do not cause ds­breaks in DNA.101 ,102 Mechanistic Studies Decatenation Inhibition The relative ability of the pyridoacridines to inhibit topoll was measured in vitro using the decatenation inhibition assay. Figure 3.8 shows the abilities of etoposide and 39 to effect dose dependent inhibition of topoll-catalyzed kDNA decatenation in vitro. In this ABCDEFGHIJKLMNOPQRST LN ... MC ... Figure 3.8. Agarose gel showing inhibitory effects of etoposide and diplarnine on topoH catalyzed decatenation of kDNA. Lane A, linear monomer-length 2.5 kb DNA marker; Lane B, control reaction (kDNA + topoH); Lanes C-K, 8.5, 28.3,42.5,56.6, 84.9,283.2,424.7, 566.3, and 849.5 flM etoposide, respectively; Lanes L-T, 1.4, 4.6, 6.8, 9.1, 13.8, 45.8,68.5,91.3, and 137.0 flM 39, respectively. LN, Linear DNA; MC, Monomer Circle DNA. 44 assay, 39 (IC9o = 9.2 flM)138 is a more potent inhibitor of topoH catalytic activity than the classic cleavable complex stabilizing etoposide (lC90 = 68 flM). A relative ranking of the pyridoacridines' ability to inhibit topoH catalyzed decatenation is provided in Table 3.4. The table gives the concentrations at which topon inhibition was apparent (approximately 90% inhibition). Although all compounds are capable of inhibiting decatenation, 35 and 39 are the most potent (IC90 = 8.4 and 9.2 flM, respectively) inhibitors whereas 34 and 38 are among the least potent (IC90 = 138 and 118 flM, respectively). This is consistent with their relative cytotoxicity and intercalative abilities. The concentrations of compound required for enzyme inhibition are greater than those needed to inhibit HCT cell growth. This may be due to the large amount of DNA required for visualization and quantification, and consequently the large amount of enzyme used for DNA cleavage in this assay. Cleavable Complex Formation The involvement of cleavable complex formation in cell death was not supported by the results from the BRlIxrs-6 assay. However, the K+-SDS assay was carried out to definitively determine if inhibition of topon-mediated kDNA decatenation was associated 45 with cleavable complex formation. Compound 39 produced no detectable protein-DNA complex above 1.4 flM, the concentration at which it caused 50% inhibition of HCT cells. Lower concentrations of 39 were used to exclude the possibility of self inhibition 116 due to disruption of the enzyme-substrate interaction; however, cleavable complex formation was not demonstrated at any level. Although it is not clear exactly how these compounds kill cells, protein-DNA covalent links or lesions resulting from disintegration of "cleavable complexes" are evidently not responsible for cytotoxicity. This and other results suggest that DNA, not the protein-DNA complex, is the target of the pyridoacridines. Pyridoacridines Effect on DNA, RNA, and Protein Synthesis Assays to determine the effects of compounds 35 and 39 on cellular macromolecule synthesis were performed using 90-95% lethal drug concentrations (IC90). These assays were performed in Prof. Barrows laboratory. Dramatic effects on RNA and DNA synthesis were observed. RNA synthesis fell to approximately 50% of control levels in the first hour and continued to fall to approximately 10% of control levels by 9 h, while DNA synthesis fell from 100% of control levels in the first hour to below 50% by the third hour and to approximately 10% by h 6 and 9 of drug treatment. No effect was observed on protein synthesis during the 9 h exposure period. These effects are similar to those observed for other DNA or topoisomerase-targeting drugs (e.g., actinomycin D, mitoxantrone and camptothecin) and are consistent with effects expected of topoisomerase inhibition. Summary Based on the observed correlations between HCT cytotoxicity, topoll inhibition, and DNA intercalation, it is hypothesized that the pyridoacridines bring about cell death by inhibiting DNA interactive proteins (e.g., topoll) following intercalation. The studies indicate that the pyridoacridines inhibit proliferation ofHCT cells by interfering with DNA synthesis. This interference is likely due to disruption of topoll enzyme function. It is hypothesized that, by intercalating into DNA, the pyridoacridines disrupt the interaction between topol! and its DNA substrate; consequently, the enzyme cannot carry out its normal functions during replication. The strong correlations between DNA intercalation, topoll inhibition and cytotoxicity, in addition to the ability of these compounds to disrupt DNA synthesis, strongly support this hypothesis. 46 The hypothesis suggests one possible cytotoxicity mechanism for the pyridoacridine alkaloids and provides a reasonable explanation for our observations and those of others,85 that these compounds inhibit the enzyme topoll. These findings are consistent with the report that the pyridoacridine alkaloid dercitin (17), does not significantly stabilize cleavable complexes.12 The fact that these compounds showed neither enhanced toxicity towards DNA ds-break repair-deficient eRO cell lines nor produced cleavable complexes, suggests that they inhibit topoll catalytic activity not by producing a cleavable complex, but by binding to DNA itself. Since the pyridoacridines intercalate into DNA with high affinity and may change the topology of the molecule, it is likely that they inhibit other DNA binding enzymes necessary for replication. Therefore, other enzy:ues such as polymerases or topoI may be unable to bind or function properly due to the presence of intercalator molecules in the DNA. The relative cytotoxic, intercalative, and topoll inhibitory activities of the pyridoacridines give some indication of the effects of structural variation within this class. Inspection of the pyridoacridine structures reveals that 35 and 39 have only four rings and are iminoquinones. They are the best intercalators, topoll inhibitors and the most potent cytotoxins of the series. The diminished potencies of the other pyridoacridines may result in part from steric effects of the additional ring(s), although the electronic effect of the iminoquinone may also be important. Since intercalation appears to be crucial for cytotoxicity, it is understandable why 34 and 38 are among the least active compounds in all the assays. Although the topoll-DNA complex is considered the primary target of a number of 47 DNA intercalators, 112 the results of show that the pyridoacridines do not act by stabilizing cleavable complexes; thus, intercalator induced cytotoxicity involving topoll inhibition does not always involve cleavable complex formation. Therefore, in evaluating cytotoxic compounds that act as topoll poisons, neither inhibition of enzymatic activity, cytotoxic potency, or intercalative ability alone can be taken as suitable criteria for predicting mechanism of action or chemotherapeutic potential. In view of its activity in the BIA, 23 is one of the more interesting pyridoacridines. Compound 23 intercalates efficiently into DNA as evidenced by its fluorescent properties in the absence and presence of DNA and by its ability to displace ethidium bromide from DNA. (See Figures 3.6 and 3.7.) Compound 23 was also reported to chelate a metal ion;90 thus, like other phenanthroline compounds, 23 may cleave'DNA by producing DNA-destructive hydroxyl radicals following intercalation. 137,139,140 Other pyridoacridines may be less capable of generating hydroxyl radicals due in part to their diminished capacities to chelate metal ions. It is likely that the pyridoacridines as a class do not act by a single mechanism in causing cell death. By inhibiting topoII, the pyridoacridines can impair DNA synthesis, gene expression, chromosome segregation and ultimately cell proliferation. Evidence that topoll inhibition plays a part in pyridoacridine cytotoxicity has been provided. Despite the fact that many clinically useful anticancer drugs are intercalators, the mechanisms by which many of these compounds act are not well understood. It is known, for example, that not all intercalators give rise to topoII-mediated DNA breaks, despite binding to the same macromolecule (i.e., DNA). The pyridoacridines, which have a reasonably well-defined mode of action, may be used in conjunction with other compounds, as tools for probing the mechanism of ceIl death in aberrant eukaryotic cells. With this enhanced understanding, we may be able to design better anticancer compounds. 48 Review of Pyridoacridines The pyridoacridines are among the growing number of DNA intercalating agents that contain a planar electron rich ring system. Some of these compounds, like dercitin (17), are tethered to a basic side-chain which may assist with DNA interaction via the phosphate backbone. Dercitin is the only pyridoacridine thus far reported to show in vivo activity (TIC of 170% at 5 mg/kg against P388 leukemia)?3 The structure of dercitin has been revised. 125 The pyridoacridines have been extensively reviewed in a recent publication. 141 CHAPTER 4 THE CHEMISTRY OF THE ASCIDIAN LISSOCLINUM PATELLA L. patella from Singapore The first part of this chapter describes the isolation and characterization of a new cyclic peptide, patellamide E (40), which was isolated from Lissoclinum patella (Order: Aplousobranchia, Family: Didemnidae) collected at Pulau Salu, Singapore. Patellamide E is the newest member of a class of cyclic octapeptides possessing oxazoline and thiazole amino acids. 142 Isolation of Patellamide E and Other Peptides Purification of the CHCl3 extract of L. patella using silica gel flash chromatography followed by silica gel HPLC led to patellamide E (40), patellamides A (41) and B (42),143,144 and ulithiacyclamide (43).145 Structure Elucidation of Patellamide E A molecular formula of C39H50N S06S2 for patellamide E (40) was provided by positive ion HR F AB mass spectral analysis. IR bands at 3371, 3326, 1666 and 1537 cm- 1 were indicative of amide NH and carbonyl stretches for peptides. The 1 Hand l3C NMR spectra (Appendix B) of 40 revealed striking similarities to the other patellamides (e.g., patellamide B (42)). A 14 mass unit difference between them suggested that 40 was a homolog of 42. The l3C NMR spectrum of 40 contained 37 resonances, including two corresponding to the degenerate phenyl carbons C17 & CI7' (129.17 ppm) and C18 & C18' (128.55 ppm). This was in agreement with the molecular formula provided by mass 36 37 '-35/ :4 0 ~ O¥~~~Sh32 8 ~ N (1) N 7 6 ~ 0 27 28 11 (3) 24 29 0=:r 0 25 26 12rs~~~o 18~15 0 :23 19~7 18 40 42 spectrometry. Table 4.1 contains the NMR assignments for 40. 41 43 PS-DQF-COSy146 data established a phenylalanine, an isoleucine, two methyl-oxazoline and two valine residues as constituents of 40. Quaternary carbon atoms resonating at 168.09 (C5) and 168.21 (C24) supported the presence of two methyl-oxazoline rings. Broad proton singlet resonances at 7.46 and 7.52 ppm (H12 and H32, 50 51 Table 4.1 NMRa Assignments for Patellamide E in CDCl3 Atom no. () l3C (mult.)b () IH (mult., 1 (Hz)) HMBCc correlations 1 172.97 (s) 2 73.83 (d) 4.29 (d, 4.0) C1, C5 3 82.21 (d) 4.88 (m) C1, C5 4 21.52 (q) 1.40 (d, 6.5) 5 168.09 (s) 6 54.67 (d) 4.56 (dd, 8.0, 8.0) C5, C10 7 28.74 (d) 2.17 (m) C5 8 18.98 (q) 0.87 (d, 6.5) 9 18.99 (q) 0.92 (d, 6.5) 10 160.96 (s) 11 147.77 12 123.81 (d) 7.46 (bs) C11, C13 13 170.20 14 52.11 (d) 5.43 (ddd, 9.5, 9.5, 6.5) C11,d C13, C16, C20 15 41.24 (t) 3.37 (dd, 14.0, 9.5) C13, C16, C17 3.20 (dd, 14.0, 6.5) C13, C16, C17 16 136.18 (s) 17 129.17 (d) 7.25 (m) 18 128.55 (d) 7.25 (m) 19 126.97 (d) 7.19 (m) C16, C17, C18 20 172.57 (s) 21 73.54 (d) 4.16 (d, 3.5) C20,C24 22 82.11 (d) 4.86 (m) C24 23 21.06 (q) 1.35 (d, 6.5) 24 168.21 (s) 25 53.00 (d) 4.66 (dd, 9.0, 8.0) C24, C30 26 34.30 (d) 2.08 (m) 27 24.90 (t) 1.46 (m) 1.26 (m) 28 9.30 (q) 0.76 (t, 7.5) 29 14.99 (q) 0.88 (d, 6.9)e 30 161.37 (s) 31 148.39 (s) 32 123.03 (d) 7.52 (bs) C31, C33 33 170.51 (s) 34 55.68 (d) 5.14 (dd, 10.5, 4.5) C1, C31, C33 35 32.23 (d) 2.26 (m) C33 36 19.95 (q) 1.06 (d, 7.0) 37 17.06 (q) 1.07 (d, 7.0) N1 7.20 (d, 10.5)e N2 7.69 (d, 8.0)e N3 7.65 (d, 9.5)e C13, C14, C20 N4 7.64 {d% 9.02e C30 a Proton and carbon data were acquired at 500 and 125 MHz, respectively. b From a DEPT experiment. C The HMBC experiment was optimized to observe nlCH couplings of 8.5 Hz. d Weak correlation. e Obtained from a PS-DQF-COSY experiment because of severe overlap in the IH spectrum. respectively) were suggestive of two thiazole rings, which were further supported by carbon resonances at 147.77 (C11), 123.81 (C12), and 170.20 (C13); 148.39 (C31), 123.03 (C32), and 170.51 (C33). The DEPT147 experiment was used to establish the multiplicities of the carbon resonances while the HMQC148 experiment permitted assignment of their attached protons. The structural units in Figure 4.1 accounted for all the mass of 40 except that required for four carbonyl groups. The partial structures in Figure 4.1 and the carbonyl groups accounted for all but one degree of unsaturation (the macrocyclic ring) required by the molecular formula. 52 Assignment of the quaternary carbon resonances using the HMBC131 experiment resulted in the larger structural units A and B shown in Figure 4.2. Unfortunately, neither HMBC experiments, optimized for n JCH of 2 to 15 Hz, nor similarly optimized selective INEPT149 experiments were able to establish the C10-C11 or C30-C31 bonds. This apparently leaves two possible ways of connecting structural units A and B. However, the upfield chemical shifts of the C 10 and C30 amide carbonyl resonances supported conjugation to the thiazole rings and allowed partial structures A and B to be connected in only one way to yield the structure shown for 40. Foster et al. have recently shown, using proton Tl measurements, that the aromatic thiazole protons in the bistratamides (e.g., bistratamide C (44)) had longitudinal relaxation times in excess of 5 s and that an HMBC experiment optimized with a sufficiently long recycle time could detect the requisite three­bond correlations. 150 44 1.40 4.8~ 82.21 o 4.29 )=N 73.83 2.26 1.06 32.23 1.07 19.95~7.06 ...... N 5.14 H 55.68 7.20 "-.JSj/46 \~-\3.81 1.35 7.25 4.8~ 129.17 82.11 o 4.16 3.37, 3.20 )=N73.54 41.24 ....... N 5.43 H 52.11 7.65 2.17 0.87 28.74 0.92 1:8X 8 . 99 0.88 14.99 N 4.56 H54.67 7.69 "-.J S'j) ~ .52 \~--\3.03 0.76 9.30 1.46, 1.26 24.90 Figure 4.1. Proton and COSY derived amino acid spin networks for patellamide E. 53 : 0 Figure 4.2. Partial structures and HMBC correlations for patellamide E. Absolute Stereochemistry of Patellamide E The absolute stereochemistry of 40 was established by comparing the amino acids obtained from acid hydrolysis with standard amino acids, both suitably derivatized for HPLC analysis. lSI The presence of L-threonine, L-valine, and L-isoleucine in the hydrolysate of 40 was established by this procedure. The absolute configurations of the thiazole amino acids were determined by a procedure greatly simplified from that previously published. lS2 This procedure employed ozone in the destruction of the aromaticity of the thiazole in order to facilitate hydrolysis and prevent racemization. This procedure involves bubbling ozone through a solution of peptide, followed by acid hydrolysis, derivatization, and HPLC analysis. In addition to the previously identified 54 amino acids, D-phenylalanine and D-valine were found in the hydrolysate of ozonized 40, establishing the presence of (D-phenylalanine)-thiazole and (D-valine)-thiazole. 55 Biological Activity of Patellamide E Patellamide E was weakly cytotoxic (ICso 125 Jlg/mL) against HCT cells in vitro. L. oatella from the Philippines Continuing investigations of the chemistry of L. patella, have led to the isolation of tawicyclamides A and B (45-46), two new cyclic peptides from a Philippine collection of the ascidian. This section describes the structure determination of these peptides by a combination of NMR spectroscopy, oxidation studies, and tandem mass spectrometry (MSIMS). X-ray crystallography confirmed the structure of tawicyclamide B (46) and also furnished a three-dimensional conformation. This new family lacks the characteristic oxazoline rings normally present in lissoclinum peptides, but possesses a thiazoline ring and a proline (existing in the cis conformation) that facilitates an unusual three-dimensional conformation. A conformational reorganization occurring upon oxidation of the thiazoline ring to a thiazole prompted an investigation of the conformations of tawicyclamide B (46) and its oxidized analog, dehydrotawicyclamide B (47) by molecular modeling. These studies established the solution conformations of both 46 and 47, and by analogy, also established the conformations of tawicyclamide A (45) and dehydrotawicyclamide A (48). The tawicyclamides assume a conformation in which the valine-proline peptide bonds are cis while the dehydrotawicyclamides assume an all-trans amide bond conformation. Isolation of Tawicyclamides A and B The MeOH extract of L. patella was concentrated and partitioned between a series of solvents of increasing polarity. Repeated silica gel flash chromatography of the CHCl3 soluble fraction, followed by reversed phase HPLC, yielded the new peptides tawicyclamides A (45) and B (46), and the known peptides patellamides A (41), and B 143,144 (42), and ulithiacyclamide145 (43). 56 0 25 Zl 25 o - 24 S~ N~{:!J S21 -4:N H N 2) 31 (3) R ~ 19 0 NH (4) (2) H~5W18 o 1 ~_ \ 3 4 5 N 9 12 S 11 78 0 6 33 35 45 R= ~U33 ~ '3l 47 R= 33 35 46 R=~I. 48 Structure Elucidation of Tawicyclamide A HR FAB mass spectral analysis showed a protonated molecular ion at m/z 807.3135 for tawicyclamide A (45). In agreement with the molecular formula C39HSIN80SS3 (~ 1.0 mmu), the 13C NMR spectrum of 45 contained 37 resonances, including signals for two degenerate phenyl carbons at 130.18 and 129.08 ppm. A DEPT147 experiment established the multiplicities of the carbon resonances while an HMQC148 experiment permitted assignment of the attached protons. Characteristic peptide resonances in the I H NMR spectrum of 45 (Appendix B) included doublets at 7.41, 7 .. 59,8.00, and 8.34 ppm, attributable to amide NH protons, and doublets of doublets between 5.89 and 4.62 ppm corresponding to peptide a protons. Further evidence establishing 45 as a peptide was provided by major stretches in the IR spectrum at 3362 cm-l (indicative of secondary amide NH stretching vibrations), amide I bands at 1666 and 1641 cm-l and amide II bands at 1536 and 1514 cm-l . The absence ofIR bands corresponding to a carboxylate or 57 ammonium ion suggested that 45 was cyclic or had end terminal modifications that rendered it nonpolar. The former proved correct when the partial- structures of 45, as discussed below, accounted for all but one degree of unsaturation required by the molecular formula. A singlet resonance at 7.02 ppm (H3) and a fine doublet at 7.42 ppm (H21; J = 0.8 Hz) in the proton spectrum of 45 implied the presence of two thiazole rings. A PS-DQF­COSyl46 experiment established the presence of phenylalanine, isoleucine, proline and two valine residues. An additional spin network consisting of an a. proton (5.14 ppm; dd, J = 9.2, 1.3 Hz; (5 13C = 77.94) coupled to geminal diastereotopic ~ protons (4.03 ppm, dd, J = 11.3, 1.3 Hz and 3.05 ppm, dd, J = 11.3, 9.2 Hz; (5 13C = 38.47) was attributed to a thiazoline ring, which was confirmed by nickel peroxide oxidation of 45 to form dehydrotawicyclamide A (48, Figure 4.3). Figure 4.4 contains the partial structures established from proton and COSY data. Assignments of the quaternary carbon atoms of 45 were accomplished using the HMBC experiment (Figure 4.5). As evident from Figure 4.5, potential correlations that could establish the C 19-C20 and the proline nitrogen-C 14 bonds are not observed, thus, substructure C remains unconnected to the remainder of the molecule. A very strong ROESY correlation between HID and H15 of the proline and valine, respectively, suggests that these residues are adjacent. Since tawicyclamide A is cyclic, there is only one way of combining the partial structures in Figure 4.5. Carbon atoms 14 and 19 are assigned based on their chemical shifts. Of the remaining unassigned carbonyl resonances, the upfield one (160.32 ppm) can be assigned to C19, which is adjacent to and in conjugation with the aromatic thiazole ring. Ni02/CsHS 23h 25°C .. Figure 4.3. Nickel peroxide oxidation of tawicyclamide A. 3.05,4.03 1.34, 1.60 7.36 38.47 22.89 Y 1.77, 2'1~.27' 3.84 3.66 32.24 47.07 38.61 - N 77.94 4.62 63.32 N'. "- N 5.89 H 56.13 8.34 1.89 2.20 0.51 36.59 0.91 0.83 33.60 0.76 1~8X9.84 1~OX9.60 1.29 15.46 N 5.37 N 4.97 H 57.08 H56.22 7.41 8.00 ~S/;/.02 \~-\4.02 ~S/;lJ.42 \~-\4.17 7.06 129.08 0.96 12.66 1.35, 1.59 28.14 Figure 4.4. Proton and COSY derived amino acid spin networks for tawicyclamide A. 58 Figure 4.5. HMBC correlations for tawicyclamide A. Complete NMR assignment of 45 based on COSY, HMQC, COLOC,153 and HMBC131 data are provided in Table 4.2. Despite abundant NMR data, the sequence of 59 the amino acid residues in 45 remains not proven by direct spectroscopic evidence. Table 4.3 contains the NMR assignments for the oxidized analog dehydrotawicyclamide A (48). Tandem Mass Spectrometry of Tawicyclamide A With its constituent amino acids and tentative sequence in hand, tandem mass spectrometry (MSIMS) was used as an alternative sequencing method for definitively proving the structure of 45. Fast atom bombardment (FAB) is a soft ionization technique that is particularly useful for generating protonated molecular ions of non-volatile compounds. In MSIMS the first mass spectrometer is used for mass separation and selection. Gas phase degradation resulting from collisional activation of the mass selected ion and scanning the second mass spectrometer result in a collision induced dissociation (CID) spectrum. The combination ofFAB ionization and MSIMS can provide structural and sequence information that can be obtained by examining the CID spectra of structurally significant ions. 154,155 60 Table 4.2 NMRa Assignments for Tawicyc1amide A in C6D6 Atom no. o l3C (mult.)b o IH (mult., J (Hz)) HMBCc correlations 1 161.57 (s) 2 148.69 (s) 3 124.02 (d) 7.02 (s) C1, C2, C4 4 170.44 (s) 5 57.08 (d) 5.37 (dd, 6.7, 3.3) C9 6 36.59 (d) 1.89 (m) C4 7 16.88 (q) 0.51 (d, 6.8) 8 19.84 (q) 0.91 (d, 6.8) 9 171.19 (s) 10 63.32 (d) 4.62 (dd, 8.5, 2.1) C9 11 32.24 (t) 1.77 (m) C9 2.10 (m) C9 12 22.89 (t) 1.34 (m) 1.60 (m) 13 47.07 (t) 3.27 (m) 3.84 (m) 14 174.24 (s) 15 56.22 (d) 4.97 (dd, 10.0, 8.5) C14, C19 16 33.60 (d) 2.20 (m) C14 17 19.10 (q) 0.83 (d, 6.8) 18 19.60 (q) 0.76 (d, 6.8) 19 160.32 (s) 20 149.62 (s) 21 124.17 (d) 7.42 (d,0.8) C20, C22 22 172.66 (s) 23 54.55 (d) 5.79 (dd, 10.2, 2.8) C20, C21, C22, C28 24 40.71 (d) 1.87 (m) C22 25 28.14 (t) 1.35 (m) 1.59 (m) 26 12.66 (q) 0.96 (t, 7.3) 27 15.46 (~ 1.29 (d, 6.8) 28 172.50 (s 29 77.94 (d) 5.14 (dd, 9.2, 1.3) C28, C31 30 38.47 (t) 3.05 (dd, 11.3, 9.2) C28 4.03 (dd, 11.3, 1.3) C28, C31 31 177.42 (s) 32 56.13 (d) 5.89 (dd, 8.0, 8.0) C1, C31 33 38.61 (t) 3.66 (m) C31, C34 34 138.18 (s) 35 130.18 (d) 7.36 (dd, 8.2, 1.3) C35, C37 36 129.08 (d) 7.06 (ddd, 8.2, 7.5, 1.3) C34, C36 37 127.46 (d) 6.97 (dt, 7.5, 1.3) C35 N1 7.41 (d, 6.7) C4,C9 N2 8.00 (d, 10.0) C19 N3 7.59 (d, 10.2) C28 N4 8.34 ~dz 8.02 C1 a Proton and carbon data were acquired at 500 and 125 MHz, respectively. b From a DEPT experiment. C The HMBC experiment was optimized to observe nJCH couplings of 8.5 Hz. 61 Table 4.3 NMRa Assignments for Dehydrotawicyclamide A in C6D6 Atom no. 8 13C 8 1H (mult., 1 (Hz» HMBCb correlations 1 160.46 2 149.95 3 123.30 7.43 (s) C2,C4 4 167.90 5 56.47 5.14 (dd, 7.9, 6.5) . C4, C9 6 34.83 2.19 (m) C4 7 18.59 0.73 (d, 6.8) 8 18.74 0.81 (d, 6.8) 9 169.94 10 60.78 4.42 (d, 7.9) C9, C14 11 25.47 0.74 (m) 2.51 (dd, 12.3, 6.1) C9 12 24.81 1.14 (m) 1.57 (m) 13 47.69 2.93 (m) 14 173.59 15 55.69 5.00 (dd, 9.0, 6.7) C14, C19 16 33.25 2.25 (m) 17 18.13 1.20 (d, 6.8) 18 19.99 1.15 (d, 6.8) 19 160.79 20 149.64 21 123.46 7.73 (s) C20, C22 22 169.35 23 56.12 5.47 (dd, 7.9, 4.7) C22, C28 24 41.54 1.97 (m) 25 25.48 1.51 (m) 26 11.75 0.62 (t, 7.3) 27 15.10 0.72 (d, 6.8) 28 160.48 29 150.72 30 124.11 7.77 (s) C29, C31 31 170.59 32 53.15 5.64 (ddd, 8.4, 7.4, 7.4) C1, C31 33 41.37 3.35 (dd, 13.5, 7.4) C31, C34 3.44 (dd, 13.5, 8.4) C31, C34 34 137.16 35 129.34 6.99 (m) C35, C37 36 128.69 6.93 (m) C34, C36 37 127.03 6.90 (m) C35 N1 8.55 (d, 7.9) C4,C9 N2 8.70 (d, 9.0) C14, C19 N3 8.86 (d, 7.9) C22, C28 N4 8.26 (d, 7.4) C1, C32, C33 a Proton and carbon data were acquired at 500 and 125 MHz, respectively. b The HMBC experiment was optimized to observe nlCH couplings of 8.5 Hz. 62 The unambiguous structure of tawicyclamide A was established by interpreting the CID spectra of its protonated molecular ion and several fragment ions resulting from unimolecular dissociations. The CID spectrum of the protonated molecular ion of 45 (m! z 807) was dominated by ions resulting from the fragmentation of the linear acylium ion depicted in Figure 4.6. This acylium ion resulted from protonation at the proline nitrogen followed by scission of the protonated N-acyl bond, and its fragmentation was the outcome of successive cleavages at the C-terminal. The more basic nature of the prolyl nitrogen -relative to the other amide nitrogen atoms- directed protonation, yielding a much simplified fragnlentation pattern. The major fragmentation of the (M+H)+ ion of 45 involved loss of C-terminal fragments to produce m! z 779, 708, 529, 512,484, 399, and 297 ions (Figures 4.6, 4.7, and 4.8). Subsequent fragmentation of these ions results in the additional peaks in the CID spectrum. Another dominant set of progeny ions can be seen in the CID spectra of both the m! z 807 and 779 ions (Figures 4.7, 4.8, and Appendix C). These progeny ions result from decarbonylation of the m! z 807 acylium ion to produce the m! z 779 immonium ion, which ultimately loses the N-terminal proline residue to produce the m! z 682 ion. Collision induced fragmentation of the m! z 682 ion accounts for the progeny ions of this second series (Figure 4.9). Appendix C contains the CID spectra and structures of the m! z 779, 611,557,529,484,399,297,285, 251, 223, 188, and 180 ions of 45. Tandem mass spectral results are fully consistent with structure 45. Figure 4.6. Linear acylium ion of tawicyclamide A resulting from scission of the proline N -CO bond. W () Z « 0 z :J CD « w > r«- ---1 w a: 100 779 807 a x x ~ I\) I\) 0 0 75 9 50 c 708 399 e 682 d 512 484 f 25 529 692 400 500 600 m/z 700 800 Figure 4.7. High mass region (mlz 825-370) from the CID spectrum of the (M+H)+ ion of tawicyclanride A, mlz 807. Letters correspond to the fragmentation depicted in Figure 4.6. 63 w 100 «oz Cl z 75 :::J co « w > 50 I- ::s UJ a: 25 124 100 251 180 231 200 b 297 285 300 c e 399 512 d f 336 484 529 400 500 600 m/z Figure 4.8. Low nlass region (mJz 625-80) from the CID spectrum of the (M+H)+ ion oftawicyclamide A, mJz 807. Letters correspond to the fragmentation depicted in Figure 4.6. 64 w 100 (z«) o z 75 ::> III « w > t- 50 « ....J w a: 25 I lL_ ____e_____________I____ J i ~, 01 II II "\X N II+~ I N II ~ I liN H I II II I H II I H I II S hi I II I ) ~I----------------J / \ (-H) ,/ (-H)\~ f 9 i (+H) d i IL _ ________________________________________________________________ J I 285 b (+3H) 126 100 h c 180 231 k a 197 9 166 251 200 300 j 356 m/z f . e 415 486 400 500 d 557 i 611 600 Figure 4.9. CID spectrum and structure of the miz 682 ion of tawicyc1amide A. 65 66 Similar MSIMS analysis of the oxidized peptide, 48, supported its proposed structure. The CID spectrum of mlz 805, its protonated nl0lecular ion, showed a fragmentation pattern virtually identical to that of 45 but with many fragments differing by two mass units. Absolute Stereochenlistry of Tawicyclamide A The absolute stereochenlistry of 45 was determined by comparing the 1-fluoro-2,4- dinitrophenyl-5-L-alanineamide (FDAA) derivatized amino acids from the acid hydrolysate of the peptide with similarly derivatized standard amino acids by HPLC according to the method of Marfey.151 This procedure established L-proline, L-valine and L-phenylalanine as constituents of 45. The absolute configurations of the thiazole amino acids were determined142,152 to be (D-isoleucine)-thiazole and (L-valine)-thiazole. Structure Elucidation of Tawicyclamide B HR FAB MS provided an (M+H)+ at mlz 773.3345 for tawicyclamide B (46), consistent with the molecular formula C36HS3NgOSS3 (il4.4 mmu). The l3C spectrum of 46 showed 36 unique resonances in agreement with this formula. Compound 46 showed remarkable spectral similarities to 45; the most significant differences were the absence of phenylalanine resonances and the presence of two new methyl resonances (0.98 and 22.99; and 0.88 and 23.06 ppm in the IH and 13C spectra, respectively (Appendix B)). Tables 4.4 and 4.5 contain the NMR assignments for the parent peptide, 46 and its oxidized analog, 47, respectively. These data, coupled with a 34 mass unit decrease in the molecular weight, allowed the conclusion that 46 differed from 45 by replacement of phenylalanine in 45 with leucine to form 46. COSY data supported a leucine spin network with long range coupling to the thiazoline ring in 46 (Figure 4.10). This thiazoline ring was also confirmed by oxidation to a thiazole to give dehydrotawicyclamide B (47). 67 Table 4.4 NMRa Assignments for Tawicyclamide B in C6D6 Atom no. B 13C (mult.)b B IH (mult., J (Hz» HMBCc correlations 1 161.66 (s) 2 149.01 (s) 3 124.07 (d) 7.12 (s) C1, C2, C4 4 170.68 (s) 5 57.34 (d) 5.38 (dd, 6.5, 3.0) C4 6 36.53 (d) 1.94 (m) 7 17.05 (q) 0.54 (d, 6.5) 8 19.84 (q) 0.91 (d, 6.5) 9 171.29 (s) 10 63.23 (d) 4.57 (dd, 8.5, 1.5) C9 11 32.21 (t) 1.72 (m) C9 2.06 (m) 12 22.80 (t) 1.30 (m) 1.58 (m) 13 46.83 (t) 3.22 (m) 3.79 (m) 14 174.06 (s) 15 56.14 (d) 4.95 (dd, 10.0, 8.0) C14, C19 16 33.60 (d) 2.12 (m) C14 17 19.02 (q) 0.80 (d, 6.5) 18 19.57 (q) 0.72 (d, 6.5) 19 160.31 (s) 20 149.69 (s) 21 124.13 (d) 7.44 (s) C19, C20, C22 22 172.73 (s) 23 54.59 (d) 5.83 (dd, 10.5, 2.5) C20, C22 24 40.82 (d) 1.90 (m) C22 25 28.16 (t) 1.39 (m) 1.65 (m) 26 12.67 (q) 1.00 (t, 7.3) 27 15.48 (~ 1.32 (d, 7.0) 28 172.61 (s 29 78.02 (d) 5.28 (dd, 9.3, 1.3) C28, C31 30 38.39 (t) 3.12 (dd, 11.3, 9.3) C28 4.04 (dd, 11.3, 1.3) C28, C29, C31 31 178.31 (s) 32 53.20 (d) 5.73 (dd, 8.0, 8.0) C1, C31 33 41.22 (t) 2.25 (m) C31 34 25.87 (d) 1.83 (m) 35 22.99 (q) 0.98 (d, 6.5) 36 23.06 (q) 0.88 (d, 6.5) N1 7.43 (d, 6.5) C4 N2 7.96 (d, 10.0) C15, C19 N3 7.64 (d, 10.5) C21, C22, C28 N4 8.26 (dz 8.02 C1 z C32z C33 a Proton and carbon data were acquired at 500 and 125 MHz, respectively. b From a DEPT experiment. c The HMBC experiment was optimized to observe nJCH couplings of 10.0 Hz. 68 Table 4.5 NMRa Assignments for Dehydrotawicyclamide Bin C6D6 Atom no. 3 13C 3 lH (mulL, J (Hz» HMBCb correlations 1 160.45 2 150.20 3 123.26 7.49 (s) C2,C4 4 167.96 5 56.45 5.15 (dd, 7.9, 6.6) C4,C9 6 34.94 2.16 (m) C4 7 18.52 0.71 (d, 6.8) 8 18.73 0.79 (d, 6.8) 9 170.05 10 60.74 4.44 (d, 7.9) C9, C14 11 25.49 0.72 (m) 2.51 (dd, 12.2, 6.4) C9 12 24.77 1.18 (m) 1.53 (m) 13 47.58 2.91 (m) 14 173.36 15 55.55 5.03 (dd, 8.8, 6.2) C14, C19 16 33.15 2.25 (m) C14 17 18.35 1.31 (d, 6.8) 18 20.01 1.22 (d, 6.8) 19 160.83 20 149.52 21 123.39 7.71 (s) C20, C22 22 169.16 23 56.20 5.39 (dd, 7.7, 4.6) C22, C28 24 41.40 1.96 (m) 25 25.35 1.52 (m) 26 11.75 0.60 (t, 6.8) 27 15.13 0.70 (d, 6.5) 28 160.51 29 150.97 30 123.68 7.85 (s) C28, C29, C31 31 171.16 32 49.60 5.54 (ddd, 8.4, 7.4, 7.4) C1, C31 33 43.93 1.95 (m) C31 2.04 (m) C31 34 25.28 1.51 (m) 35 22.08 0.74 (d, 6.6) 36 22.35 0.69 (d, 6.6) N1 8.46 (d, 7.9) C4,C9 N2 8.67 (d, 8.8) C14, C19 N3 8.85 (d, 7.7) C22, C28 N4 8.05 (d, 7.4) C1, C32 a Proton and carbon data were acquired at 500 and 125 MHz, respectively. b The HMBC experiment was optimized to observe nJCH couplings of 8.5 Hz. Figure 4.10. Partial structure of tawicyclamide B showing five bond correlation between leucine and thiazoline. Absolute Stereochemistry of Tawicyclamide B 69 HPLC analysis of the FDAA derivatized hydrolysate of 46 revealed the presence of L-proline, L-valine and L-leucine with a significant amount of D-leucine that resulted from partial racemization of the L-leucine from (L-leucine )-thiazoline. The UD-leucine ratio was 1.38. The 16% ee of the L isomer did, however, support its presence in the natural product; and whereas partial racemization of L-1eucine from (L-leucine )-thiazoline occurred upon acid hydrolysis of 46, analysis of the hydrolysate of ozonized 46 revealed only L­leucine. A similar trend was observed for 45. X-Ray Structure of Tawicyclamide B Single crystal X-ray analysis was carried out on tawicyclamide B (46) in the laboratory of Prof. Jon Clardy. Since only the relative stereochemistry could be deternrined in the X-ray experiment, the absolute configuration shown in Figure 4.11 was set by the L-proline, L-valine, and L-leucine residues. The X-ray analysis confirmed the spectroscopically determined structure and also revealed the three-dimensional structure. The thiazo1e rings are essentially parallel and are separated by only 3.7 A, a typical aromatic stacking distance. The resulting conformation has the hydrophobic side chains pointing away from the internal cavity; the isopropyl group of one valine points up, while the other valyl and leucyl side chains point down. This conformation also has a cis-valine-proline peptide bond. A weak intramolecular hydrogen bond from NH4to 03 of 2.15 A (148°) suggests that hydrogen bonding may also playa role in stabilizing the solid state conformation of 46. 70 C36 Figure 4.11. X-ray model of tawicyclamide B. Molecular Modeling Studies on the Tawicyclamides Nickel peroxide oxidation of the tawicyclamides appeared to result in conformational changes in remote parts of the molecules. For example, several pieces of NMR data suggested that, in C6D6 solution, 47 adopted a drastically different conformation than its parent peptide, 46. The dissimilarity between the ROESy129,130,156 and 13C NMR data of these two peptides, especially data associated with the proline residue, indicated a significant change in the chemical environment of the atoms. These changes suggested isomerization of the valine-proline amide bond from cis to trans upon oxidation of the thiazoline to a thiazole. These observations prompted the investigation of the conformations of 46 and 47 by molecular modeling. 71 Tawicyclamide B Molecular Modeling Studies Compound 46 was modeled starting from the X-ray coordinates using a minimization-molecular dynamics-minimization (min-MD-min) procedure. 157-159 Initial energy minimization using the ABNR (Adopted Basis-set Newton-Raphson) technique was run to eliminate unfavorable constraints brought about by crystal packing forces and to locate the lowest energy conformer near the starting geometry. A MD simulation at 300 K was run using this low energy conformer. In order to evaluate the conformational space available to the molecule under experimental conditions, the effect of solvent was approximated by a constant dielectric (E 2.284 for benzene @ 20°C). Minimization of several dynamics datasets resulted in convergence to an average structure practically identical to the starting low energy conformer and the crystal structure. Figure 4.12 shows the optimal conformation for 46, which has a cis-valine-proline amide bond (C15-C14-N5-C10 dihedral = -10.3°), and an intramolecular hydrogen bond from NH4 to 03 of 2.05 A (145°), Figure 4.12. Stereo drawing of the solution structure of tawicyclamide B. 72 The stability of this conformation of 46 was illustrated by the fact that both stacking interaction and hydrogen bonding persisted throughout the min-MD-min procedure. Furthermore, proline allowed for the possibility of cis-trans isomerism, yet only one conformer appeared to be present in C6D6 solution-as evidenced by a single set of NMR resonances. The activation energy barrier to rotation about the valine-proline amide bond was determined to be 14.9 kcal/mol, a value comparable to those reported for X-valine (X = any amino acid) rotation barriers. 160 The solution structure of 46, obtained from modeling studies, was in excellent agreement with NMR data. For example, a distance of 2.2 A between HID and H15 in this model was consistent with the very strong ROESY crosspeak observed between these two protons (Figure 4.13). This key nOe also implied a cis orientation about the valine-proline amide bond. Figure 4.14 provides additional evidence for the proposed solution conformation of 46. The figure shows that a strong nOe exists between one thiazole aromatic proton (H3) and the y-methyl group of isoleucine (H27) which, in the model, are separated by 2.5 A at closest approach. The relative downfield position of H27 is due primarily to anisotropic deshielding by the thiazole ring and is fully consistent with this model. Further evidence supporting the proposed solution conformer of 46 is provided by an nOe between one leucine () methyl group (H35) and the proline () protons (HI3) which approach each other to within 2.4 A. A 10 ps MD simulation at 1000 K, carried out to escape local minimum energy wells, allowed for sampling of a larger conformational space and generated a reasonable number of structures for subsequent minimization. Following minimizations, convergence to an average structure similar to the starting conformer and to the crystal structure was again observed. The side chains were somewhat more flexible, showing larger rms deviations from the average structure, than the main chain atoms. The average energy of these minimized conformers was, however, 3.5 kcallmole higher than the starting low energy F1 , /J ~ H10-H15 If) 00 4.8 4.9 5.0 5.1 5.2 5.3 5.3 5.1 4.9 4.7 4.5 F2 (ppm) Figure 4.13. Region of a 300 ms ROESY spectrum of tawicyclamide B showing the HIO-HIS crosspeak. 73 4.3 Figure 4.14. Region of a 300 ms ROESY spectrum of tawicyclamide B showing the H3-H27 crosspeak. 74 75 conformer as a result of large deviations of the amino acid side chains. Dehydrotawicyclamide B Molecular Modeling Studies The starting model for 47 was taken from the x-ray structure of 46 by removal of two hydrogen atoms and explicit aromatization of the thiazoline ring. Modeling was again carried out using a constant dielectric corresponding to benzene. Energy minimizations were run to locate the lowest conformer near the starting geometry. This structure, not surprisingly, retained a cis-valine-proline amide bond after minimization with the ABNR technique. The molecule was first heated to 1000 K for 1.0 ps, allowed to equilibrate at this temperature for 1.0 ps, and finally simulated at 1000 K for 10 ps. Following minimizations of select structures throughout the trajectory, two populations of conformers were obtained: one was similar to the starting cis structure and the other was dissimilar, having a trans-valine-proline amide bonds. Interestingly, convergence to the trans structures occurred toward the end of the lOps simulation. Although this trans isomer appeared to fit the NMR data, its energy was higher, by 0.1 kcal/mole, than the cis isomer. Since the calculations failed to give results energetically consistent with experimental data, another approach to find a starting conformation was taken. The valine-proline amide bond was constrained to 1800 and subjected to an ABNR-MD-ABNR protocol. The dihedral constraint was removed after the first series of minimizations and the MD simulations were run to let the whole molecule relax. The molecule was heated to 300 K for 0.3 ps, equilibrated for 0.3 ps, and finally simulated at 300K for 3 ps. After minimization, the resulting structure contained the trans valine-proline amide bond and was very similar to that obtained from dynamics simulation starting from the cis dehydrotawicyclamide B conformer. Compared to the cis, this trans conformer of 47 was more stable by 0.9 kcallmole. This result suggests that the final model obtained from the modeling protocol is sensitive to the starting geometry. The lowest energy trans conformer of 47 is shown in Figure 4.15. The model shows that in 47, the all trans peptide backbone forms a rectangle, with the three thiazole Figure 4.15. Stereo drawing of the solution structure of dehydrotawicyclamide B. 76 rings and the proline defining its comers. This shape is similar to that for ascidiacyclamide (49), although somewhat distorted from the "saddle" shape described.161-163 The isoleucine side chain protrudes below the plane of the macrocyclic ring while the leucine and both valine side chains extend above the ring. An interesting feature of this conformer is that all the NH bonds point toward the center of the ring, away from the hydrophobic environment of the solvent. 77 The stacking of the thiazole rings and the stabilizing hydrogen bonding present in the parent peptide, 46 are not present in the dehydro isomer, 47. The structure obtained from modeling studies agrees quite well with NMR data. For example, a strong ROESY crosspeak between both () proline protons (HI3) and the a valine proton (HI5), supports this trans conformer (Figure 4.16). In the modeled structure, these protons are separated by 2.2 A and 2.5 A respectively. The model also shows that HI0 and NHI approach each other close enough (2.8 A) to lead to the strong crosspeak observed in the ROESY spectrum (Figure 4.17). The absence of a crosspeak between HI0 and H15 in the ROESY spectrum of 47 further supports the proposed trans conformer in which these hydrogen atoms are separated by 4.4 A. These two hydrogen atoms were 2.2 A apart in 46 and gave rise to a strong nOe as discussed earlier (see Figure 4.13). Interestingly, upon changing the thiazoline ring to a thiazole, the resulting aromatization causes the adjacent carbonyl to come into resonance with the aromatic ring and become coplanar with the ring. This undoubtedly contributes to the drastic conformational changes observed upon oxidation. Added proof of the proposed conformational changes come from the chemical shift differences between the proline ~ and 1 carbons (~()py), which supported the cis valine-proline amide bond configuration for the tawicyclamides and the trans configuration for the dehydrotawicyclamides. Siemion et al. had shown that for X-Pro (X = any amino acid), there was a linear dependence of the difference in the chemical shifts of the ~ and 1 carbons of proline (~()~1) with dihedral angle 9 = ('V - 60°) according to the equation; ~()~1 = 0.081191 + 2.47 for a cis orientation about the X-Pro amide bond, or ~()~1 = 0.036191 + 0.73 for a trans orientation. 164 The observed ~()~1 value of 9.5 for 46 placed it in the cis­X- Pro series. A 9 (02-C9-CI0-Cl1) angle of -79.7° in the modeled cis conformer, predicted a ~()~1 value of 8.9, whereas a ~()~1 value of 0.7 placed 47 in the trans-X-Pro series. A 9 angle of -8.2° in the trans conformer predicted a ~()~1 value of 1.0. Similar values were obtained for tawicyclamide A (45) and dehydrotawicyclamide A (48). F1 (ppm) 2.1 - 2.2 ., H5-H6 2.3 H16-H15 2.4 2.5 2.6 2.7 2.8 2.9 3.0 H13-H15 3.1 3.2 5.22 5.18 5.14 5.10 5.06 5.02 4.98 4.94 1'2 (ppm) Figure 4.16. Region of a 300 ms ROESY spectrum of dehydrotawicyclamide B showing the H13-H15 crosspeak. 78 011 • 4.5 NH1-H10 4.6 4.7 4.8 4.9 5.0 0 5.1 NH2-H15 e 5.2 NH1-H5 5.3 5.4 CD 5.5 NH3-H23 • 5.6 NH4-H32 5.7 8.8 8.6 8.4 8.2· 8.0 1'2 (ppm) Figure 4.17. Region of a 300 ms ROESY spectrum of dehydrotawicyc1amide B showing the NH 1-H 10 crosspeak. 79 80 The empirical observation that carbon atoms syn to the carbonyl oxygen of amides are shielded relative to those that are anti165 is borne out by the NMR data. The proline a carbon of the trans conformer 47 is shielded relative to the cis conformer 46. Likewise, the proline 0 carbon atom of 46 is shielded relative to 47 as predicted. Biological Activity of the Tawicyclamides The tawicyclamides (45-46) and their dehydro analogs, 47 and 48, are weakly cytotoxic against human colon tumor cells, all with 1CSO values of 31 Ilg/mL. These results are consistent with structure-activity studies that show that the oxazoline ring is essential for the cytotoxicity of this class of compounds. 166 Review of Cyclic Peptides Cyclic Peptides from Ascidians Ascidians have proven to be a rich source of bioactive amino acid-derived secondary metabolites. 18,167 The prolific Didemnidae family has produced several families of peptide metabolites such as the didemnins59 and the lissoclinum peptides.142-144,150,168-175 The lissoclinum peptides from L. patella, characterized by the presence of thiazole and oxazoline amino acids, fall into two general groups-the heptapeptide lissoclinamides and the octapeptide patellamides/ulithiacyclamides. These peptides exhibit in vitro cytotoxicity with the presence of the oxazoline ring proving important to their potency. 166 Ulithiacyclamide is the most potent lissoclinum peptide. A large number of cyclic peptides have been isolated from ascidians.68,142-145,150,168-174,176-181 These will not be explicitly reviewed here since amino acid derived metabolites from ascidians have been extensively reviewed in a recent publication.182 Three-Dimensional Structure of Cyclic Peptides from Ascidians The three-dimensional structures of several ascidian derived cyclic peptides