UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Chemical studies of Pacific Ocean sponges Coleman, John Edward 1998

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_1998-271226.pdf [ 12.61MB ]
Metadata
JSON: 831-1.0061561.json
JSON-LD: 831-1.0061561-ld.json
RDF/XML (Pretty): 831-1.0061561-rdf.xml
RDF/JSON: 831-1.0061561-rdf.json
Turtle: 831-1.0061561-turtle.txt
N-Triples: 831-1.0061561-rdf-ntriples.txt
Original Record: 831-1.0061561-source.json
Full Text
831-1.0061561-fulltext.txt
Citation
831-1.0061561.ris

Full Text

CHEMICAL STUDIES OF PACIFIC O C E A N SPONGES. by John Edward Coleman B.Sc, Bishop's University, Lennoxville, P .Q . , 1992 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF T H E REQUIREMENTS FOR THE D E G R E E OF DOCTOR OF PHILOSOPHY in T H E F A C U L T Y OF G R A D U A T E STUDIES (Department of Chemistry) We accept this thesis as conforming to the required standard The University of British Columbia December 1997 © John Edward Coleman, 1997 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of The University of British Columbia Vancouver, Canada Date <?7- 12.22-DE-6 (2/88) A B S T R A C T A n investigation into the chemistry of two species of marine sponges has led to the isolation of eighteen new and twelve previously known secondary metabolites. The structures of the new compounds were elucidated by a combination of spectroscopic analysis and chemical degradation. A study of the organic extracts of a pink sponge collected of the coast of British Columbia led to the identification of botanones A to C (55-57), botanicol (58), cadlinolide C methyl acetal (59), and cadlinoglycine (60). Botanones A to C (55-57) were the first examples of diterpenoids with the degraded "labdane" skeleton to be isolated from a marine sponge. Cadlinoglycine (59) and cadlinolide C methyl acetal (60) and were both thought to be isolation artifacts. The novel secondary metabolites isolated from the pink sponge were useful in identifying the sponge as a previously undescribed Aplysilla species. The novel cyclic depsipeptides geodiamolides H to N (89-95) and P (97), as well as the known cytotoxins geodiamolides A to G (77-83), were isolated during the bioassay guided fractionation of the cytotoxic extracts of the marine sponge Cymbastela sp. Geodiamolides J to N (91-95) and P (97) represented the first examples in the geodiamolide family in which a serine residue had been incorporated. Extracts of Cymbastela sp. also yielded the novel cytotoxic peptides hemiasterlin A (98), hemiasterlin B (99), criamide A (100), and criamide B (101), in addition to the known peptide hemiasterlin (86). Chemical degradation, followed by Marfey's amino acid analysis of hemiasterlin (86), combined with single crystal X-ray diffraction analysis of hemiasterlin methyl ester (141), established the absolute configuration of 86. Using a combination of chemical degradation and C D analysis, the absolute configurations of hemiasterlin A (98) and hemiasterlin B (99) were also determined. ii The hemiasterlins were shown to be potently cytotoxic in vitro against several cell lines, and hemiasterlin (86) demonstrated promising activity in vivo against murine leukemia P388 in mice. Studies on the mechanism of activity of hemiasterlin A (98) suggested that it exerted its cytotoxic affects by inhibiting spindle microtubule dynamics. Both hemiasterlin (86) and hemiasterlin A (98) were more potent cytotoxins and mitotic blockers than the known microtubule inhibitors paclitaxel (113) and vinblastine (114). The total synthesis of hemiasterlin (86) was accomplished. Using the methodology developed in the total synthesis, several analogs were synthesized in the beginning of an extensive structure activity study. i i i Compound R2 X Compound X 91 geodiamolide J C H 2 O H Me I 89 geodiamolide H Br 92 geodiamolide K C H 2 O H Me Br 90 geodiamolide I CI 93 geodiamolide L C H 2 O H Me CI 94 geodiamolide M Me C H 2 O H I 95 geodiamolide N Me C H 2 O H Br 97 geodiamolide P C H 2 O H H I 100 R = H 101 R = Me J T A B L E O F C O N T E N T S ABSTRACT ii TABLE OF CONTENTS V LIST OF FIGURES vii LIST OF TABLES xviiiii LIST OF ABREVIATIONS XV LIST OF SCHEMES xviii ACKNOWLEDGMENTS xviii 1 GENERAL INTRODUCTION 1 1.1 M A R I N E N A T U R A L PRODUCTS CHEMISTRY 1 1.2 SPONGE BIOLOGY 8 1.3 ENDNOTES (CHAPTER 1: GENERAL INTRODUCTION) 1 0 2 DlTERPENOIDS FROM THE MARINE SPONGE APL YSILLA SP. 12 2.1 INTRODUCTION 1 2 2 . 2 RESULTS 2 0 2.2 .1 Known Diterpenoids 2 4 2 . 2 . 2 Novel Diterpenoids 2 6 2 .2 .3 Discussion 5 3 2 . 2 . 4 Isolation Artifacts 5 6 2 .2 .5 Discussion 7 3 2.3 CONCLUSIONS 7 7 2 . 4 EXPERIMENTAL 7 9 2.4 .1 General 7 9 2 . 4 . 2 Materials 81 2 .5 ENDNOTES (CHAPTER 2 : DITERPENOIDS FROM THE M A R I N E SPONGE APLYSILLA SP.) 8 7 3 NEW GEODIAMOLDDES FROM THE MARINE SPONGE CYMBASTELA SP. 89 3.1 INTRODUCTION 8 9 3.2 RESULTS AND DISCUSSION 9 5 3.2.1 Geodiamolide H (89) 9 7 3 .2 .2 Geodiamolide I (90) 1 0 0 3.2.3 Geodiamolide J (91) 1 0 2 3 .2 .4 Geodiamolide K (92) 1 0 7 3.2 .5 Geodiamolide L (93) 1 1 0 3 .2 .6 Geodiamolide M (94) 1 1 2 3 .2 .7 Geodiamolide N (95) 1 1 7 V 3 .2 .8 Geodiamolide P (97) 1 2 0 3.3 CONCLUSIONS 1 2 4 3.4 EXPERIMENTAL 1 2 7 3.4.1 Materials 1 2 7 3 .4 .2 Acid Hydrolysis of Geodiamolides 1 3 5 1.1. ENDNOTES (CHAPTER 3 : GEODIAMOLIDES FROM CYMBASTELA SP.) 1 3 7 4 PEPTIDES F R O M THE MARINE SPONGE CYMBASTELA SP. 138 4.1 INTRODUCTION 13 8 4 .1 .1 Mitosis 141 4 . 1 . 2 Antimitotic Natural Products 1 4 4 4 .1 .3 The vinca domain 153 4 . 1 . 4 Cytotoxic peptides 1 5 4 4 . 2 RESULTS AND DISCUSSION 1 5 5 4.2 .1 The Hemiasterlins 1 5 7 4 . 2 . 2 TheCriamides 1 7 9 4 .2 .3 Biological activity of the hemiasterlins and criamides 1 8 6 4.3 CONCLUSIONS 193 4 . 4 EXPERIMENTAL 1 9 5 4.4 .1 Materials 1 9 5 4 .5 ENDNOTES (CHAPTER 4 : PEPTIDES FROM CYMBASTELA SP.) 2 0 2 5 SYNTHESIS OF HEMIASTERLIN (86) AND STRUCTURE ACTIVITY STUDY. 205 5.1 INTRODUCTION 2 0 5 5.2 RESULTS AND DISCUSSION 2 1 3 5.2.1 TV-Methyl Homo Vinylogous Valine (MHVV) (173) 2 1 3 5 .2 .2 (-)-Hemiasterlin (86) 2 1 5 5.3 STRUCTURE ACTIVITY STUDY 2 1 8 5.4 EXPERIMENTAL 2 2 2 5 .4 .2 Structure Activity Study 2 2 9 5.5 ENDNOTES (CHAPTER 5: SYNTHESIS) 2 4 3 6 APPENDIX: SELECTED N M R SPECTRA 244 vi L I S T O F F I G U R E S Figure 1. Anticancer leads with significant selective cytotoxic activity in the NCI pre-clinical anti-tumour drug discovery screens. 4 Figure 2. Percentage of specimens per phylum with IC50 < 4 (organic fractions). 4 Figure 3. Body plan of a typical sponge. 10 Figure 4. Carbon skeletons encountered in Aplysilla sp. sponges. 13 Figure 5. Phylogenic classification of the marine sponge Aplysilla sp. 21 Figure 6. Photograph of the marine sponge Aplysilla sp. 22 Figure 7. Map the of collection site for Aplysilla sp. 23 Figure 8. 'H NMR spectrum of botanone A (55). 27 Figure 9. 1 3 C NMR spectrum of botanone A (55). 28 Figure 10. COSY NMR spectrum of botanone A (55). 29 Figure 11. HMQC NMR spectrum of botanone A (55). 30 Figure 12. HMBC NMR spectrum of botanone A (55). 31 Figure 13. Selected HMBC correlations for botanone A (55). 33 Figure 14. Selected isolated spin systems in botanone A (55). 34 Figure 13. Homo-nulcear decoupling experiment on botanone A (55). 35 Figure 16. Selected HMBC correlations for botanone A (55). 36 Figure 17. Results of selected difference nOe experiments for botanone A (55). 37 Figure 18. [ H NMR spectrum of botanone B (56). 38 Figure 19. 1 3 C NMR chemical shifts for ring A in botanone A (55) and botanone B (56). 40 Figure 20. Selected HMBC correlations for botanone B (56). 41 Figure 21. Selected isolated spin systems in botanone B (56). 41 Figure 22. Results of selected difference nOe experiments for botanone B (56). 42 vii Figure 23. 'H NMR spectrum of botanone C (57). 43 Figure 24. 1 3 C NMR chemical shifts for the C9/C11 olefin and for ring A in botanone C (57). 45 Figure 25. Selected isolated spin systems in botanone C (57). 46 Figure 26. Selected HMBC correlations for botanone C (57). 46 Figure 27. Results of selected difference nOe experiments for botanone C (57). 47 Figure 28. *H NMR spectrum of botanicol (58). 50 Figure 29. Selected isolated spin systems in botanicol (58). 51 Figure 30. Selected HMBC correlations from botanicol (58). 51 Figure 31. Results of selected nOe difference experiments on botanicol (58). 53 Figure 32. ] H NMR spectrum of cadlinoglycine (59). 59 Figure 33. Selected isolated spin systems in cadlinoglycine (59). 58 Figure 34. Selected HMBC correlations for sub-structure A of cadlinoglycine (59). 60 Figure 35. Selected HMBC correlations from cadlinoglycine methyl ester (72). 64 Figure 36. lH NMR spectrum of cadlinoglycine methyl ester (72). 62 Figure 37. Expansion of HMBC NMR spectrum of cadlinoglycine methyl ester (72). 63 Figure 38. 1 3 C NMR chemical shifts for ring A of cadlinoglycine (59) and cadlinolide A (47). 64 Figure 39. Selected isolated spin systems in the A ring of cadlinoglycine (59). 65 Figure 40. Selected HMBC correlations for ring A of cadlinoglycine (59). 66 Figure 41. Selected nOe enhancements observed for cadlinoglycine (59). 66 Figure 42. *H NMR spectrum of cadlinolide C methyl acetal (60). 69 Figure 43. Isolated spin systems from the COSY spectrum of 60. 70 Figure 44. Selected HMBC correlations for cadlinolide C methyl acetal (60). 71 Figure 45. nOe Enhancements observed for cadlinolide C methyl acetal (60). 72 Figure 46. [ H NMR spectrum of compound 61. 74 v i i i Figure 47. Phylogenic classification of sponges that produce metabolites in the jaspamide/geodiamolide families. 93 Figure 48. Photograph of Cymbastela sp. sponge. 94 Figure 49. Map of the collection site for Cymbastela sp. 96 Figure 50. ] H NMR spectrum of geodiamolide H (89). 98 Figure 51. *H NMR spectrum of geodiamolide I (90). 101 Figure 52. 'H NMR spectrum of geodiamolide J (91). 103 Figure 53. Selected HMBC correlations for the alanine unit in 91. 105 Figure 54. Fragmentation of geodiamolide J (91) in the electrospray MS/MS spectrum. 106 Figure 55. 'H NMR spectrum of geodiamolide K (92). 108 Figure 56. 'H NMR spectrum of geodiamolide L (93). 111 Figure 57. 1 H NMR spectrum of geodiamolide M (94). 113 Figure 58. Fragmentation of geodiamolide M (94) in the electrospray MS/MS spectrum. 116 Figure 59. *H NMR spectrum of geodiamolide N (95). 118 Figure 60. *H NMR spectrum of geodiamolide P (97). 121 Figure 61. Fragmentation of geodiamolide P (97) in electrospray MS/MS mass spectrum. 123 Figure 62. Compounds in the geodiamolide family. 125 Figure 63. Stages of the cell cycle. 142 Figure 64. Stages of mitosis. 143 Figure 65. Addition of tubulin dimers to a microtubule. 144 Figure 66. Disruption of tubulin/microtubule equilibrium by vinca alkaloids. 146 Figure 67. A proposed model for the vinca domain of tubulin. 153 Figure 68. 'H NMR spectrum of hemiasterlin (86). 158 Figure 69. 1 3 C NMR spectrum of hemiasterlin (86). 159 Figure 70. Chemical shift differences in hemiasterlin (86). 161 ix Figure 71. Computer generated ORTEP drawing of hemiasterlin methyl ester (144). 165 Figure 72. 'H NMR spectrum of hemiasterlin A (98). 167 Figure 73. Selected HMBC correlations for the N\/?,/krimethyltryptophan fragment (A). 170 Figure 74. Selected HMBC correlations for the terMeucine residue (B). 170 Figure 75. Selected HMBC correlations for ./V-methyl homo vinylogous valine residue (C). 171 Figure 76. HMBC correlations establishing the connectivity of hemiasterlin A (98). 172 Figure 77. CD spectra of hemiasterlin (86) and hemiasterlin A (98). 173 Figure 78. lH NMR spectrum of hemiasterlin B (99). 175 Figure 79. COSY and HMBC correlations for the valine residue in hemiasterlin B (99). 177 Figure 80. HMBC correlations establishing the connectivity of hemiasterlin B (99). 177 Figure 81. CD spectra of hemiasterlin (86) and hemiasterlin B (99) 178 Figure 82. *H NMR spectrum of criamide A (100). 180 Figure 83. COSY correlations for the arginine residue in criamide A (100). 182 Figure 84. HMBC correlations establishing the connectivity of criamide A (100). 183 Figure 85. XU NMR spectrum of criamide B (101). 185 Figure 86. Cytotoxicity of the hemiasterlins and microtubule inhibitors. 189 Figure 87. Antimitotic activity of the hemiasterlins and microtubule inhibitors. 191 Figure 88. Microtubule and chromosome distribution. 192 Figure A l . 1 3 C NMR spectrum of botanone B (56). 245 Figure A2. COSY NMR spectrum of botanone B (56). 246 Figure A3. HMQC spectrum of botanone B (56). 247 Figure A4. HMBC spectrum of botanone B (56). 248 Figure A5. 1 3 C NMR spectrum of botanone C (57). 249 Figure A6. COSY NMR spectrum of botanone C (57). 250 Figure A7. HMQC spectrum of botanone C (57). 251 x Figure A8. HMBC spectrum of botanone C (57). 252 Figure A9. 1 3 C NMR spectrum of botanicol (58). 253 Figure A10. COSY NMR spectrum of botanicol (58). 254 Figure A l l . HMQC NMR spectrum of botanicol (58). 255 Figure A12. HMBC NMR spectrum of botanicol (58). 256 Figure A13. 1 3 C NMR spectrum of cadlinoglycine (59). 257 Figure A14. COSY NMR spectrum of cadlinoglycine (59). 258 Figure A15. HMQC NMR spectrum of cadlinoglycine (59). 259 Figure A16. HMBC NMR spectrum of cadlinoglycine (59). 260 Figure A17. 1 3 C NMR spectrum of cadlinolide C methyl acetal (60). 261 Figure A18. COSY NMR spectrum of cadlinolide C methyl acetal (60). 262 Figure A19. HMQC NMR spectrum of cadlinolide C methyl acetal (60). 263 Figure A20. HMBC NMR spectrum of cadlinolide C methyl acetal (60). 264 Figure A21. 1 3 C NMR spectrum of geodiamolide J (91). 265 Figure A22. COSY NMR spectrum of geodiamolide J (91). 266 Figure A23. HMQC NMR spectrum of geodiamolide J (91). 267 Figure A24. HMBC NMR spectrum of geodiamolide J (91). 268 Figure A25. COSY NMR spectrum of geodiamolide M (94). 269 Figure A26. HMQC NMR spectrum of geodiamolide M (94). 270 Figure A27. HMBC NMR spectrum of geodiamolide M (94). 271 Figure A28. 1 3 C NMR spectrum of hemiaserlin A (98). 272 Figure A29. COSY NMR spectrum of hemiaserlin A (98). 273 Figure A30. HMQC NMR spectrum of hemiaserlin A (98). 274 Figure A31. HMBC NMR spectrum of hemiaserlin A (98). 275 Figure A32. 1 3 C NMR spectrum of hemiaserlin B (99). 276 xi Figure A33. COSY NMR spectrum of hemiaserlin B (99). 277 Figure A34. HMQC NMR spectrum of hemiaserlin B (99). 278 Figure A35. HMBC NMR spectrum of hemiaserlin B (99). 279 Figure A36. 1 3 C NMR spectrum of criamide A (100). 280 Figure A37. COSY NMR spectrum of criamide A (100). 281 Figure A38. HMQC NMR spectrum of criamide A (100). 282 Figure A39. HMBC NMR spectrum of criamide A (100). 283 Figure A40. COSY NMR spectrum of criamide B (101). 284 xii L I S T O F T A B L E S Table 1. NMR data for botanone A (55) 32 Table 2. NMR data for botanone B (56) 39 Table 3. NMR data for botanone C (57) 44 Table 4. NMR data for botanicol (58) 48 Table 5. NMR data for cadlinoglycine (59) 57 Table 6. NMR data for cadlinolide C methyl acetal (60) 68 Table 7. 'H NMR data for geodiamolide G (83), geodiamolide H (89), and geodiamolide I (90). 99 Table 8. NMR data for geodiamolide J (91). 104 Table 9. *H NMR data for geodiamolide J (91), geodiamolide K (92), and geodiamolide L (93) 109 Table 10. NMR data for geodiamolide M (94). 114 Table 11. *H NMR data for geodiamolide M (94) and geodiamolide N (95). 119 Table 12. 'H NMR data for geodiamolide D (80) and geodiamolide P (97). 122 Table 13. HPLC retention times for serine Marfey's derivatives in minutes. 136 Table 14. *H and 1 3 C NMR data for hemiasterlin (86). 160 Table 15. Effect of pH on the 1 3 C chemical shifts glycine. 161 Table 16. NMR data for hemiasterlin A (98). 168 Table 17. NMR data for hemiasterlin B (99). 176 Table 18. NMR data for criamide A (100). 181 Table 19. NMR data for the criamides. 186 Table 20. in vitro Cytotoxicities for the hemiasterlins and criamides. 188 Table 21. HPLC retention times for Marfey's amino acid derivatives in minutes. 200 Table 20. Activities of hemiasterlin analogs. 220 xiii L I S T O F A B B R E V I A T I O N S [OC]D specific rotation at wavelength of sodium D line (T °C) Ac acetyl AIDS acquired immunodeficiency syndrome APT attached proton test ax axial Boc terr-butyloxycarbonyl br broad c concentration (g/100 mL) °C degrees Celsius CD circular dichroism cm"1 wavenumbers COSY correlation spectroscopy 8 NMR chemical shift in parts per million d doublet Da Daltons dd doublet of doublets ddd doublet of doublets of doublets AM difference in mass DMAP 4-(dimethylamino)pyridine DMSO dimethyl sulfoxide dt doublet of triplets ED5o effective dose resulting in 50% response eq equatorial Et ethyl Et 20 diethyl ether EtOAc ethyl acetate FABMS fast atom bombardment mass spectrometry FDAA 5-fluoro-2,4-dinitrophenyl-Z,-alanineamide xiv Fmoc 9-fluorenylmethyloxycarbonyl FTIR Fourier transform infrared g gram GC gas chromatography h hour HREIMS high resolution electron impact mass spectrometry HRDCIMS high resolution desorption chemical inonization mass spectrometry HRFABMS high resolution fast atom bombardment mass spectrometry HMBC heteronuclear multiple bond connectivity HMQC heteronuclear multiple quantum coherence HPLC high performance (pressure) liquid chromatography Hz hertz i signal due to an impurity IC50 inhibitory concentration resulting in 50% response J scalar coupling constant X wavelength LH-20 Sephadex LH-20 l.r. long range M molar m multiplet M + molecular ion m/z mass to charge ratio Me methyl mg milligram MHz megahertz mL milliliter mmol millimole mmu millimass units p micro (10"6) N normal n nano (10"9) nOe nuclear Overhauser effect ppm parts per million XV r.t. room temperature q quartet Rf ratio to front s signal due to solvent s singlet SCUBA self contained underwater breathing apparatus sp. species t triplet td triplet of doublets T/C test versus control TFA trifluoroacetic acid TLC thin layer chromatography UV ultraviolet w signal due to water Z benzyloxycarbonyl xvi L I S T O F S C H E M E S Scheme 1. Proposed biogenesis of botanicol (58) from ambliofuran (20). 55 Scheme 2. Proposed biogenesis of botanone B (56) from the intermediate carbocation 66. 55 Scheme 3. Proposed formation of cadlinolide C (49) and cadlinolide C methyl acetal (60) from 61. 75 Scheme 4. Proposed formation of cadlinoglycine (59) from 62. 76 Scheme 5. Chemical degradation of hemiasterlin (86). 164 Scheme 6. Proposed mechanism of PyBroP-mediated peptide coupling. 208 Scheme 7. Formation of 7V-Me-NCA (166) from JV-Boc-JV-Me-Val-OH (163). 209 Scheme 8. Retrosynthetic analysis of hemiasterlin (86). 211 Scheme 9. Synthesis of iV'-Boc-iV,A '^,yf7,y5-tetramethyltryptophan (170). 212 Scheme 10. Schreiber's route to vinylogous amino acids. 213 Scheme 11. Synthesis of MHVV-TFA (193). 214 Scheme 12. Synthesis of (-)-hemiasterlin (86). 216 xvii A C K N O W L E D G M E N T S First, I would like to express my appreciation to my research supervisor, Professor Raymond J. Andersen, for his endless encouragement, excellent guidance, and support. For their assistance in collecting and photographing sponge samples, I thank Michael LeBlanc and David Williams for Cymbastela sp. and Julia Kubanek for Aplysilla sp. Michel Roberge and Hillary Anderson must be thanked for their help with the MCF-7 and antimitotic bioassays, and also for their many helpful impromptu microbiology lectures. The staff of Dr. Theresa Allen's laboratory, specifically Sarah Halleran and Jennifer Sandrowski, I must also thank for their help with all the other bioassays, both in vitro and in vivo. For their advice on synthesis, I thank Serge Boulet, Debra Wallace, James Nieman, Patricia Gladstone, and Todd Schindeler in Dr. Edwards Piers' laboratory. The staff of the NMR and Mass Spectroscopy facilities, as well as Brian Patrick in James Trotter's X-ray laboratory, must be thanked for their assistance in collecting data. Finally, I am indebted to Julia Kubanek, Pat Gladstone, and Michel Roberge for their assistance in proofreading this thesis. xviii 1 G E N E R A L I N T R O D U C T I O N i 1.1 MARINE NATURAL PRODUCTS CHEMISTRY As the human species approaches the new millenium, it is faced with the increased threat of diseases. While cancer is currently the leading cause of death in industrialized countries,1 humans are also subject to other deadly diseases, both new and re-emerging, for which we possess no effective treatment. These include Acquired Immune Deficiency Syndrome (AIDS), malaria, tuberculosis, and a host of multi-drug-resistant bacterial infections. These threats reinforce the need to identify novel active leads for drug development. Since the discovery of 2 3 effective antibiotics isolated from bacterial cultures in the 1940's and 1950's, ' scientists have successfully discovered new, structurally diverse compounds with exciting biological activities from natural sources which have been developed as pharmaceuticals, agrochemicals, growth regulating substances and molecular probes.1 A recent review, based on data published in Annual Reports of Medicinal Chemistry from 1984 to 1995, indicated that over 60% of the approved drugs and candidates for pre-New Drug Application were natural products or compounds derived from natural products.4 This, combined with the fact that close to half of the best selling pharmaceuticals in 1991 were either natural products or their derivatives, shows striking evidence of the importance of natural products.4 Natural products are secondary metabolites whose role in the daily existence of an organism is not fully understood. The most commonly hypothesized role for these compounds is self-defense, but anti-microbial, and UV-protective functions have also been proposed. " Living 2 organisms investigated as sources of natural products can be divided into three main groups: plants, animals, and microbes. Historically, natural products and their derivatives developed by the pharmaceutical industry have come from terrestrial plant and microbial sources.4 Yet the oceans, which are home to representatives of 26 of the 28 principal animal phyla (8 being exclusively aquatic), whose bio-diversity of species and habitat make them possibly the greatest source of natural products, have only begun to be explored. Since its explosion in the early 1970's, research in the field of marine natural products has yielded thousands of compounds isolated from marine invertebrate animals, many with novel o structures and exciting pharmacological applications. Marine organisms that are soft-bodied, slow growing, either sessile or slow moving, those that have bright coloration or that live in exposed habitats are especially appropriate to target for investigation. The rationale for this is these organisms often rely on chemical defenses to avoid predation, and that structurally complex organic molecules may play a role in this defense mechanism.* The main groups of organisms that fit these criteria include sponges, corals, bryozoans, coelenterates, ascidians, and gastropods. Sponges were the first marine invertebrates to be investigated as a possible source of new compounds, and to date have yielded the greatest number of novel compounds across a very diverse chemical spectrum.8 The beginning of marine natural products chemistry can be traced to the 1950's when Bergmann isolated two ara-nucleosides from the Caribbean sponge Crypotethya crypta.10 These nucleosides have ribose replaced by an arabinose unit (1, 2). The discovery of this new class of compounds stimulated the synthesis of a series of arabinosyl nucleosides, such as ara-A (3) and ara-C (4). Ara-A (3), which has been used as a treatment for * Recent studies have shown there is no correlation between toxicity and palatability in sponges; unpalatable extracts were about equally likely to be toxic or nontoxic, and some palatable extracts were quite toxic (although no very palatable extracts were also toxic). Hence, Pawlik has suggested that biologists and natural product chemists should abandon the practice of inferring that the toxicity of extracts or compounds has ecological significance.9 3 Herpes encephalitis since the 1970's, was subsequently isolated from the Mediterranean gorgonian coral Eunicella cavolini.11 Ara-C (4) is widely used today for the treatment of leukemias.12 The discovery of these unusual nucleosides was responsible for the initial interest 13 in sponges, and the oceans in general, as a potential source of new chemical compounds. OH OH 1 R = C H 3 3 4 2 R = H Recent statistical data from the National Cancer Institute (Figure 1) indicates that a much larger percentage of bio-active leads are resulting from the screening of marine animals than from terrestrial plants or microorganisms.14 Of over 6 000 animals screened, close to 2% showed significant cytotoxic activity, whereas screening of over 18 000 terrestrial plants and over 8 000 microorganisms resulted in a hit rate of less than 1%. When the marine data are separated into phyla (Figure 2), it is clear that sponges are unquestionably the richest source of anti-cancer leads.14 One of the challenges for a chemist studying marine natural products is to find renewable sources of compounds displaying promising biological activity. Marine secondary metabolites are commonly present in concentrations of parts-per-million by weight in the individual marine animals. It is therefore difficult to obtain the quantities of compounds needed for clinical trials without plundering the oceans by repeatedly collecting huge numbers of animals, a practice both 4 Terrestrial Animals Terrestrial Plants 434 Absolute number screened in italics 18293 Marine Animals 6540 Marine Plants 1872 Microorganisms 0 8246 1 1 Percentage 2 Figure 1. Anticancer leads with significant selective cytotoxic activity in the NCI pre-clinical anti-tumour drug discovery screens.14 12 v o LT> u y x o u CP 8 h 0 56 1041 154 100 Phylum Figure 2. Percentage of specimens per phylum with IC50 < 4 (organic fractions) 14 5 ecologically irresponsible and non-sustainable. Total laboratory synthesis has been one of the most popular methods of overcoming this problem. The advantages to this approach include the potential of an infinite supply of the compound, the opportunity to make analogs and an independent confirmation of the compound's structure and activity. The importance of this last point was illustrated recently by Pettit and Taylor in their synthesis of stylopeptide (5), a cycloheptapeptide isolated from the marine sponge Stylotella aurantium which was reported to be cytotoxic against P388 lymphocytic leukemia cells.15 Synthetic stylopeptide (5), which was spectroscopically identical to the natural product, was inactive when tested for cytotoxicity, while natural samples maintained their activity, suggesting the presence of an unknown active contaminant in the natural samples.15 This example illustrates the need to verify biological activity. 5 Unfortunately not all natural products are currently possible to synthesize because of their complex structures: either the synthesis is not economically viable because it involves too many steps and the yields are too low, or it is not possible with current technology. An alternative source of the desired compounds is aquaculture. Organisms suitable for farming can be grown and the natural products isolated on a large scale. Bryostatin-1 (6), one of a group of macrocyclic lactones isolated from the marine bryozoan Bugula neritina, is currently used in clinical trials in the U.S. for the treatment of cancer, and represents one of the few aquaculture success stories in marine natural products.16 The small quantities of bryostatin-1 (6) available from the naturally occurring Bugula neritina make this bryozoan a non-viable source of the compound, and at present an efficient synthetic route to 6 does not exist. Aquaculture has proven to be key to the success of 6 as a potential drug. Bugula neritina larvae were sown onto perforated plastic plates, stacked on submerged towers in the open ocean, where they were allowed to grow for 5 months. An innovative isolation scheme designed for the maximum recovery of 6 was used after the harvest and after 10 months, 18 g of pharmaceutical-grade bryostatin-1 (6) were produced.16 Another renewable source of desired compounds from marine organisms may be from associated microorganisms that are harbored within the animal. Recent studies have suggested that many secondary metabolites isolated from marine invertebrates such as sponges, coelenterates, and mollusks, are products of associated microorganisms rather than products of the invertebrate's metabolism.17 These microorganisms, present in or on the invertebrate, are 1 S usually zooxanthellae, cyanobacteria, or photosynthetic filamentous bacteria. In sponges, the volume represented by the associated microorganisms is frequently of the same order as the 1 & sponge tissue itself. As is the case with aquaculture, there are many challenges involved in reproducing the exact conditions required for these microbes to flourish, making the culturing of marine microorganisms obtained from invertebrates tissues very difficult. Sponges are sedentary, appear to have few physical defenses, and have few predators. To protect themselves against predation, these animals have developed a remarkable arsenal of chemical defenses which deter prospective attackers and assure success in the battle for space on the ocean floor. Among the few organisms that do feed on sponges, primarily opistobranch mollusks, some have developed the ability to sequester defensive substances from their sponge diet.19 A better understanding of sponge secondary metabolism may be used to help clarify sponge classification, based on an hypothesis that each order, genus and possibly each species has a signature suite of metabolites.13 Some sponge metabolites have been shown to have chemotaxonomic significance. For example, brominated amino acid metabolites have been repeatedly isolated from sponges of the order Verongida.13 This work was pioneered by Bergmann, who attempted to show a relationship between classification of marine sponges and their steroids.10 8 1.2 SPONGE BIOLOGY Sponges represent the simplest and most primitive multicellular animals (metazoans). Recognized as animals since 1825, sponges belong to the phylum Porifera, meaning pore-bearing. Currently there are about 15 000 recognized species of sponges. They are sessile animals that are found most abundantly in shallow coastal waters, attached to hard substrates such as rocks, pilings, or animal shells, and can be found from the intertidal zone to the dark abyss. Approximately 200 species of sponges have adapted to fresh water. Their ecological success can be partly attributed to their simplicity, which allows for adaptation to an infinite 18 number of ecological niches. The morphological diversity of sponges is spectacular. Some sponge colonies are shaped like large urns and exceed a meter in height and diameter, while other encrusting sponges are barely a millimeter thick.18 Although there are some sponges that are radially symmetrical, 18 sponge colonies more commonly conform to the shape of the substrate. Taxonomists have divided sponges into four classes, according to their skeletal elements (which are made up of flexible spongin fibers or mineralized spicules): Calcarea, Hexactinellidae, Demospongiae, and Sclerospongiae classes. Sponges with CaCC»3 spicules belong to the class Calcarea. Species in this class are the smallest of all sponges, reaching less than 10 cm in height. They are found in all the oceans of the world, commonly in protected 20 intertidal and shallow-water habitats such as crevices, overhangs, caves and tunnels. Class Hexactinellidae sponges contain hexagonal siliceous (SiC^) spicules. Some of the spicules are fused to form a lattice-like skeleton built of long siliceous fibers, and hence, are commonly called "glass sponges". These sponges average 10 to 30 cm in height and form symmetrical cup, vase, and urn-like shapes. Hexactinellidae sponges are found in tropical water at depths between 400 and 950 meters.20 9 The largest class is the Demospongiae, encompassing over 95% of all sponges. These sponges may contain either siliceous spicules or spongin fibers, or a combination of both. Demosponges occur in all marine habitats, from intertidal to abyssal depths of 5 000 meters, as 90 well as freshwater rivers, streams, and lakes. Sponges in the class Sclerospongiae build a skeleton that contains both silica and calcite spicules, as well as spongin. This rarely-encountered class of sponge prefers caves, tunnels, 90 crevices and deep-water overhangs on coral reefs. Unlike other metazoans, sponges lack nervous systems and have no true musculature. In 21 fact, sponges do not have any organs at all. As filter feeders, they have specialized in the retention of very fine particles, for which there is very limited competition. The body of a sponge is pierced with small dermal or incurrent pores called ostia. Water, including suspended particulate matter and oxygen, is brought into the sponge through its many ostia, and exits from the fewer large excurrent pores —called oscula— by way of a complex system of passageways and cavities containing flagellated collar cells known as choanocytes (Figure 3). Organic particles and microbes are the sponge's primary food source and are ingested mainly by the choanocytes. Respiration occurs by osmotic transfer of oxygen from the incoming flow of water across the cell membranes. The outflowing water carries away waste and may also be used to 90 move sperm and eggs out of the sponge. In some places, sponges play an important ecological role. For example, it has been estimated that on the exterior slopes of some Caribbean reefs, sponges are so abundant and active that over a 24-hour period they are able to filter the overlying 1 R 25- to 40- meter water column. Figure 3. Body plan of a typical sponge. 1.3 ENDNOTES: CHAPTER 1: GENERAL INTRODUCTION 1) Hay, M . W.; Fenical, W. Oceanography 1996, 9, 10. 2) Fleming, A. Brit. J. Exp. Pathol 1929,10, 226. 3) Waksman, S. A. Streptomycin; Williams and Wilkins: Baltimore, 1949. 4) Cragg, G. M ; Newman, D. J.; Snader, K. M. J. Nat. Prod. 1997, 60, 52. 5) McClintock, J. B.; Gauthier, J. J. Antarct. Sci. 1992, 4, 179. 6) Paul, V. J. Chemical defenses of benthic marine invertebrates. In Ecological Roles of Marine Natural Products; Paul, V. J., Ed.; Comstock Publishing: Ithaca, New York, 1992, pp 164-188. 7) Pawlik, J. R. Oceanogr. Mar. Biol. Ann. Rev. 1993, 30, 273. 8) Faulkner, D. J. Nat. Prod. Rep. 1997,14, 259. 11 9) Pawlik, J. R.; Chanas, B.; Toonen, R. J.; Fenical, W. Mar. Ecol. Prog. Ser. 1995, 727, 183. 10 11 1 2 13 14 15 16 17 18 19; 20 21 22 Bergmann, W.; Feeney, R. J. J. Amer. Chem. Soc. 1950, 72, 2809. Cinino, F.; De Rosa, S.; De Stefano, S. Experientia 1984, 40, 339. Pratt, W. B.; Ruddon, R. W.; Ensminger, W. D.; Maybaum, J. The Anticancer Drugs; 2 n d ed.; Oxford University Press: New York, 1994, p 96. Bergquist, P. R. Sponges; University of California Press: Los Angeles, 1978, pp 203-216. Garson, M . J. The biosynthsis of sponge secondary metabolites: Why it is important? In Sponges in Time and Space; van Soest, R. W. M. , van Kempen, Th. M . G., Braekman, J. C , Eds.; A. A. Balkema Publishers: Rotterdam, Netherlands, 1993, pp 427-440. Pettit, G. R.; Taylor, S. R. J. Org. Chem. 1996, 61, 2322. Schaufelberger, D. E.; Koleck, M . P.; Beutler, J. A.; Vatakis, A. M. ; Alvarado, A. B.; Andrews, P.; Marzo, L. V.; Muschik, G. M . Journal of Natural Products 1991, 54, 1265. Kobayashi, J.; Ishibashi, M . Chem. Rev. 1993, 93, 1753. ' de Vos, L.; Riitzler, K.; Boury-Esnault, N. ; Donadey, C ; Vacelet, J. Atlas of Sponge Morphology; Smithsonian Institution Press: Washington, 1991. Cimino, G.; Sodano, G. Transfer of sponge secondary metabolites to predators. In Sponges in Time and Space; van Soest, R. W. M. , van Kempen, Th. M . G., Braekman, J. C , Eds.; A. A. Balkema Publishers: Rotterdam, Netherlands, 1994, pp 459-472. Meglitsch, P. A.; Schram, F. R. Invertebrate Zoology; 3 r d ed.; Oxford University Press: New York, 1991, pp 54-67. Pechenik, J. A. Biology of the Invetebrates; Prindle, Weber & Schmidt: Boston, 1985, pp 75-85. Microsoft Microsoft Encarta 97 Encyclopedia [CD-ROM]; 1997. 12 2 D I T E R P E N O I D S F R O M T H E M A R I N E S P O N G E APLYSILLA SP. 2.1 INTRODUCTION Marine natural products chemists investigating sponges collected from around the world have isolated thousands of novel compounds belonging to a variety of different classes, from simple mono-terpenoids to large cyclic peptides.1 Of all the different classes of compounds, terpenoids are the most abundant non-steroidal secondary metabolites isolated from marine sponges, representing approximately 45 % of the total sponge metabolites.1 The natural product chemistry of Aplysilla sp. follows this trend and is characterized by diterpenoid derivatives of the hypothetical 'spongian'2 skeleton 7 (see Figure 4). To date, over forty compounds have been isolated from Aplysilla sponges, of which twenty-seven are diterpenoids representing nine carbon skeletons (Figure 4). The following is a brief review of these diterpenoids. Karuso's studies of the sponge Aplysilla sulphurea resulted in the isolation of the novel spongian-derived diterpenoid aplysulphurin (18), the structure of which was determined by spectroscopic means and confirmed by single-crystal X-ray diffraction analysis. No biological data was reported for aplysulphurin (18). OAc O 18 13 spongian 7 degraded spongian 8 marginatane 9 macfarlandin 10 aplysulphurane 11 norrisane 12 dendrillane 13 glaciane 14 gracilane 15 labdane 16 degraded labdane 17 Figure 4. Carbon skeletons encountered in Aplysilla sp. sponges. Kazlauskas et al. reported the isolation of the novel diterpenoid triacetate aplysillin (19) from the sponge Darwinella sp. (order Dendroceratida; previously referred to as Aplysilla rosea), collected in New Zealand.4 The structure of 19 was solved by a combination of spectroscopic analyses and the relative stereochemistry determined by single crystal X-ray diffraction analysis. Re-investigation of the Darwinella sp., identified as Darwinella oxeata, collected in the same 14 locality, revealed no trace of aplysillin (19), but rather afforded the known diterpenoids aplysulphurin (18) and ambliofuran (20), and the novel diterpenoids tetrahydroaplysulphurins 1-3 (21-23).5 9 A c OAc Investigation of an Australian Aplysilla sp. by Mol insk i et al. resulted in the isolation of ambliofuran (20) and two new diterpenoid lactones 24 and 25 with the 'spongian' skeleton, 6 which were closely related to the lactone 26 isolated from the sponge Igernella notibilis by Schmitz et a l . 7 The structure of ambliofuran (20) was confirmed by comparison of its spectroscopic data with the literature values, while the structures of lactones 24 and 25 were determined by interpretation of their spectroscopic data. Bobzin and Faulkner's investigation of the Californian marine sponge Aplysilla polyrhaphis resulted in the isolation of nine diterpenoids, 27-34.9 O f these nine compounds, five were known nudibranch or sponge metabolites while four were novel. The major metabolite, macfarlandin E (27), was previously reported from the dorid nudibranch Chromodoris macfarlandi,10 while the analogous mono-acetate derivative, aplyviolene (28), was first isolated 15 from the sponge Chelonaplysilla violacea.11 Norrisolide (29) was originally found in the nudibranch Chromodoris norrisi, shahamin C (30) was reported from an undescribed species of Dydidea sponge13, and the y-lactone 31 was isolated from the sponge Aplysilla rosea.14 The structure of the novel compounds, polyrhaphins A-D (32-35), were determined by extensive spectroscopic analysis. 27 R = OAc 29 28 R = H OAc OAc 30 31 Macfarlandin E (27), norrisolide (29), shahamin C (30), and polyrhaphin A (32) have also been isolated from specimens of the dorid nudibranch Chromodoris norrisi that were collected in 16 the same locality as Aplysilla polyrhaphis, which is the presumed dietary source of these compounds. 9 The observation that C. norrisi sequesters metabolites from A. polyrhaphis implies that some or all of these terpenoids 2 7 - 3 5 may have ecological significance to the nudibranch. In an anti-feedant bioassay, shahamin C ( 3 0 ) , the y-lactone 3 1 , and polyrhaphin C ( 3 4 ) were found to be active, deterring the feeding of rainbow wrasse Thalassoma lucasunum at concentrations of 100 u,g metabolite/mg food. Macfarlandin E ( 2 7 ) , aplyviolene ( 2 8 ) , norrisolide ( 2 9 ) , and polyrhaphin A ( 3 2 ) were all inactive. Both aplyviolene ( 2 8 ) and polyrhaphin C ( 3 4 ) have been shown to possess moderate anti-microbial activity. 9 OAc Polyrhaphin D ( 3 5 ) was mistakenly described as the first example of a diterpenoid with the 'isospongian' skeleton. The 'isospongian' skeleton is identical to the 'marginatane' skeleton (9 ) first encountered in marginatafuran ( 3 6 ) , a diterpenoid isolated from the dorid nudibranch Cadlina leutomarginata}5 17 36 Investigation of the marine sponge Aplysilla rosea by Karuso and Taylor led to the isolation of ten diterpenoids, the known compounds ambilofuran (20) and hexahydroambliofuran (37), the novel compounds aplyroseol-1 to 6 (38-43), and the lactones 44 and 45.14 The structures of the novel diterpenoids were established by interpretation of their spectroscopic data. Although the absolute stereochemistry of the aplyroseols has not been proven, it is assumed to be the same as that of isoagatholactone (46), whose absolute stereochemistry is known.16 No biological testing was done on these compounds. 37 O aplyroseol-1 aplyroseol-2 aplyroseol-3 aplyroseol-4 aplyroseol-5 aplyroseol-6 (38) (39) (40) (41) (42) (43) OCOPr OAc OCOPr OCOPr OH OAc H H OH OAc OCOPr OCOPr O O 44 R = OAc 45 R = II 46 18 Tischler's investigations of the marine sponge Aplysilla glacialis resulted in the isolation of seven novel diterpenoids: the five 'spongian'-derived terpenoids cadlinolide A to C (47-49), aplysillolide A (50), and aplysillolide B (51), and a 'marginatane' diterpenoid, marginatone (52). As well, glaciolide (53), a rearranged 'spongian' diterpenoid possessing the 'glaciane' carbon skeleton (14) was isolated.17"19 Cadlinolide C (49) was thought to be an isolation artifact of cadlinolide A (47).19 Cadlinolides A to C (47-49) are structurally similar to the tetrahydroaplysulphurins-1 to 3 (21-23). Minor amounts of tetrahydroaplysulphurin-1 (21), cadlinolide A (47), and glaciolide (53) were isolated from the dorid nudibranch C. luteomarginata found feeding on A glacialis.11'n The origin of tetrahydroaplysulphurin-1 (21) in the skin extracts of C. luteomarginata has been explained by the possibility that the nudibranch has developed the ability to convert cadlinolide 1 R B (48) into its acetate 21, since 48 was not found in the extracts of A. glacialis. There is also the possibility that tetrahydroaplysulphurin-1 (21) is present in the sponge, but at very low concentrations, and C. luteomarginata selectively sequesters 21 and not cadlinolide B (48). 51 52 53 19 The following section will describe the chemical study of an Aplysilla sp. sponge collected at Botanical Beach, on Vancouver Island, B.C. Extracts from the Aplysilla sp. sponge have yielded the two known sponge diterpenoids ambliofuran (20) and hexahydroambliofuran (37), the two known nudibranch diterpenoids marginatafuran (36) and 9,11-dihydrogracillin A (54), and six novel diterpenoids (55-60). Botanones A to C (55-57) possess the rare degraded 'labdane' skeleton 17, botanicol (58) has the 'labdane' skeleton 16, while cadlinoglycine (59) and cadlinolide C methyl acetal (60) both possess the 'aplysulphurane' skeleton 11. The following section describes the structure elucidation of the novel compounds 55 to 60. 59 60 2.2 RESULTS 20 Andersen et al. have reported the isolation of seven diterpenoids and two secosteroids 17 1 R 7fi from the marine sponge Aplysilla glacialis. ' ' Pika observed the presence of additional minor terpenoid containing fractions in extracts of A. glacialis, prompting our investigation of a pink encrusting sponge that was initially characterized as Aplysilla glacialis.21 Upon investigation of the pink sponge's chemistry, it became apparent that the initial classification of the sponge was incorrect. Although the sponge did produce diterpenoids, they were not characteristic of Aplysilla glacialis, but instead of a different species of sponge, possibly an undescribed species of Aplysilla. Re-examination of the pink sponge specimen by Bergquist, in light of its secondary metabolites, indicated it was a previously undescribed species of Aplysilla22 (Figures 5 and 6). The Aplysilla sp. sponge was collected and investigated twice over a four year period. In May of 1991, Aplysilla sp. was collected by hand (60 g wet weight) from the high intertidal zone at Botanical Beach on Vancouver Island, British Columbia (see map, Figure 7). Freshly collected sponge was immersed in methanol and stored at 4°C for one and a half years. Removal of the methanol by filtration followed by repeated extraction with mefhanol/CHiCb (1:1, 2 x 250 mL) and concentration in vacuo of the combined organic extracts yielded a green oil (3.8 g). Sequential partitioning of the oil between hexanes and brine, then ethyl acetate and brine, followed by drying of the organic layers over anhydrous Na2S04, and concentration in vacuo afforded a non-polar hexanes fraction (1.2 g) and a polar ethyl acetate fraction (0.74 g). Further fractionation of the hexanes fraction by silica gel flash chromatography and the ethyl acetate fraction by reversed-phase flash chromatography yielded several fractions containing mixtures of fats, pigments, steroids, and terpenoids as detected by analytical TLC and 'H NMR analysis. HPLC purification of terpenoid fractions yielded the known diterpenoid marginatafuran (36), the novel diterpenoids botanones A to C (55-57), botanicol (58), and two suspected artifacts, cadlinoglycine (59) and cadlinolide C methyl ester (60). Kingdom Phylum Class Subclass Order Family Genus Metazoa (multi-cellular animals) Porifera (sponges) Demospongiae Ceractinomorpha Dendroceratida Aplysillida Apllysilla Figure 5. Phylogenic classification of the marine sponge Aplysilla sp. according to Berquist (1993) 22 22 Figure 6. Photograph of the marine sponge Aplysilla sp. Recollection of the Aplysilla sp. sponge in September 1994 was made in an attempt to determine if cadlinoglycine (59) and cadlinolide C methyl ester (60) were isolation artifacts resulting from the 1.5 year methanol extraction of the initial collection. Freshly collected sponge was immediately immersed in methanol, where it remained for two days, after which time the methanol was decanted off and the sponge re-extracted with a solution of methanol/CF^CL (1:1, 2 x 250 mL). After filtration and concentration of the organic layers in vacuo, progression through the same isolation procedure described above yielded marginatafuran (36), botanones A to C (55-57), and botanicol (58). The known diterpenoids ambliofuran (20), hexahydroambliofuran (37), and 9,11-dihydrogracillin A (54), which were not present in the first Aplysilla sp. extract, were also isolated. Although cadlinoglycine (59) and cadlinolide C methyl acetal (60) were not found in the extract, there was 'H NMR evidence for 61, a compound that 2 3 could possibly be their precursor, as well as a precursor to cadlinolide A (47) and the suspected isolation artifact cadlinolide C (49). 0 61 Copies of 1 3 C NMR spectra and two dimensional NMR spectra (COSY, HMBC, and HMQC) for compounds 5 5 to 60 not included in the main body of the thesis can be found in Appendix A. 49.5 N 48.5 N 125 W 123 W Figure 7. Map of the collection site for Aplysilla sp. 24 2.2.1 Known Diterpenoids Marginataruran (36) and 9,11-dihydrogracillin A (54) are known nudibranch metabolites which have been found repeatedly in skin extracts of the dorid nudibranch C. luteomarginata}5'211'24 Both 36 and 54 isolated from Aplysilla sp. were identified by comparison of their spectroscopic data with the literature values.15'23 Discovery of marginatafuran (36) and 9,11-dihydrogracillin A (54) in the sponge Aplysilla sp. represents the first time that a dietary sponge source has been found for these compounds 2 4 9,11-Dihydrogracillin A (54) has been reported to possess moderate anti-microbial activity.23 The known sponge metabolites ambliofuran (20) and hexahydroambliofuran (37) were also identified by comparison of their spectroscopic data with the literature values.8 36 54 The isolation of ambliofuran (20), hexahydroambliofuran (37), and 9,11-dihydrogracillin A (54) exclusively from the second collection of Aplysilla sp., possibly could be explained by seasonal variations, since the two collections were made in May and October, respectively. An 25 alternative explanation comes from observations made by Pika when studying the marine sponge Pleraplysilla sp.25 Pika found that repeated collections of the same patch of sponge resulted in simplification of the array of isolated secondary metabolites.25 Two hypotheses were proposed to explain this phenomenon. The first hypothesis suggested that the sponge's growth rate and metabolism are sufficiently slow that it cannot elaborate its full array of secondary metabolites in the time between collections.25 As a result, the metabolites that are isolated are the precursors to the final products that the sponge produces. A second theory, based on the fierce competitiveness of the ocean floor, suggested that the sponge suffers reduced competitive effectiveness because of repeated collections, and responds by producing compounds which are the most effective against predation and competing animals.25 The isolation of the linear diterpenoids 20 and 37 is in accordance with the first hypothesis since ambliofuran (20) is a logical starting point in the biogenesis of all of the diterpenoids isolated from this sponge. The presence of the anti-microbial 9,11-dihydrogracillin A (54), a rearranged and functionally modified diterpenoid, is consistent with the latter hypothesis. An alternative explanation for 20, 37, and 54 not being isolated from the first collection of Aplysilla sp. is that they were degraded during the one and half years that the sponge sat immersed in methanol. Without further investigation, involving careful repeated collections of the sponge, it would be difficult to determine if any of the above hypotheses are correct. 26 2.2.2 Novel Diterpenoids 2.2.2.1 Botanone A (55) Botanone A (55) was obtained as a colourless oil that gave an (M+H)+ peak in the HRDCIMS at m/z 360.2306, corresponding to a molecular formula of C22H32O4, which requires seven degrees of unsaturation. Table 1 provides a summary of the NMR data obtained for compound 55. Resonances for the twenty-two carbon atoms are well resolved in the 1 3 C NMR spectrum (Figure 9), and the HMQC experiment (Figure 11) indicated that all 32 hydrogen atoms are attached to carbon atoms. The infrared spectrum contains two carbonyl stretches at 1772 and 1741 cm"1, indicating the presence of two carbonyl groups in the molecule. The presence of two resonances in the C NMR spectrum of 55 at 5 170.1 and 204.1 confirmed the presence of two carbonyl moieties and identified them as ester and ketone functionalities, respectively. A disubstituted double bond was identified by two carbon resonances at 8 144.8 (C9) and 5 124.7 (Cl 1) in the 1 3 C NMR spectrum, which were correlated to the proton resonances at 5 5.37 (d, J = 16 Hz: H9) and 5.34 (m: HI 1), respectively, in the HMQC spectrum of 55. Four deshielded 1 3 C NMR resonances at 5 139.6 (CH20), 143.2 (CH15), 124.3 (C13) and 111.6 (CH14), and the three attached protons (8 7.19 (s: H20), 7.18 (br s: H15) and 6.30 (br s: HI4)), were characteristic of a mono-substituted furan ring system. The ester, ketone, furan, and 27 28 30 55 H15 H20 U oo u u u In Mel7 Mcl9 <* o O 'i i i i i i i i i i i i i i i i i i i i i i i i | i i i i i i i i i | i i i i i i i i i | i i i i i i i i i | i i i i i i i i i | (Ppm) 6 5 4 3 2 1 a a . 20 40 60 80 100 120 140 Figure 11. HMQC NMR spectrum of botanone A (55). Recorded in C 6D 6 at 500 MHz. 31 55 H15 H20 |H11 H 9 .11 | H 7 0 1 O A c M e l 7 M e l 9 0 0 OD I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I I 40 60 80 100 120 140 160 180 200 (ppm) Figure 12. H M B C N M R spectrum of botanone A (55). Recorded in C 6 D 6 at 500 MHz. 32 Table 1. NMR Data for botanone A (55) (Recorded in C 6 D 6 at 500 MHz.) Carbon No. 5 1 3 C S ' H COSY H M B C 1 40.8 1.23 m HI', 2', 2 H9 r 1.31 m HI, 2' 2 19.0 1.45 ttJ= 13, 3 Hz HI, 2', 3, 3' 2' 1.28 m HI, 1', 2,3,3' 3 42.2 1.07 dtJ= 13, 3 Hz H2, 2', 3' H16 3' 1.26 m H2, 2', 3 4 34.5 H3', 16, 17 5 48.9 1.19 dd J=6,3 Hz H6, & HI, 1', 6', 16, 17 6 28.2 1.65 dddJ= 15,6, 5, Hz H5, 6', 7 H5 6* 1.55 ddd J= 15, 10, 4 Hz H5, 6, 7 7 80.3 5.16 ddJ= 10, 5 Hz H6, 6' H5,6' 8 204.1 H18 9 144.8 5.37 d J = 16 Hz H l l H12, 19 10 40.7 HI, 1', 6', 19 11 124.7 5.34 m H9, 12 H12 12 28.9 3.06 d .7=6 Hz H l l , 20 (l.r.) H9 13 124.3 H12, 14, 15,20 14 111.6 6.30 br s H15 H12, 15,20 15 143.2 7.18 brs H14 H14, 20 16 22.1 0.78 s 17 33.9 0.86 s 18 25.6 1.81 s 19 18.0 0.85 s 20 139.6 7.19 s H15 H15 21 170.1 22 20.2 1.72 s carbon-carbon double bond functionalities accounted for all four oxygen atoms in the molecular formula, and six of the seven required sites of unsaturation. No other unsaturated functional groups were indicated by the 1 3 C NMR data; therefore the remaining site of unsaturation had to be present as a ring. HMBC correlations from a pair of methyl singlet resonances at 8 1.72 (Me22) and 1.81 (Mel8) to the carbonyl resonances at 8 170.1 (C21) and 204.1 (C8), respectively, identified an acetate and a methyl ketone functionality (Figure 13). Presence of the acetate was also confirmed by an ion measuring m/z 300 in the mass spectrum, representing the loss of 60 mass 33 units from the molecular ion. The deshielded chemical shift of the carbon resonance at 8 80.3 (C7), correlated to a proton at 8 5.16 (dd, J = 10, 5 Hz: H7) in the HMQC spectrum, identified it as the site of attachment of an oxygen atom. Strong HMBC correlations between the methine resonance at 8 5.16 (H7) and the carbons assigned to the acetate (C21: 8 170.1) and ketone (C8: 8 204.1) carbonyl groups established the C7/C8/C18 connectivity, and placed the acetate group on C7. Figure 13. Selected HMBC correlations for botanone A (55). COSY and HMQC data identified an isolated spin system starting from the acetoxy methine (H7: 8 5.16), which correlated to a pair of methylene resonances at 8 1.65 (ddd, J = 15, 6, 5 Hz: H6) and 1.55 (ddd, J = 15, 10, 4 Hz: H6'), which in turn were both correlated to the proton resonances H5 (8 1.19) and H6' (8 1.55) to the carbon resonance C7 (8 80.3) confirmed the C5/C6/C7/C8/C18 connectivity of the methyl ketone side chain. COSY correlations were observed from an allylic methylene resonance at 8 3.06 (d, J = 6 Hz: H12, H12') to the olefinic resonance at 8 5.34 (HI 1), which was further coupled to another olefinic resonance at 8 5.37 (H9), establishing the C9/C11/C12 connectivity. A long methine resonance at 8 1.19 (dd, J =6,3 Hz: H5 ) (Figure 14). HMBC correlations from the 34 range COSY correlation from the H12/H12' (5 3.06) resonance to the furan resonance at 5 6.30 (H14), combined with HMBC correlations from the H12/H12' (8 3.06) resonance to carbon resonances assigned to C9 (5 144.8), C l l (5 124.7), C13 (8 124.3), C14 (8 111.6), and C20 (8 139.6), confirmed the connectivity of the furan side chain. Figure 14. Selected isolated spin systems in botanone A (55). A homo-nuclear decoupling experiment in which the resonance at 8 3.06 (H12/H12') was selectively irradiated resulted in the transformation of the olefinic resonances at 8 5.34 (HI 1) and 8 5.37 (H9) into two clean doublets, each with a coupling constant of 16 Hz, appropriate for the E stereochemistry of the C9-C11 double bond (Figure 15). The three additional methyl singlet resonances (8 0.78 (Me 16), 0.86 (Mel7), and 0.85 (Me 19)) proved useful in establishing the connectivity of the remainder of the molecule. HMBC correlations from Mel9 (8 0.85) to the carbon resonances at 8 48.9 (C5), 144.8 (C9) and the quaternary carbon resonance at 8 40.7 (C10) established the C5/C10/C9 connectivity (Figure 16). HMBC correlations between the two geminal methyls Mel6 (8 0.78) and Mel7 (8 0.86) and carbons assigned to C3 (8 42.2), C4 (8 34.5), and C5 (8 48.9) established the C3/C4/C5 connectivity. An isolated five proton spin system in the COSY spectrum that started with coupling between the HI/HI' geminal methylene resonances at 8 1.23 and 1.31 and a resonance 35 36 at 5 1.45 (H2), which in turn was further coupled to its geminal partner at 5 1.28 (H21) and the methylene resonances at 5 1.07 (H3) and 8 1.26 (H3'), established the C1/C2/C3 connectivity in ring A. HMQC correlations between the HI/HI' resonances at 8 1.23 and 1.31 and a methylene carbon resonance at 8 40.8, between the H2/H2' resonances at 8 1.28 and 1.45 and a methylene carbon resonance at 8 19.0, and between the H3/H3' resonances at 8 1.07 and 1.26 and a methylene carbon resonance at 8 42.2 (C3), identified the three remaining methylene carbon resonances. An HMBC correlation from Mel9 (8 0.85) to the carbon resonance assigned to CI (8 40.8), combined with HMBC correlations from HI (8 1.23) and HI' (8 1.31 ) to the carbon resonances assigned to C5 (8 48.9) and C10 (8 40.7), completed the connectivity of the ring and thereby accounted for the remaining site of unsaturation in the molecule. Figure 16. Selected HMBC correlations for botanone A (55). A series of difference nOe experiments established the relative stereochemistry in botanone A (55) (Figure 17). Irradiation of 8 0.78 (Mel6) induced nOe enhancements in the resonances assigned to H2 (8 1.45) and Mel9 (8 0.85), demonstrated that H2, Mel6, and Mel9 are in axial orientations. Irradiation of 8 1.19 (H5) induced nOe enhancements in HI (8 1.31), 37 H3 (8 1.07) and H9 (8 5.37), demonstrating that HI, H3, and H5 are also in axial orientations and the CIO olefinic side-chain is in the equatorial orientation. Figure 17. Results of selected difference nOe experiments for botanone A (55) 2.2.2.2 Botanone B (56) 56 Botanone B (56) was obtained as a colourless oil that gave a molecular ion at m/z 302.2254 in the HREIMS spectrum, which corresponds to a molecular formula of C20H30O2 requiring six degrees of unsaturation. Inspection of the NMR data obtained for botanone B (56), summarized in Table 2, revealed that it was closely related to botanone A (55) and differed only 39 Table 2. NMR Data for botanone B (56) (Recorded in CDC13 at 500 MHz.) Carbon No. S 1 3 C 5 lU COSY H M B C 1 40.6 1.18m HI 1, 2, 2' H3', 19 1' 1.32 m HI, 2, 2' 2 18.7 1.55 m HI, 1', 2', 3, 3' 2' 1.45 m HI, 1', 2,3,3' 3 42.1 1.12m H2, 2', 3' H16, 17 3' 1.39 m H2, 2', 3 4 34.6 H3, 3', 16, 17 5 53.2 0.82 m H6, 6' H3', 6', 7, 9, 16, 17, 19 6 21.0 1.49 m H5,6' H7 6' 1.33 m H5,6 7 47.2 2.32 m H7 H6', 18 7' 2.36 m 8 209.1 H7, 18 9 144.8 5.27 dJ= 15 Hz H l l H l l , 12, 19 10 40.7 H9, 11, 19 11 123.7 5.38 m H9, 12 H9, 12 12 28.3 3.09dJ=6Hz H9, 11 13 124.1 H l l , 12, 14, 15,20 14 111.0 6.20 s H15 H12, 15,20 15 142.6 7.32 t J = 2 H z H14, 20 H14, 20 16 21.9 0.87 s H3, 17, 17 33.6 0.84 s H16 18 29.6 2.00 s 19 18.0 0.98 s H9 20 138.8 7.14 s H15 H12, 14, 15 in the constitution of the methyl ketone side chain. Loss of 58 mass units in the mass spectrum, and lack of an acetate methyl in the lH NMR spectrum of 56 (Figure 18), suggested that it is simply the C7 deacetoxy analogue of botanone A (55). A carbonyl stretch at 1743 cm - 1 in the 1T infrared spectrum and a carbonyl resonance at 8 209.1 (C8) in the C NMR spectrum confirmed the ketone functionality. The four sp2 hybridized carbons at 8 124.1 (C13), 111.0 (CH14), 142.6 (CH15), and 138.8 (CH20), and the three attached protons at 8 6.20 (H14), 7.32 (H15), and 7.14 (H20) were assigned to the mono-substituted furan ring. HMBC correlations between the allylic resonance at 8 3.09 (d, J = 6 Hz: H12/H12') and carbons assigned to C13 (8 124.1), C14 (8 111.0), C20 (8 138.8), C9 (8 144.8), and CI 1 (8 123.7), identified a furan side-chain identical to that in botanone A (55). An isolated spin system in the COSY spectrum, with correlations between H12/H12' (8 3.09) and an olefinic resonance at 8 5.38 (m: H l l ) , and between HI 1 (8 5.38) and an olefinic resonance at 8 5.27 (d, J =16 Hz: H9), confirmed the C9/C11/C12 connectivity. Assignment of the E configuration to the double bond was based on the H9/H11 coupling constant of 15 Hz. A comparison between the 1 3 C chemical shifts for the ring A functionality in 56 with those found in botanone A (55) revealed that the two ring systems are identical (Figure 19). 55 56 Figure 19. 1 3 C NMR chemical shifts for ring A in botanone A (55) and botanone B (56). All that remained to assign was the constitution of the ketone side chain. An HMBC correlation between the methyl singlet *H NMR resonance at 8 2.00 (Mel 8) and the 1 3 C ketone resonance at 8 209.1 (C8) identified the methyl ketone terminus of the side-chain (Figure 20). An isolated spin system in the COSY spectrum, involving a methine resonance at 8 0.82 (H5) coupled to a pair of geminal methylene proton resonances at 8 1.49 and 1.33 (H6/H6'), which in turn were both coupled to the deshielded resonance at 8 2.32 (H7), established the C5/C6/C7 connectivity (Figure 21). The deshielded chemical shift of H7 (8 2.32) suggested that it is a to a carbonyl group, which was confirmed by an HMBC correlation from the H7 resonance to the carbonyl resonance assigned to C8 (8 209.1). HMQC correlations between the H6/H6' resonances at 8 1.49 and 1.33 and a methylene carbon resonance at 8 21.0 (C6), and between the 41 H5 (8 0.82) resonance and the methine carbon resonance at 8 53.2 (C5), combined with two HMBC correlations between the H6' resonance at 8 1.33 and the carbon resonances assigned to C5 (8 53.2) and C7 (47.2), confirmed the connectivity of the methyl ketone side-chain. Figure 20 summarizes the HMBC data for botanone B (56). Figure 20. Selected HMBC correlations for botanone B (56). Figure 21. Selected isolated spin systems in botanone B (56). The relative stereochemistry of the A ring was determined from a series of nOe experiments (Figure 22). Irradiation of the methyl resonance at 8 0.98 (Me 19) induced nOe enhancements in resonances assigned to HI 1 (8 5.38) and Mel6 (8 0.87). Irradiation of the methyl resonance at 8 0.87 (Mel6) induced an nOe enhancement in Mel9 (8 0.98), indicating 42 that the A ring must be in the chair conformation with both Me 16 and Me 19 in axial orientations. Although there are no nOe data indicating that H5 (5 0.82) is axial, it was assumed that 56 would have the same relative configuration at C5 as botanone A (55). This is supported by the similarity of the C NMR shifts assigned to the A ring carbons in 55 and 56 (Figure 19). 0.82 J-J Figure 22. Results of selected difference nOe experiments for botanone B (56). 2.2.2.3 Botanone C (57) 57 Botanone C (57) was isolated as a colourless oil that gave a molecular ion at m/z 318.2204 in the HREIMS corresponding to a molecular formula of C 2 0 H 3 0 O 3 , requiring six sites of unsaturation. A number of features of the NMR data obtained for 57 (Table 3) suggested that 44 Table 3. NMR data for botanone C (57) Recordec in C 6 D 6 at 500 MHz. Carbon No. 5 1 3 C 5*H COSY HMBC 1 40.4 1.32 m HI' r 0.92 m HI, 2, 2' 2 18.9 1.27 m HI, 2', 3, 3* 2' 1.49 dt 7= 3, 10 Hz HI, 2, 3, 3' 3 41.6 1.29 m H2, 2', 3' H16, 17 3' 1.10ddJ= 13,4Hz H2, 2', 3 4 33.8 H5,6', 16, 17 5 46.4 2.12 ddJ= 18, 4 Hz H6, 6' H6', 16, 17, 18 6 35.5 2.16ddJ=6,4Hz H5, 6' 6' 2.04 ddJ = 18, 6 Hz H5, 6 7 211.8 H6, 6' 18 8 72.7 3.95 m HI8, OH H18 9 144.2 5.27 m H l l H12, 19 10 40.7 H15, 19 11 125.4 5.26 m H9, 12 H12 12 28.4 2.88dJ=5Hz H9, 11 13 124.1 H15,20 14 111.4 6.08 s H15 H20 15 143.1 7.111 J = 2 H z H14, 20 H14, 20 16 22.1 0.70 s H5, 17 17 33.5 0.65 s H16 18 20.3 1 .08dJ=7Hz H8 H5 19 18.2 0.82 s 20 139.3 7.02 s H15 H12, 14, 15 OH 3.48dJ=5Hz H8 it is closely related to the co-occurring diterpenoids botanone A (55) and botanone B (56). Al l the carbons in the 1 3 C NMR spectrum of 57 are resolved, and the HMQC experiment indicated that 29 of the 30 protons are attached to carbon atoms. A carbonyl stretch at 1711 cm - 1 in the infrared spectrum, along with a carbonyl resonance at 5 211.8 (C7) in the C NMR spectrum, confirmed the presence of a ketone functionality in the molecule. A hydroxyl group was identified by an OH stretch at 3550 cm" in the infrared spectrum. Vom sp hybridized carbons at 5 124.1 (C13), 111.4 (CH14), 143.1 (CH15), and 139.3 (CH20) in the 1 3 C NMR spectrum and the three attached protons (8 6.08 (s: H14), 7.11 (t, J = 2 Hz: H15) and 7.02 (s: H20)) were 45 assigned to a mono-substituted furan ring. HMBC correlations between the allylic proton resonance at 8 2.88 (d, J = 5 Hz: H12/H12') and the carbon resonances assigned to C13 (8 124), C14 (8 111.4), C20 (8139.3), C9 (8 144.2), and CI 1 (8 125.4) identified an olefinic side-chain with a terminal furan ring as found in botanones A (55) and B (56). A comparison of the 1 3 C chemical shifts between 57 and botanone A (55) revealed that ring A and the olefinic side-chain functionalities in 57 and 55 are identical (Figure 24). The E stereochemistry of the A 9 ' 1 1 double bond was assigned on the basis of the similarity of the chemical shifts of C9/C11 in 57 and botanone A (55). The hydroxyl, ketone, furan ring, carbon-carbon double bond, and ring A functionalities account for all three oxygen atoms and the six sites of unsaturation required by the molecular formula of 57. What remained was to identify the location of the ketone and hydroxyl functionalities on the side-chain. A methine carbon at 8 72.7 (C8), with the attached proton at 8 3.95 (m: H8), has a chemical shift characteristic of an oxygen-bearing carbon, and therefore it was assigned as the point of attachment of the hydroxyl group. An isolated spin system in the COSY spectrum, consisting of a methyl resonance at 8 1.08 (d J = 7 Hz: Mel 8) coupled to the deshielded methine H8 (8 3.95), which was further coupled to a resonance at 8 3.48 (d, J = 5 Hz: OH), established the C18/C8/OH connectivity (Figure 25). A second spin system, consisting of a methine proton resonance at 8 2.12 (dd, J = 18, 4 Hz: H5) coupled to a pair of geminal methylene proton resonances at 8 2.16 (dd, J = 6, 4 Hz: H6), and 2.04 (dd, J = 18, 6 Hz: H6') was also identified in the COSY spectrum. HMBC correlations between H6/H6' (8 2.16, 2.04) and Mel8 (8 1.08) and the carbon resonance assigned to C7 (8 211.8) established the C6/C7/C8 connectivity, and completed the assignment of the hydroxy-ketone side-chain (Figure 26). Figure 26. Selected HMBC correlations for botanone C (57). A series of difference nOe experiments established the relative stereochemistry of the A ring of botanone C (57) (Figure 27). Irradiation of the methyl resonance at 8 0.82 (Me 19) induced nOe enhancement in the resonance assigned to Mel 6 (8 0.70); and irradiation of the methyl resonance at 8 0.70 (Mel6) induced an nOe enhancement in Mel9 (8 0.82), indicating 47 that the A ring must be in the chair conformation with both Me 16 and Me 19 in axial orientations. Although there was no positive nOe data to indicate the stereochemistry at C5, it was assumed, 13 based on the similarity of the C NMR chemical shifts assigned to the A ring carbons in 55 and 57 (Figure 24), that H5 (8 2.12) in 57 has the same axial orientation as H5 in botanone A (55). 2.12 Figure 27. Results of selected difference nOe experiments for botanone C (57). 2.2.2.4 Botanicol (58) Botanicol (58) was obtained as a colourless oil that gave a molecular ion peak in the HREIMS at m/z 362.2457, appropriate for a molecular formula of C 2 2 H 3 4 O 4 and requiring 6 degrees of unsaturation. The mass spectrum showed a very weak molecular ion peak at m/z 362 with a strong peak at m/z 302, indicating facile loss of acetic acid. Table 4 summarizes the NMR 48 Table 4. NMR data for botanicol (58) Recordec in C 6 D 6 at 500 MHz. Carbon 5 1 3 C 8*H COSY H M B C No. 1 40.1 0.74 m Hl*,2, 2' H19 r 1.67 dm J= 13 Hz HI, 2' 2 18.8 1.29 m HI, 2' 2' 1.54 m HI, 2, 3' 3 42.0 1.27 dm J= 13 Hz H3' H16, 17 3' 1.05 td J= 12, 6 Hz H2', 3 4 33.4 H5, 16, 17 5 56.1 0.74dJ=3Hz H6, 6' H7', 16, 17, 19 6 20.4 1.48 m H5, 6', 7, 7' 6' 1.13qdJ=7,3Hz H5,6 7 45.4 1.77 dt J = 12,3 Hz H6, 7' HI 8, OH 7' 1.51 m H6, 7 8 73.7 HI 8, OH 9 63.4 1.58dJ=4Hz H l l H19, 18 10 39.1 H19 11 72.9 5.67 dtJ= 10, 5 Hz H9, 12, 12' H12 12 35.0 2.86 ddJ= 15, 9 Hz H l l , 12' 12* 2.80 ddJ= 15, 5 Hz H l l , 12 13 121.5 H12, 14, 15,20 14 111.8 6.17 s H15 H12, 15,20 15 143.1 7.08 t J = 2 H z H14, 16 H14, 20 16 21.6 0.72 s H H17 17 33.5 0.78 s H16 18 27.0 1.33 s 19 16.2 0.92 s H5 20 140.6 7.10 s H15 H12, 14, 15 21 169.1 H22 22 21.0 1.55 s O H 2.49 s weak data obtained for 58. Resonances for all twenty-two carbon atoms are well resolved in the C NMR spectrum, and the HMQC experiment indicated that 31 of the 32 hydrogen atoms are attached to carbon atoms. An infrared band at 1740 cm"1 indicated the presence of a carbonyl moiety in the molecule, which was identified as an ester functionality by the presence of a resonance at 8 169.1 in the 1 3 C NMR spectrum. An HMBC correlation between a methyl resonance at 8 1.55 (Me22) and the carbonyl resonance at 8 169.1 (C21) established that the ester 49 is an acetate group. The broad infrared band at 3425 cm"1 identified a hydroxyl functionality in the molecule and a 'H NMR resonance at 5 2.49 (s: OH), which was not correlated to a carbon, 1 ^ confirmed this assignment. Apparent in the C NMR spectrum of 60 are four deshielded resonances at 8 121.5 (C13), 111.8 (CH14), 143.1 (CH15), and 140.6 (CH20), characteristic of a mono-substituted furan ring. HMQC correlations between three aromatic lH NMR resonances at 8 6.17 (s: H14), 7.08 (tj = 2 Hz: H15), and 7.10 (s: H20) and carbon resonances assigned to C14 (8 111.8), C15 (8 143.1), and C20 (8 140.6), respectively, supported the presence of a furan ring in the molecule. The acetoxy, furan, and hydroxyl functionalities account for the four oxygen atoms and four of the six sites of unsaturation required by the molecular formula of 58. No additional unsaturated functionalities were evident in the 1 3 C NMR spectrum of 58; therefore two additional rings must be incorporated into the structure to account for the remaining two sites of unsaturation. The *H NMR spectrum (Figure 28) of botanicol (58), which is very well dispersed and extremely informative, contains a deshielded resonance at 8 5.67 (dt, J = 10, 5 Hz: HI 1) which is correlated to a carbon at 8 72.9 (Cl 1) in the HMQC spectrum. The deshielded chemical shift of C l 1 (8 72.9) identified it as the site of attachment of one of the oxygen atoms. The COSY and HMQC spectra identified a four proton spin system, starting at a methine resonance at 8 1.58 (d, J = 4 Hz: H9), correlated to a carbon at 8 63.4 (C9), and continuing through HI 1 (8 5.67), before terminating in a pair of geminal methylene resonances at 8 2.86 (dd, J = 15, 9 Hz: H12) and 2.80 (dd, J = 15, 5 Hz: HI2') (Figure 29). The deshielded chemical shifts of H12 (8 2.86) and H12' (8 2.80), correlated to a carbon resonance at 8 35.0 (C12), implied that they are allylic. HMBC correlations between H12/H12' and carbon resonances at 8 72.9 (Cl 1), 121.5 (C13), 111.8 (C14), and 140.6 (C20) confirmed that H12/H12' are allylic and established the 51 C9/C11/C12/C13 connectivity, indicating that the furan ring attached Xo'QXl is at the terminus of the side-chain (Figure 30). 1.58 Figure 29. Selected isolated spin systems in botanicol (58). The deshielded chemical shift of a quaternary carbon resonance at 8 73.7 (C8) identified the site of attachment of a second oxygen atom. An HMBC correlation between the hydroxyl resonance 8 2.49 (OH) and the carbon resonance at 8 73.7 (C8) indicated that the hydroxy moiety is bonded to C8 (Figure 30), requiring that the acetate group be attached to CI 1 (8 72.9), the other oxygenated carbon. Figure 30. Selected HMBC correlations from botanicol (58). 52 An isolated spin system which begins with coupling between the methine resonance at 8 0.74 (d, J = 3 Hz: H5) and a pair of geminal methylene resonances at 8 1.48 (m: H6) and 1.13 (qd, J =7,3 Hz: H6'), and terminates with coupling between H6' (8 1.13) and a second pair of geminal methylene resonances at 8 1.77 (dt, J = 12, 3 Hz: H7) and 1.52 (m: H7'), established the C5/C6/C7 connectivity (Figure 29). Three contiguous methylenes were identified by an isolated spin system in the COSY spectrum. A pair of geminal methylene proton resonances at 8 0.74 (m: HI) and 8 1.67 (dm, J = 13 Hz: HI'), correlated to a carbon at 8 40.1 (Cl), were correlated to each other and to a resonance at 8 1.29 (m: H2). H2 (8 1.29) was coupled to its geminal partner at 8 1.54 (m: H2'), which was further coupled to the resonance at 8 1.05 (td, J = 12, 6 Hz: H3'). Finally, H3' (8 1.05) was coupled to its geminal partner at 8 1.27 (dm, J= 13 Hz: H3) (Figure 29). HMBC correlations consistent with a decalin ring system were observed between Me 19 (8 0.92) and carbon resonances at 8 40.1 (Cl), 56.1 (C5), 63.4 (C9), and 39.1 (C10), between Mel6 and Mel7 (8 0.72 and 0.78) and carbons resonances at 8 42.0, (C3), 33.4 (C4), and 56.1 (C5), and between H5 (8 0.74) and carbon resonances at 8 33.4 (C4) and 16.2 (C19) (Figure 30). The combination of the above structural information allowed the total structure of botanicol (58) to be determined. A series of difference nOe experiments established the relative stereochemistry of 58 (Figure 31). Irradiation of Mel9 (8 0.92) induced nOe enhancement in resonances assigned to Mel6 (8 0.72), Mel8 (1.33) andH6' (1.13). Irradiation of Me 18 (8 1.33) induced nOe enhancement in resonances assigned to Mel9 (8 0.92), H6' (1.13), and H7 (1.77), while irradiation of the OH resonance (8 2.49) induced nOe enhancement in resonances assigned to H9 (8 1.58), Mel 8 (1.33), and Me22 (1.55). The results of the difference nOe experiments are consistent with 58 possessing a trans decalin ring system with H9, Me 16, Mel8, and Me 19 in 53 axial orientations and the C7 hydroxyl group and C9 furan containing side-chain in equatorial orientations. H5 has been assigned to the axial orientation as required by the trans decalin ring system in 58. Figure 31. Results.of selected nOe difference experiments on botanicol (58). 2.2.3 Discussion The degraded labdane carbon skeleton encountered in the botanones is quite rare in terpenoid chemistry. Chrysothame (62), isolated from the rabbit-bush Chrysothamnus nauseosus 7ft collected in Nevada, and the terpenoid glyceride esters 63 and 64, isolated from the Antarctic 77 dorid nudibranch Austrodoris kerguelensis, represent the only natural products previously isolated that possess this carbon skeleton. o 62 63 R.! = Ac, R 2 = H 64 Rx = H, R 2 = Ac The presence of the labdane diterpenoid 65 in extracts of A. kerguelensis gave some insight into the possible biogenetic origins of 63 and 64. Davies-Coleman and Faulkner demonstrated that O Q the oxidative cleavage of the A ' bond in ring B of compound 65 results in the formation of diketone 63.2 7 65 63 The proposed biogenesis of botanicol (58) and botanones A to C (55-57) from ambliofuran (20) are outlined in Scheme 1 and Scheme 2, respectively. The proton initiated cyclization of ambliofuran (20) affords the 'labdane' carbocation intermediate 66. Nucleophilic attack of the carbocation 66 by H2O, followed by loss of a proton, results in the formation of botanicol (58). Scheme 1. Proposed biogenesis of botanicol (58) from ambliofuran (20). The biogenetic path to form botanones A to C (55-57), outlined in Scheme 2, is based on 77 the observations made by Davies-Coleman and Faulkner. After ambliofuran (20) has cyclized to form intermediate 66 (Scheme 1), the C9 proton may be lost, forming a A 8 ' 9 bond (67). Oxidative cleavage of the A ' bond in 67 affords the diketone intermediate 68, which can undergo further functional group modifications to yield the botanones A to C (55-57). Scheme 2. Proposed biogenesis of botanone B (56) from the intermediate carbocation 66. 2.2.4 Isolation Artifacts. 56 2.2.4.1 Cadlinoglycine (59) 59 Cadlinoglycine (59) was isolated as a yellowy oil that exhibits a molecular ion at m/z 389.2208 in the HREIMS, which corresponds to a molecular formula of C22H31NO5, requiring 8 degrees of unsaturation. Table 5 provides a summary of the NMR data collected for compound 59. The twenty-two carbon resonances were identified by analysis of the C NMR and HMQC spectra. The HMQC spectrum indicates that all 31 hydrogen atoms are attached to carbon. Infrared bands at 1674, 1725, and 1777 cm"1 implied that there are three carbonyl groups in the molecule, one of which is a carboxylic acid, as indicated by a broad band from 2500-3300 cm"1. 1 "X Three deshielded resonances in the , J C NMR spectrum at 8 172.4, 175.1, and 169.7 confirmed the presence of three carbonyl functionalities in 59. The frequency of the carbonyl stretching vibration at 1777 cm"1 indicated that one of the carbonyls is a y-lactone. A deshielded carbon resonance at 8 88.3 (C15) in the C NMR spectrum, attached to a proton at 8 5.65 (d, / = 5 Hz: HI 5), is characteristic of an aminal carbon (a carbon attached to an oxygen and a nitrogen atom). The only oxygen atom available to be bonded to the sp3 hybridized aminal carbon is the ester alkoxy oxygen; therefore it and the nitrogen atom must be attached to the same carbon. Since four oxygen atoms have been accounted for in the acid and ester functionalities, the remaining carbonyl must be present as an amide to account for the shielded chemical shift of all of the 57 Table 5. NMR data for cadlinoglycine (59) Recorded in DMSO-flf6 at 500 MHz. Carbon 5 B C 5 COSY H M B C No. 1 38.1 1.11 m HI', 2, 2' H5,5', 19 r 1.89 m HI, 2, 2' 2 20.6 1.47 m H1,T, 2,3,3' HI 2* 1.47 m HI, 1', 2', 3, 3' 3 39.1 1.13 m H2, 2', 3' H16, 17 3' 1.22 m H2, 2', 3 4 32.2 H5, 16, 17 5 49.7 1.11 d J = 14 Hz H5' H16, 17, 19 5' 1.76dJ = 14 Hz H5 6 18.1 1.32dJ = 7 Hz H7 H7 7 40.3 3.87qJ = 7 Hz H6 H6 8 122.3 H6, 7, 11', 14, 15 9 143.1 H5, 5', 7, 11, 11', 12', 14, 19 10 39.8 H5, 5', 19 11 23.0 1.90 m H l l ' , 12, 12' 11* 2.26 ddJ-= 13,2, Hz H l l , 12, 12' 12 19.7 1.48 m H l l , 11', 12', 13 12' 1.99 dm J = 14 Hz H l l , 11', 12, 13 13 39.2 3.25 m H12, 12', 14 14 35.6 3.53 m H13, 15 H7 15 88.3 5.65 d J = 5 Hz H14 H21,21' 16 27.7 0.74 s H5, 17 17 31.3 0.86 s H5, 5', 16 18 172.4 H6, 7, 15,21,21* 19 31.8 1.02 s 20 175.1 H13, 14 21 46.4 3.76 dJ = 17 Hz H21' 21' 4.26 dJ = 17 Hz H21 22 169.7 H21,21' carbonyl resonances in the C NMR. The final unsaturated functionality, which was apparent 1 T from the C NMR data, is a tetra-substituted double bond, identified by two quaternary olefinic resonances at 5 122.3 and 143.1 in the 1 3 C NMR spectrum. Al l five oxygen atoms and one nitrogen atom, as well as four sites of unsaturation have been accounted for; therefore, three additional rings must be incorporated into the structure of cadlinoglycine (59) in order to satisfy the degrees of unsaturation required by the molecular formula. 58 The deshielded resonance at 8 5.65 (HI5) in the ! H NMR spectrum of 59 (Figure 32) is the start of a seven proton spin system, as determined by interpretation of the COSY spectrum, which continued through two contiguous methine resonances at 8 3.53 (m: H14) and 3.25 (m: HI 3) before terminating in a pair of adjacent methylene resonances at 8 1.48 (m: HI 2) and 1.99 (dm,J= 14Hz:H12'),and8 1.90 (m: H l l ) and 2.26 (dd,7= 13, 2 Hz: HI 1') (Figure 33). The deshielded chemical shift of the HI 1/H11' (8 1.90, 2.26) resonances, which were correlated to a carbon at 8 23.0 (CI 1), implied that they are allylic. Strong HMBC correlations between HI 1/H11' (8 1.90, 2.26) and the olefinic carbon resonances at 8 122.3 (C8) and 143.1 (C9), combined with HMBC correlations between the proton resonance at 8 3.53 (HI4) and the olefinic carbon resonances at 8 122.3 (C8) and 143.1 (C9), established that CI 1 and C14 are connected by the tetra-substituted olefin, thus establishing the C8/C9/C11/C12/C13/C14 connectivity and forming a cyclohexene ring (Figure 34). A second spin system, identified in the COSY spectrum, involving a methyl resonance at 8 1.32 (d, J = 1 Hz: Me6) coupled to a deshielded methine resonance at 8 3.87 (q, J = 7 Hz: H7), established the C6/C7 connectivity (Figure 33). The chemical shift of H7 (8 3.87), which is correlated to a carbon at 8 40.3 (C7), is characteristic of a methine situated between a double bond and a carbonyl. Figure 33. Selected isolated spin systems in cadlinoglycine (59). 59 60 An isolated methylene was identified by the COSY correlations between a pair of deshielded resonances at 6 3.76 (d, J = 17 Hz: H21) and 4.26 (d, J = 17 Hz: H21') (Figure 32). These proton resonances are both correlated to a carbon resonance at 5 46.4 (C21) in the HMQC spectrum. The chemical shifts of H21/H21' (5 3.76, 4.26) are characteristic of the cc-protons of glycine. A Figure 34. Selected HMBC correlations for sub-structure A of cadlinoglycine (59). HMBC correlations that establish the C18/C7/C8/C14/C15 connectivity are observed between Me6 (8 1.32) and the carbon resonances assigned to C8 (8 122.3) and C18 (8 172.4), between H7 (8 3.87) and the carbon resonances assigned to C8 (8 122.3), C9 (8 143.1), and C18 (8 172.4), and finally between H15 (8 5.65) and the carbon resonances assigned to C8 (8 122.3) and C18 (8 172.4) (Figure 34). HMBC correlations between the two resonances for H21/H21' (8 3.76, 4.26) and the carbon resonances at 8 88.3 (C15), 172.4 (C18), and 169.7 (C22) indicated that the nitrogen atom of the glycine residue is attached to the carbonyl carbon C l 8, thus completing the ring to form a 8-lactam (Figure 34). The position of the remaining carbonyl was determined by HMBC correlations between the proton resonances H13 (8 3.25) and H14 (8 3.53) 61 and the carbon resonance assigned to C20 (8 175.1), indicating that the remaining carbonyl carbon is attached to C13 (Figure 34). The two possible tricyclic sub-structures B and C, shown in Figure 35, both satisfy the HMBC data summarized in Figure 34. Unfortunately, there are no HMBC correlations between HI5 (8 5.65) and the lactone carbonyl group, making it impossible to differentiate between these two sub-structures. To aid in distinguishing between the two possible structures, cadlinoglycine (59) was converted to the methyl ester (72) with diazomethane, and an additional HMBC experiment was performed with 72. The appearance of a new methyl ester resonance at 8 3.22 (s: Me23) in the ' H NMR spectrum of 72 (Figure 36), combined with an increase of 14 mass units in the molecular ion in the mass spectrum of 72, indicated that the reaction proceeded smoothly to yield the methyl ester. An HMBC correlation between Me23 (8 3.22) and the carbonyl carbon resonance at 8 169.2 (C22) established the position of the methyl ester as attached to C22 (8 169.2) (expansion shown in Figure 37). HMBC correlations between the geminal methylene resonances at 8 4.66 (H21) and 3.42 (H2T) and the carbon resonances assigned to C20 (8 175.1) and C22 (8 169.2) indicate that fragment D is the correct sub-structure. Therefore, the alkoxy oxygen involved in the y-lactone must also be attached to C20 (8 175.1). The tri-cyclic lactone-lactam fragment of 59 resembles the bis-lactone fragment in the known compounds aplysulphurin (18), tetrahydroaplysulphurin-1 (21) and cadlinolide A (47), differing only in that the 8-lactone alkoxy oxygen is replaced by the nitrogen of glycine to form a 8-lactam. The hydrocarbon segment of 59, which remains to be identified, contains three methyl groups, four methylenes, two quaternary carbons, and one site of unsaturation. Comparison of the NMR data for ring A in cadlinoglycine (59) and the NMR data for ring A in cadlinolide A 63 19 j 1 1 / 9 ^ !?••• 4"''., 17 H21 H 14 o 2^0 \ o o 8 H l 5 | " H 7 1 8 / N ^ 2 1 ^ V 2 3 O 72 H7 H21' JUL Me23 -, 1 1 1 1 1 1 1 1 1 1 1 1 | i | r-(ppm) 4.4 4.2 4.0 3.8 3.6 3.4 3.2 F- 168 169 F- n o 171 172 F- 173 h 174 Figure 37. Expansion of H M B C NMR spectrum of cadlinoglycine methyl ester (72). Recorded in C 6 D 6 at 500 MHz.. 64 (36), revealed that the two molecules have identical A ring systems, and led to the overall structure of cadlinoglycine (59) (Figure 38). COSY and HMBC data supported this assignment. D E Figure 35. Selected HMBC correlations from cadlinoglycine methyl ester (72). 59 47 Figure 38. C NMR chemical shifts for ring A of cadlinoglycine (59) and cadlinolide A (47). Three contiguous methylenes were identified by a six proton spin system in the COSY spectrum, starting with the resonances at 5 1.11 and 1.89 (m: HI; m: HI') and continuing uninterrupted through 5 1.47 (m: H2, H2'), before terminating at 8 1.13 and 1.33 (m: H3; m: H3') 65 (Figure 39). A second isolated methylene was identified in the COSY spectrum by a spin system involving two geminal resonances at 8 1.11 and 1.76 (d, J = 14 Hz: H5; d, J = 14 Hz: H5'), which were attached to the carbon at 8 49.7 (C5) (Figure 39). Figure 39. Selected isolated spin systems in the A ring of cadlinoglycine (59). HMBC correlations consistent with the proposed structure were observed between Me 19 (8 1.02) and the carbon resonances assigned to CI (8 38.1), C5 (8 49.7), C9 (8 143.1), and CIO (8 39.8), between two geminal methyl resonances at 8 0.74 (Mel6) and 0.86 (Mel7) and the carbon resonances assigned to C3 (8 39.1), C4 (8 32.2), and C5 (8 49.7), and between H5 (8 1.11) and H5' (8 1.76) and the carbon resonances assigned to CI (8 38.1), C4 (8 32.2), C9 (8 143.1), and CIO (8 39.8) (Figure 40). The relative stereochemistry of the tri-cyclic portion of cadlinoglycine (59) was established by a series of difference nOe experiments. Irradiation of H15. (8 5.65) induced nOe enhancements in the proton resonances assigned to H14 (8 3.53) and Me6 (8 1.32), and irradiation of H14 (8 3.53) induced nOe enhancements in the proton resonances assigned Me6 (8 1.32) and HI5 (8 5.65), establishing that Me6, H14, and HI5 are all cis with respect to each other, as in cadlinolide A (47) (Figure 41). The assignment of H13 (8 3.25) as cis with respect to HI4 (8 3.53) in 59 was based on the assumption that the stereochemistry at C13 would be the 66 same as that found in cadlinolide A (47) and the tetrahydroaplysulphurins 1 to 3 (21-23). Unfortunately, no nOe data were obtained concerning the relative stereochemistry of the A ring in cadlinoglycine (59); but, based on the similarity of the I 3 C NMR chemical shifts assigned to the A ring of 59 and those of the corresponding carbons in cadlinolide A (47) (Figure 38), it was assumed that the A ring of 59 has the same stereochemistry as the A ring of 47. Figure 40. Selected HMBC correlations for ring A of cadlinoglycine (59). 59 Figure 41. Selected nOe enhancements observed for cadlinoglycine (59). 67 2.2.4.2 Cadlinolide C methyl acetal (60) O 20 o . 22 OMe X: 16 17 0 60 Cadlinolide C methyl acetal (60) was isolated as a colourless oil, exhibiting a parent ion in the HREIMS at m/z 378.2402 corresponding to a molecular formula of C22H34O5, which requirs six units of unsaturation. All twenty-two carbon resonances were identified by analysis of the 1 3 C NMR and HMQC spectra, and the HMQC experiment indicated that all 34 hydrogen atoms are attached to carbon. Infrared bands at 1771 and 1733 cm"1 indicated the presence of two ester functionalities, which was supported by resonances in the 1 3 C NMR spectrum at 8 174.2 and 179.2, accounting for four oxygen atoms. The carbonyl stretching frequency at 1771 cm"1 indicated that one of the ester functionalities is present as a y-lactone. An HMBC correlation between a methyl singlet resonance at 8 3.68 and the carbon resonance at 8 174.2, identified the other carbonyl group (1733 cm"1) as a methyl ester. A second methyl singlet resonance at 8 3.40, attached to a carbon at 8 56.9, identified a methyl ether and accounted for 13 the remaining oxygen atom. Two quaternary carbon resonances in the C NMR spectrum at 8 127.4 and 146.8 identified a tetra-substituted double bond, which is the last unsaturated functionality apparent in the molecule. Therefore, three rings must be incorporated into the structure of cadlinolide C methyl acetal (60), in order to account for the remaining sites of unsaturation required by the molecular formula. The ! H NMR spectrum of 60 (Figure 42), which is very well dispersed, contains a deshielded methine resonance at 8 4.99 (d, J = 4 Hz: HI5), which is coupled to a ketal carbon 68 Table 6. NMR data for cadlinolide C methyl acetal (60) Recorded in CDC13 at 500 MHz. Carbon No. 5 1 3 C 5*11 C O S Y H M B C 1 39.0 1.27 m HI', 2, 2' H5, 19 r 2.14dmJ=6Hz HI, 2, 2' 2 19.8 1.50 m HI, 1', 2,3,3' 2' 1.86 qmJ= 11 Hz HI, 1', 2', 3, 3' 3 39.9 1.18m H2, 2', 3' HI', 16, 17 3' 1.86 qmJ= 11 Hz H2, 2', 3 4 31.2 H5, 5', 16, 17 5 50.6 0.99 d J = 14 Hz H5' H2', 16, 17, 19 5' 1.75 d7= 14 Hz H5 6 16.4 1.18dJ=7Hz H7 H7, 18,21 7 41.8 4.25qJ=7Hz H6 H6 8 127.4 H6, 7, 14, 15 9 146.8 H5, 7, 11', 14 10 42.0 H5, 5', 19 11 26.6 1.49 m HIT, 12, 12' 11* 2.28 ddJ= 16, 2 Hz H l l , 12, 12' 12 27.2 1.30 m H l l , 11', 12', 13 H13 12' 2.04 dt J= 13, 3 Hz H l l , 11', 12, 13 13 40.6 2.87tm7=8Hz H12, 12', 14 H14 14 44.0 2.94 ddJ= 10, 3 Hz H13, 15 H7 15 110.2 4.99dJ = 4Hz H14 H14, 22 16 26.9 0.85 s H5, 17 17 32.8 0.86 s H5, 16 18 174.2 H6, 7, 21 19 30.4 1.04 s 20 179.2 H13, 15 21 51.8 3.68 s 22 56.9 3.40 s H15 at 5 110.2 (C15) in the HMQC spectrum. The C l 1/C12/C13/C14/C15 connectivity was established by a seven-proton spin system identified in the COSY spectrum, which starts at the ketal methine resonance at 5 4.99 (HI 5) and continues uninterrupted through two methine resonances at 8 2.94 (dd, J = 10, 3 Hz: H14) and 2.87 (tm, J=8 Hz: HI 3), before ending at a pair of adjacent methylene resonances at 8 1.30 (m: H12) and 2.04 (dt, J = 13,3 Hz: H12') and at 8 1.49 (m: HI 1) and 2.28 (dd, J = 16, 2 Hz: HIT) (Figure 43). HMBC correlations observed between HI 1' (8 2.28) and the carbon assigned to C9 (8 146.8), and between H14 (8 2.94) and s 69 (D <L> CN CN <0 (N N o o 13 O Q U <D O o rt o +-» I CD a u T3 SO O 1 o o o s 70 the carbons assigned to C8 (127.4) and C9 (8 146.8), indicated that C l 1 and C14 are connected by the tetrasubstituted double bond, thus assigning the C8/C9/C11/C12/C13/C14 cyclohexene ring (Figure 44). Figure 43. Isolated spin systems from the COSY spectrum of cadlinolide C methyl acetal (60). A second spin system, involving a methyl resonance at 8 1.18 (d, J = 7 Hz: Me6) coupled to a deshielded methine resonance at 8 4.25 (q, J = 7 Hz: H7), established the C6/C7 connectivity (Figure 43). The deshielded chemical shift of H7 (8 4.25) is characteristic of a methine situated between a double bond and a carbonyl group. Both this spin system and the seven-proton spin system described above resemble the systems found in cadlinoglycine (59) and cadlinolides A (36), B (37), and C (38). Further similarities between 60 and cadlinolide C (38) were established by a series of HMBC correlations. An HMBC correlation between Me6 (8 1.18) and the carbons assigned to C7 (8 41.8), C8 (8 127.4), and C18 (8 174.2) identified the allylic methyl ester side chain, which was also present in 49. HMBC correlations consistent with the presence of the y-lactone are observed between HI3 (8 2.87) and the carbon assigned to C20 (8 179.2), between HI4 (8 2.94) and the carbons assigned to C13 (8 40.6) and C15 (8 110.2), and between H15 (8 4.99) and the carbons assigned to C20 (8 179.2) and C22 (8 56.9) (Figure 44). HMBC correlations between HI5 (8 4.99) and 71 the methyl ether carbon C22 (5 56.9), and between Me22 (5 3.40) and the ketal carbon (5 110.2), established that the alkoxy oxygen of the lactone and the oxygen of the ether are attached to the ketal carbon (Figure 44). Comparison between the remaining NMR data for 51 and the NMR data for cadlinolide C (49) indicated that the A ring portions of these molecules are identical, and that 60 is simply the methyl acetal of 49. Figure 44. Selected HMBC correlations for cadlinolide C methyl acetal (60). A series of difference nOe experiments established that the relative stereochemistry of 60 is the same as cadlinolide A (47) (Figure 45). Irradiation of HI 5 (5 4.99) induced nOe enhancements in the proton resonances assigned to H14 (8 2.94), Me21 (8 3.68) and Me22 (8 3.40), irradiation of HI4 (8 2.94) induced nOe enhancements in the proton resonances assigned to Me6 (8 1.18), H13 (8 2.87), H15 (8 4.99), and Me22 (8 3.40), irradiation of Me6 (8 1.18) induced nOe enhancements in the proton resonances assigned to H7 (8 4.25) and HI4 (8 2.94), and irradiation of H7 (8 4.25) induced nOe enhancements in the proton resonances assigned Me6 (8 1.18) and Mel9 (8 1.04). This data established that Me6, H13, H14, and H15 all have a cis relationship with respect to each other, similar to cadlinolide A (47). Figure 45. nOe Enhancements observed for cadlinolide C methyl acetal (60). 73 2.2.5 Discussion The unique structure of cadlinoglycine (59) combines both terpenoid and amino acid chemistry. Normal terpenoid biosynthesis does not explain the origin of this interesting compound, and the question of whether 59 is a natural product or an artifact of the isolation procedure must be addressed. In an attempt to answer this question, and the question of the origin of cadlinolide C methyl acetal (60), several experiments were performed. We believed it was possibe that glycine could react with the y-lactone of cadlinolide A (36) to form cadlinoglycine (59), although this reaction sequence seems unlikely from a mechanistic point of view. Nonetheless, the reaction of 47 with glycine was attempted. Cadlinolide A (47), readily available as one of the major metabolites of Aplysilla glacialis, was reacted with glycine under neutral, basic, and acidic conditions for time periods varying from one hour to 7 days. The result in each case was a 100% recovery of unreacted 47, indicating that cadlinoglycine (59) is likely not an artifact of cadlinolide A (47), as expected. Recollection of the Aplysilla sp. sponge, followed by an expeditious extraction and isolation of the terpenoids, was performed in an attempt to determine if 59 is an isolation artifact. It was hoped that a fast extraction and work-up would minimize the opportunity for artifact formation, and this was found to be the case. Work-up of the recollected Aplysilla sp. sponge produced diterpenoids 55 to 58, but cadlinoglycine (59) and cadlinolide C methyl acetal (60) were not present in the extract. However, there was *H NMR evidence for compound 61 (Figure 46), which was believed to be the free acid of cadlinolide C (49). Unfortunately, while it was dissolved in CDCI3, 61 spontaneously cyclized to yield cadlinolide A (47). The observation that 61 is so reactive, combined with the fact that it was not present in the first extract of Aplysilla sp., make it the likely precursor to cadlinolide A (47), cadlinolide C (49), cadlinoglycine (59), and cadlinolide C methyl acetal (60). Scheme 3 outlines a proposed reaction sequence for the formation of cadlinolide C (49) and cadlinolide C methyl acetal (60) from compound 61. Formation of cadlinolide C (49) can occur via the acid-catalyzed esterification of 61 with the methanol solvent. Cadlinolide C (49) can then undergo acid-catalyzed loss of water to form the oxonium ion 73. Nucleophilic attack of methanol on 73 followed by loss of a proton, yields cadlinolide C methyl acetal (60). Scheme 3. Proposed formation of cadlinolide C (49) and cadlinolide C methyl acetal (60) from 61. The formation of cadlinoglycine (59) may also proceed through the oxonium ion intermediate 73 (Scheme 4). Nucleophilic attack of oxonium ion 73 by glycine, followed by a proton transfer, yields intermediate 75. Nucleophilic attack on the methyl ester carbon of 75 by the amino nitrogen, followed by loss of methanol and a proton, yields the desired product 59. 76 The origin of glycine residue in 59 is presently unknown. Do sponges have free amino acids just "floating" around, or was there some sort of contamination in the original sample? In order to determine the source of the glycine and test the hypotheses described above, additional collections of the Aplysilla sp. sponge will have to be made. 59 75 Scheme 4. Proposed formation of cadlinoglycine (59) from 62. 2.3 CONCLUSIONS 77 The study of a pink sponge, initially identified as Aplysilla glacialis, resulted in the isolation of six new compounds: botanones A to C (55-57), botanicol (58), and the artifacts cadlinoglycine (59) and cadlinolide C methyl acetal (60). Identification of the 'labdane' diterpenoid chemistry of this sponge indicated it is not Aplysilla glacialis, and reexamination of the sponge in light of its secondary metabolites revealed that it is a previously undescribed 99 Aplysilla species. Botanones A to C (55-57) represent the first examples of diterpenoids with the degraded-'labdane' skeleton 17 reported from a sponge source. The biogenesis of the botanone carbon skeleton is thought to proceed via the oxidative cleavage of the A ' bond of the hypothetical labdane intermediate 67. Cadlinoglycine (59) and cadlinolide C methyl acetal (60) are both thought to be isolation artifacts resulting from the sample of sponge sitting in methanol for one and half years. An explanation for the formation of these two compounds starting from 61 has been proposed. Esterification of 61 yields cadlinolide C (49), which could either react with methanol to yield 60 (Scheme 3) or react with glycine to yield cadlinoglycine (59) (Scheme 4). The stability of cadlinolide A (47), combined with evidence for the existence of 61, suggests that cadlinolide C (49) is likely an artifact of 61. In order to determine if the reaction pathways outlined in Schemes 3 and 4 are plausible, the Aplysilla sp. will have to be recollected and the isolation procedure modified such that 61 can be isolated and studied. In addition to the new compounds, four known diterpenoids were also isolated from the extracts of Aplysilla sp.: ambliofuran (20), marginatafuran (36), hexahydroambliofuran (37), and 9,11-dihydrogracillin A (54). Marginatafuran (25) and 9,11-dihydrogracillin A (43) are known metabolites, which have been isolated repeatedly from the skin extracts of the dorid nudibranch C. leutomarginata collected off the coast of North America. 1 5 ' 2 3 ' 2 4 This is the first time that a sponge source for these compounds has been found.24 It seems reasonable to suggest that the Aplysilla sp. is the dietary source of these compounds for the nudibranch, although there is no direct evidence to support this hypothesis. It is curious that the second batch of Aplysilla sp. collected yielded three additional metabolites. That repeated collections of the sponge resulted in a modification of the sponge's secondary metabolites is a phenomenon which calls for further investigation, also involving repeated collections of the sponge. 2.4 EXPERIMENTAL 79 2.4.1 General All reactions involving air or moisture sensitive reagents were performed in oven dried glassware which was assembled while hot and cooled under a positive pressure of argon. Reaction temperatures cited refer to external cooling or heating bath temperatures, unless otherwise noted. Al l chemicals were used as received unless otherwise noted. Solvents were dried by distillation from the appropriate drying agent. Tetrahydrofuran and diethyl ether were refluxed over, then distilled from sodium benzophenone ketyl. Dichloromethane, triethylamine, and diisopropylamine were distilled from calcium hydride. Petroleum ether refers to a hydrocarbon mixture with a boiling range of 30-60°C. Proton nuclear magnetic resonance (*H NMR) spectra were recorded on either a Bruker model AC-200 (200 MHz), WH-400 (400 MHz), AM-400 (400 MHz), or AMX-500 (500 MHz) 1 ^ spectrometer. Carbon nuclear magnetic resonance ( C NMR) spectra were recorded on a Bruker AC-200 (50.3), AM-400 (100.6 MHz), or AMX-500 (125.8 MHz) spectrometer. The chemical shifts are reported in ppm downfield from the tetramethylsilane resonance with the solvent residual peaks as the references ( !H: DMSO-d 6 2.49 ppm, CD 3OD 2.49 ppm, CDC13 7.24 ppm, C 6 D 6 7.15 ppm; 1 3 C: DMSO-c/6 39.5 ppm, CD 3OD 49.0 ppm, CDC13 77.0 ppm, C 6 D 6 128.0 ppm). Coupling constants (J) are reported in hertz (Hz). Spectra were processed using both Bruker UXNMR version 930601.3 software and Bruker Windows™ compatible WIN-NMR software. 80 Infrared spectra were recorded as either thin films or as chloroform solutions on a Perkin-Elmer 1710 Fourier Transform Spectrophotometer with internal calibration. Only the strongest and/or most diagnostic bands are reported. Low and high resolution electron impact mass spectra (LREIMS and HREIMS) were recorded on a Kratos MS50/DS55SM mass spectrometer (70 eV). The low and high resolution FAB mass spectra (LRFABMS and HRFABMS) were recorded on a Kratos Concept II HQ mass spectrometer. Elemental analyses were performed on a Carlo Erba CHN elemental analyzer, Model 1106, by the UBC microanalytical laboratory. Optical rotations were measured on either a JASCO J-700 spectropolarimeter (1 cm quartz cell) or a Perkin Elmer 241 polarimeter (1 cm quartz cell) at the sodium D line (589 nm) and the values are given in 10"Idegcm2g"1. Normal and reversed phase thin layer chromatography (TLC) were carried out on commercial aluminum-backed Kieselgel 60 F254 (E. Merck, type 5554, 0.2 mm) and Whatman MKC18F reversed phase TLC plates, respectively. TLC spots were visualized by either UV light (254 nm), iodine, and/or heating the chromatograms after staining with a solution of phosphomolybdic acid (PMA) in EtOH (20% w/v, Aldrich) or a solution of vanillin in a sulfuric acid-EtOH mixture (6% vanillin w/v, 4% sulfuric acid, and 10% water v/v in EtOH). Flash chromatography was performed on 230-400 mesh silica gel (E. Merck, Silica Gel 60). TLC grade silica 10-40 p, type H silica (S-6628, Sigma) was used for some column chromatography. Gel permeation chromatography was performed using Sephadex LH-20 resin. The term in vacuo is used to describe removal of solvent via a Buchi rotary evaporator. HPLC separations were performed on one of three systems: i) Waters 501 HPLC pump equipped with a Waters 440 absorbance detector and a Perkin-Elmer LC-25 refractive index detector; ii) Waters 600E HPLC pump/system controller with a Waters 486 tunable absorbance detector; and iii) Waters 600E HPLC pump/system controller with a Waters 996 photodiode array detector. System i) was run with a standard chart recorder, while systems ii) and iii) could 81 alternately use a chart recorder or could be interfaced with a personal computer using the Millenium 2021 chromatography software. Normal phase HPLC separations used either a Waters Rad-Pak silica 10p, column or an Econosil Si 5p column. Reversed phase HPLC separations were performed on either a Whatman Partisil 10 ODS-3 magnum column, a Waters Rad-Pak C18 lOp column, or an Econosil C18 5p column. Al l HPLC solvents were Fisher HPLC grade, and were filtered and degassed prior to use. 2.4.2 Mater ia ls . May 1991: Aplysilla sp. was collected by hand (60 g wet weight) from the high intertidal zone at Botanical Beach on Vancouver Island, British Columbia. Freshly collected sponge was immersed in methanol and stored at 4°C for one and a half years. Removal of the methanol by filtration, followed by repeated extraction with methanol/CHaCb (1:1, 2 x 250 mL) and concentration in vacuo of the combined organic extracts yielded a green oil (3.8 g). Sequential partitioning of the oil between hexanes and saturated sodium chloride solution, then ethyl acetate and saturated sodium chloride solution, followed by drying of the organic layers over anhydrous Na2SC»4, and concentration in vacuo afforded a non-polar hexane fraction (1.2 g) and a polar ethyl acetate fraction (0.74 g). Further fractionation of the hexane fraction by silica gel flash chromatography and the ethyl acetate fraction by reversed phase flash chromatography yielded several fractions containing mixtures of fats, pigments, steroids, and terpenoids as detected by analytical TLC and *H NMR analysis. HPLC purification of terpenoid fractions yielded the known diterpenoid marginatafuran ( 3 6 ) , the novel diterpenoids botanones A to C ( 5 5 - 5 7 ) , botanicol ( 5 8 ) , cadlinoglycine ( 5 9 ) , and cadlinolide C methyl ester ( 6 0 ) . September 1994: Aplysilla sp. was collected by hand (50 g wet weight) from the high intertidal zone at Botanical Beach on Vancouver Island, British Columbia. Freshly collected 82 sponge was immediately immersed in methanol where it remained for two days, after which time the methanol was decanted off and the sponge was re-extracted with a solution of methanol/CFkCk (1:1, 2 x 250 mL). After filtration and concentration of the organic layers in vacuo, progression through the same isolation procedure described above yielded ambliofuran (20), marginatafuran (36), hexahydroambliofuran (37), 9,11-dihydrogracillin A (54), botanones A to C (55-57), and botanicol (58). There was also ! H NMR evidence for compound 61. 2.4.2.1 Ambliofuran (20). 20 Compound 20 was isolated as a colourless oil (4 mg); *H NMR (CDC13, 500 MHz, 8): 1.59 (br s, 9H), 1.65 (s, 3H), 2.17 (q, J = 7 Hz, 6H), 2.42 (t, J = 7 Hz, 6H), 5.03 (m, 2H), 5.11 (t, J = 1 Hz, IH), 6.17 (br s, IH), 7.12 (br s, IH), 7.24 (br s, IH); exact mass calculated for C 2oH 3 00 (M+): 286.2297; found: 286.2286. 2.4.2.2 Hexahydroambliofuran (37). Compound 37 was isolated as a colourless oil (4.5 mg); XW NMR (CDC13, 500 MHz, 8): 0.86 (d, J = 7 Hz, 3H), 0.87 (d,J=7 Hz, 9H), 1.00-1.30 (m, 16H), 1.47 (m, 3H), 2.37 (t, J = 7.5,2H Hz), 6.25 (br s, IH), 7.20 (br s, IH), 7.33 (t, J = 1.6 Hz, IH); exact mass calculated for C 2 0 H 3 6 O (M+): 292.2766; found: 292.2764. 83 2.4.2.3 Marginatafuran (36). 36 Compound 36 was isolated as a white solid (8 mg); [ct]D25 = -87° (c 0.8, CHC1 3 ) ; *H N M R ( C 6 D 6 , 500 M H z , 6): 0.64 (dd, J = 13, 4 Hz , 2H), 0.80 (dd, J = 8, 4 Hz , 1H), 0.82 (s, 3H), 0.98-1.01 (m, 1H), 1.02 (s, 3H), 1.28-1.31 (m, 1H), 1.31 (s, 3H), 1.34-1.37 (m, 1H), 1.29-1.43 (m, 3H), 1.54 (dm, J= 11 Hz , 1H), 1.74 (ddd, J= 13, 12, 6 Hz , 1H),2.06 (dd,J= 14, 6 Hz , 1H), 2.21 (ddd, J= 16, 12, 6 Hz , 1H), 2.32-2.39 (m, 3H), 2.53 (d, J= 13 H z , 1H), 5.99 (d, J= 2 Hz , 1H), 7.07 (d, J = 2 Hz , 1H); 1 3 C N M R ( C 6 D 6 , 125 M H z , 8): 19.1, 20.0, 20.5, 20.9, 23.0, 23.3, 33.7, 33.8, 37.3, 37.4, 38.6, 42.5, 48.6, 56.0, 56.5, 110.2, 113.9, 140.5, 159.5, 182.4; exact mass calculated for C20H28O3 ( M + ) : 316.2038; found: 316.2038. 2.4.2.4 9,11-dihydrogracallin A (54). Compound 54 was isolated as a white solid (2 mg); lU N M R (CDC1 3 , 400 M H z , 8): 0.89 (s, 3H), (0.96 s, 3H), 1.03 (s, 3H), 1.16 (d, 1H, J= 14 Hz), 1.65 (d, 3H , J= 7 Hz) , 1.85 (m, 1H), 2.08 (2, 3H), 2.10 (s, 3H), 2.36 (ddd, 1H, J =9, 7.7, 7.5 Hz) , 2.42 (dd, 1H, J = 6.8, 4.2 Hz), 3.14 (dd, 1H, J = 7.5, 5.5 Hz) , 5.65 (q, 1H, J = 7 Hz) , 5.97 (s, 1H), 6.44 (d, 1H, J = 5.5 Hz); exact mass calculated for C i 9 H 2 8 0 ( M + - 2AcOH) : 272.2140; found: 272.2122. OAc 54 2.4.2.5 Botanone A (55). 84 Compound 55 was isolated as a colourless oil (12 mg); [OC]D25 = -37° (c 1.1, CHCI3); IR (neat): 2972, 1770, 1733, 1377, 1244, 1230 cm"1; see Table 1 for NMR data; exact mass calculated for C22H32O4 (M+): 360.2300; found: 360.2306. 2.4.2.6 Botanone B (56). O 56 Compound 56 was isolated as a colourless oil (2 mg); IR (CHCI3): 2962, 1735, 1651 cm"1; see Table 2 for NMR data; exact mass calculated for C20H30O2 (M+): 302.2246; found: 302.2254. 2.4.2.7 Botanone C (57). OH 57 Compound 57 was isolated as colourless oil (1 mg); IR (CHC13): 3450, 2960, 1710, 1602, 1462 cm"1; see Table 3 for NMR data; exact mass calculated for C 2 0 H 3 0 O 3 (M+): 318.2195; found: 318.2165. 2.4.2.8 Botanicol (58). AcO 58 Compound 58 was isolated as a colourless oil (8 mg); [a] D 2 5 = -7.2° (c 0.5, CHC13); IR (neat) 3420, 2975, 1733, 1239 cm"1; see Table 4 for NMR data; exact mass calculated for C22H32O3 (M+): 344.2351; found: 344.2353. 2.4.2.9 Cadlinogycine (59). o OH 59 Compound 59 was isolated as a colourless oil (6 mg); IR (CHC13): 3300-2400, 2953, 1777, 1725, 1674 cm"1; see Table 5 for NMR data; exact mass calculated for C 2 2 H 3 i N i 0 5 (M+): 389.2202; found: 389.2208. 2.4.2.10 Cadlinolide C methyl acetal (60). OMe OMe 60 86 Compound 60 was isolated as a colourless o i l (6 mg); [ a ] D = -74.3° (c 0.14, CHC1 3 ) ; IR (CHC13): 2953, 1771, 1732, 1220 cm" 1; see Table 6 for N M R data; exact mass calculated for C22H34O5 ( M + ) : 378.2406; found: 378.2402. 2.4.2.11 Cadlinoglycine methyl ester (71). O To a stirred suspension of cadlinoglycine 71 (1.5 mg, 0.004 mmol) in ether (1 mL ) at room temperature was added an ethereal solution of diazomethane dropwise until the yellow color of the diazomethane persisted in the reaction mixture, and T L C analysis showed complete consumption of starting material. Excess diazomethane was removed under a stream of argon, and the remaining solvent removed in vacuo to afford ester 71 as a colorless solid (1.5 mg, ' H N M R ( C 6 D 6 , 500 M H z , 8): 0.70 (s, 3H), 0.76 (s, 3H), 0.78-0.82 (m, 1H), 0.90 (d, J = 13.5 Hz , 1H), 0.94-0.99 (m, 1H), 0.96 (s, 3H), 1.02-1.07 (m, 1H), 1.08-1.12 (m, 1H), 1.31-1.37 (m, 3H), 1.34 (d, J= 7.3 Hz , 3H), 1.62 (d, J= 13.5 Hz , 1H), 2.03-2.08 (m, 3H), 2.25-2.27 (m, 1H), 2.68-2.72 (m, 1H), 3.22 (s, 3H), 3.42 (d, J = 17.4 Hz , 1H), 4.19 (q, J = 7.3 Hz , 1H), 4.65 (d, J = 17.4 Hz , 1H), 4.89 (d, J = 5 Hz , 1H); exact mass calculated for C23H34NO5 (M+H) + : 404.24369; found: 404.24384. O 71 100%). 87 2.5 ENDNOTES: CHAPTER 2: DITERPENOIDS FROM T H E MARINE SPONGE APLYSILLA SP. 1) Faulkner, D. J. Nat. Prod. Rep. 1997,14, 259. 2) Kazlauskas, R.; Murphy, P. T.; Wells, R. J.; Noack, K.; Oberhansli, W. E.; Schonholzer, P. Aust. J. Chem. 1979, 32, 867. 3) Karuso, P.; Skelton, B. W.; Taylor, W. C ; White, A. H. Aust. J. Chem. 1984, 37, 1081. 4) Kazlauskas, R.; Murphy, P. T.; Wells, R. J.; Daly, J. J. Tetrahedron Letters 1979,10, 903. 5) Karuso, P.; Bergquist, P. R.; Cambie, R. C ; Buckleton, J. S.; Clark, G. R.; Rickard, C. E. V.Aust. J. Chem 1986, 39, 1643. 6) Molinski, T. F.; Faulkner, D. J. J. Org. Chem. 1986, 51, 1144. 7) Schmitz, F. J.; Chang, J. S.; Houssain, M . B.; van der Helm, D. J. Org. Chem. 1985, 50, 2862. 8) Walker, R. P.; Faulkner, D. J. J. Org. Chem. 1981, 46, 1098. 9) Bobzin, S. C ; Faulkner, D. J. J. Org. Chem. 1989, 54, 3902. 10 11 12 13 14 15 16 17 18 19 20 Molinski, T. F.; Faulkner, D. J.; Cun-heng, H.; Van Duyne, G. D.; Clardy, J. J. Org. Chem. 1986, 51, 4564. Hambley, T. W.; Poiner, A.; Taylor, W. C. Tetrahedron Letters 1986, 27, 3281. Hochlowski, J. E.; Faulkner, D. J.; Matsumoto, G. K.; Clardy, J. J. Org. Chem. 1983, 48, 1141. Carmely, S.; Cojocaru, M. ; Loya, Y.; Kashman, Y. J. Org. Chem. 1988, 53, 4801. Karuso, P.; Taylor, W. C. Aust. J. Chem. 1986, 39, 1629. Gustafson, K.; Andersen, R. J.; Cun-heng, H.; Clardy, J. Tetrahedron Lett. 1985,26,2581. Cimino, G.; de Rosa, D.; de Stefano, S.; Minale, L. Tetrahedron 1974, 30, 645. Tischler, M. ; Andersen, R. J. Tetrahedron Letters 1989, 42, 5717. Tischler, M. ; Andersen, R. J.; Cloudhary, M . I.; Clardy, J. J. Org. Chem. 1991, 56, 42. Tischler, M. Ph.D. Thesis, University of British Columbia, Vancouver, B.C., 1990. Pika, J.; Tischler, M. ; Andersen, R. J. Can. J. Chem. 1992, 70, 1506. 88 21) Pika, J. Personal communication, 1992. 22) Bergquist, P. Personal communication, 1993. 23) Molinski, T. F.; Faulkner, D. J. J. Org. Chem. 1987, 52, 296. 24) Dumdei, E. J.; Kubanek, J.; Coleman, J. E.; Pika, J.; Andersen, R. J.; Steiner, J. R.; Clardy, J. Can. J. Chem. 1997, 75, 773. 25) Pika, J. Ph.D. Thesis, University of British Columbia, Vancouver, B.C., 1993. 26) Hoffmann, J. J.; McLaughlin, S. P.; Jolad, S. D.; Schram, K. H.; Tempesta, M . S.; Bates, R. B. J. Org. Chem. 1982, 47, 1725. 27) Davies-Coleman, M. T.; Faulkner, D. J. Tetrahedron 1991, 47, 9743. 89 3 N E W G E O D I A M O L I D E S F R O M T H E M A R I N E S P O N G E CYMBASTELA SP 3.1 INTRODUCTION Marine sponges are a well established source of unique and biologically active peptides.1 Jaspamide (jasplakinolide) (76), a cyclic depsipeptide with a twelve carbon hydroxy acid fragment linked to three amino acid residues, was independently isolated from a Jaspis sp. (order Choristida) of sponge, one collected in Fiji 2 and the other collected in Palau.3 OH Jaspamide (76) represents the first example of a new class of cyclic depsipeptides containing a common twelve carbon hydroxy acid fragment linked to three variable amino acid residues. The closely related geodiamolides A (77) and B (78) were subsequently isolated from the Caribbean sponge Geodia sp. (order Choristida)4 and geodiamolides A to G (77-83) were isolated from the Papua New Guinea sponge Cymbastela sp.(order Axinellida)5'6 Crews et al. reported that specimens of Auletta cf. Constricta (order Choristida) collected in Papua New Guinea contained 90 jaspamide (76) and an unrelated cytotoxic peptide milnamide A (84).7 Jaspamide (76) has also been isolated from another taxonomically distant species of sponge, Hemiasterella minor, collected in South Africa, along with the novel geodiamolide TA (85) and hemiasterlin (86),8 a tripeptide closely related to milnamide A (84). 77 geodiamolide A X=I, R=Me 83 78 geodiamolide B X=Br, R=Me 79 geodiamolide C X=C1, R=Me 80 geodiamolide D X=I, R=H 81 geodiamolide E X=Br, R=H 82 geodiamolide F X=C1, R=H Neosiphoniamolide A (87), a new compound closely related to geodiamolide TA (85), was isolated from the New Caledonian sponge Neosiphonia superstes? Al l of the metabolites in the jaspamide/geodiamolide class exhibit potent antimicrobial and cytotoxic activities.3"5'8'9 Jaspamide was also reported to be insecticidal. The mechanism by which jaspamide (76) exhibits its antiproliferative effect observed in in vitro studies is by the disruption of the actin cytoskeleton.10 This effect manifests itself during mitosis (see Figure 64, Chapter 4). Following metaphase, the contractile ring that will divide the cell into two daughter cells forms in the cleavage furrow during cytokinesis by the polymerization of actin. Disruption of actin polymerization interferes with the formation of this ring, thus preventing cytokinesis and resulting in cells with two or more nuclei which will eventually die.10 91 85 R=Me 87R=H The sponges that produce the jaspamide/geodiamolide and hemiasterlin families of metabolites are taxonomically distinct (Figure 47). It seems unlikely, from an evolutionary point of view, that all of these sponges have independently developed the ability to biosynthesize these metabolites de novo. The fact that many sponges contain associated microorganisms, such as cyanobacteria, has long been known11 and there has been speculation that some known sponge natural products are not products of sponge enzymes, but are rather the products of the associated microorganisms. To account for the jaspamide and the hemiasterlin family of metabolites occurring across such a diverse group of sponges, it has been suggested that one or two species of microorganisms living within the sponges produce these metabolites.4 Recently, a group of depsipeptides structurally similar to jaspamide (76) were isolated from cultures of various strains of Chondromyces myxobacteria obtained from soil samples collected in Madeira, Spain. l 2 a , b The chondramides A to D (88-91) possess moderate anti-fungal activity and are potently cytotoxic. x 2 a , h The chondramides represent the first microbial metabolites in the jaspamide/geodiamolide family. Several groups are currently attempting to isolate marine microorganisms that produce these metabolites. O H There continues to be considerable interest in jaspamide and the geodiamolides as lead compounds in drug development and also as synthetic targets. Recently, the American National Institute of Health selected geodiamolides A (77), D (80), and E (81) for entry into in vivo Hollow Fiber Assay based upon their activity in the National Cancer Institute's 60 cell line in vitro anticancer screen. The original small collection of Pseudaxinyssa sp. sponge examined by our group did not provide the quantities of geodiamolides A to F (77-83) required for in vivo evaluation in animal models relevant to the observed in vitro cytoxicities against human cancer cell lines. 5 ' 1 3 In order Kingdom Phylum Class Subclass Order Family Genus Ceractinomorpha Halichondrida Axinellidae 93 Metazoa (multi-cellular animals) Porifera (sponges) Demospongiae Tetractinomorpha Astrophorida Hadromerida v V V Geodiidae Coppatiidae Hemiastrellida Auletta Cymbastela Geodia Jaspis | geodiamolides | |jaspamide" Hemiastrella milnamide A jaspamide hemiasterlins criamides geodiamolides hemiasterlins jaspamide geodiamolides Figure 47. Phylogenic classification of sponges that produce metabolites in the jaspamide/geodiamolide families. to obtain sufficient amounts of the geodiamolides for in vivo testing, the source sponge was recollected from the original dive sites at Motupore and Madang in Papua New Guinea. Taxonomic identification of a voucher sample of the recollected sponge verified that it was identical to the original sample; however, in the period between collections, the classification of the sponge had been changed from Pseudaxinyssa sp. to Cymbastela sp. (an undescribed species).14 Bioassay guided fractionation of extracts from the recollected Cymbastela sp. led to the isolation of the known metabolites geodiamolides A to G (77-83) and hemiasterlin (86), and the novel metabolites geodiamolide H to N and P (89-95, 97), hemiasterlin A (98), hemiasterlin B (99), criamide A (100) and criamide B(101).* Figure 47 shows a schematic of the classification of the Cymbastela sp. sponge from which the geodiamolides, hemiasterlins, and criamides described in this thesis have been isolated. The sample of Cymbastela sp. sponge used in this study had tan yellow stalks growing upright with finger-like projections. Overall size of each animal was about 20 cm high with stalks ranging from 0.5 to 1.0 cm thick (Figure 48). Figure 48. Photograph of Cymbastela sp. sponge. * Please see Chapter 4 for a complete description of the hemiasterlins and criamides. 3.2 RESULTS AND DISCUSSION 95 Specimens of Cymbastela sp. were collected by hand using SCUBA on outer reefs at -10 to -25 meters near Motupore and Madang in Papua New Guinea (see Collection Map, Figure 49). Freshly collected sponge was frozen on site and transported to Vancouver packed in dry ice. Bio-assay guided fractionation of the Cymbastela sp. sponge was performed as follows: freeze dried sponge (160 g dry weight) was extracted exhaustively with a 1:1 solution of CH 2Cl 2/MeOH (4 x 4L). Combined organic extracts were filtered and concentrated in vacuo to yield 14 g of a dark yellow/orange oil which was suspended in 250 mL of a 9:1 solution of MeOFI/H20 and extracted with hexanes (3 x 150 mL). The MeOH/H 20 layer was diluted with H 2 0 to a 4:1 MeOH/H 20 solution and extracted with CC14 (3 x 150 mL). Further dilution of the MeOH/H 20 layer with H 2 0 resulted in a 2:3 MeOH/H 20 solution, which was extracted with C H C I 3 (3 x 150 mL). Al l traces of MeOH were removed in vacuo and the resulting aqueous solution was extracted with EtOAc (3 x 150 mL). The biological activity was concentrated in the CCI4 and C H C I 3 soluble portions, which were combined and fractionated further by Sephadex LH-20 chromatography, eluting with MeOH, to yield two active fractions, one containing a mixture of geodiamolides and one containing a mixture of peptides as identified by TLC and *H NMR. Sequential application of the geodiamolide fraction to silica-gel flash column chromatography using a step gradient from 1:1 to 2:1 EtOAc/hexanes and reversed phase HPLC, eluting with 2:3 H 20/MeOH, yielded pure samples of geodiamolide H (89) (1 mg, 0.0004 % dry wt), geodiamolide I (90) (0.3 mg, 0.00012 % dry wt), geodiamolide J (91) (2 mg, 0.0008 % dry wt), geodiamolide K (92) (1.5 mg, 0.0006 % dry wt), geodiamolide L (93) (0.8 mg, 0.00032 % dry wt), geodiamolide M (94) (1.5 mg, 0.0006 % dry wt), geodiamolide N (95) (1.5 mg, 0.0006 % dry wt), geodiamolide P (97) (1 mg, 0.0004 % dry wt), and the known metabolites, 96 geodiamolides A to G (77-83). The structures of geodiamolides H to N (89-95) were determined using spectroscopic methods, particularly one and two dimensional NMR spectroscopy and mass spectrometry. The structure of geodiamolide P (97) was solved by comparison of its *H NMR and mass spectral data with the literature values for geodiamolide D (80).5 Geodiamolides A to G (77-83) were identified by comparing their HREIMS and 'H NMR data with literature values.4"6 Copies of 1 3 C NMR spectra and two dimensional NMR spectra (COSY, HMBC, and HMQC) for geodiamolide J (91) and geodiamolide M (94) can be found in Appendix A. 0* A ^ £ i \ V N L ) 0 Madang b • "ax* o ° — -i ) c; o J L ^ v r j ^ V A -X Austi S- .. J •alia x 51i IS — 30*S V V ) f ^ > ( V ' > \ h • V / n • r • . . _ J • r / j Motopore v ^ Island 8S 142 E 146 E 150 E Figure 49. Map of the collection site for Cymbastela sp. 97 3.2.1 G e o d i a m o l i d e H (89) Geodiamolide H (89) was isolated as a colourless glass which gave a parent ion cluster in the HRDCIMS at m/z M + 610.19431 and 608.19508 corresponding to a molecular formula of C28H 3 9N 307Br (AM -1.49, -3.38 ppm). A comparison of the ! H NMR spectroscopic data for compound 89 with the lH NMR data collected for geodiamolide G (83) (Table 7) revealed that the alanine-polyketide-alanine portions of the molecule are identical and the only differences are in the chemical shifts of the resonances assigned to the tyrosine moiety. The change in chemical shift of the resonance assigned to HI 7, from 8 7.45 in geodiamolide G (83) to 8 7.25 in 89, is consistent with 89 being the C18 bromo analogue of geodiamolide G (83). Table 7. *H NMR data for geodiamolide G (S3\ geodiamolide H (89), and geodiamolide I (90). (Recorded in CDC13 at 500 MHz.) Carbon No. geodiamolide G (83) geodiamolide H (89) geodiamolide I (90) 1 2 2.46 m 2.45 m 2.46 m 3 2.55 dd J= 12, 4 Hz. 2.54 dd7= 13,4 Hz 2.54 d J = 12 Hz 3' 4 2.1417 = 12 Hz 2.14 7= 12Hz nsa 5 6 2.95 nsa nsa 7 1.82 dddJ= 15,9, 3 Hz 1.82 m nsa 7* 1.61 ddd7 = 15, 11, 3 Hz 1.60 m nsa 8 0 5.11 m 5.11 m 5.11 m y 10 4.51 dq 7=7, 7 Hz 4.50 dqJ=7, 7 Hz 4.51 m NH(10) 6.35 d J = 7 H z 6.21 d7=7Hz 6.21dJ=8Hz 11 12 5.06 ddJ=9, 8 Hz 5.06 dd7=9, 8 Hz 5.06 m 13 14 4.72 dqJ=8, 7 Hz 4.71 dq7=8, 7 Hz 4.71 m NH(14) 6.19d7=7Hz 6.36 d J = 8 Hz 6.37 d J= 8 Hz 15 3.12ddJ= 15,8Hz 3.13 ddJ= 15, 8 Hz 3.14m 15' 2.90 ddJ = 15, 9 Hz 2.91 dd 7= 15, 9 Hz 2.90 m 16 17 7.45 d J = 1 Hz 7.25 d7 = 2Hz 7.12 s 18 19 20 6.88dJ=8Hz 6.91 d7=8Hz 6.91 d J = 8 H z 21 7.04 ddJ = 8, 1 Hz 7.02 dd J= 8, 2 Hz 6.98dJ=8Hz 22 1.15dJ=6Hz 1.14dJ=7Hz 1.14dJ=7Hz 23 5.90 s 5.90 s 5.90 s 23' 5.78 s 5.78 s 5.79 s 24 1.09dJ=7Hz 1.09dJ=7Hz 1.09dJ=7Hz 25 1.28dJ=6Hz 1.28d7=6Hz 1.28dJ=6Hz 26 1.32dJ=7Hz 1.31 dJ=7Hz 1.31 d7=7Hz 27 2.97 s 2.96 s 2.96 s 28 1.04 d7=7 Hz 1.03 d J = 7 H z 1.03 d J = 7 H z 'Signals not seen due to sample size and/or interference with H 2 0 peak. 100 3.2.2 Geodiamolide I (90) 90 Geodiamolide I (90) was isolated as a colourless glass which gave parent ion clusters in the HREIMS at m/z 565.23956 and 563.24011 corresponding to a molecular formula of C28H39N3O7CI (AM -4.7, -0.5 ppm). The NMR spectrum of geodiamolide I (90) (Figure 51) is nearly identical to that of geodiamolide G (83), again only differing significantly in the chemical shifts of the resonance assigned to HI 7 (Table 7). The change in the chemical shift of the resonance assigned to HI7, from 5 7.45 in geodiamolide G (83) to 5 7.12 in 90, indicates that 90 is the C l 8 chloro analog of geodiamolide G (83). 101 102 3.2.3 Geodiamolide J (91) 91 Geodiamolide J (91) was isolated as a clear colourless glass that gave a (M+H)+ ion in the HRCIMS at m/z 658.19767, appropriate for a molecular formula of C28H40O7N3I (AM -1.91 ppm). The molecular formula differed from that of geodiamolide A (77) by the addition of one oxygen atom but indicated that the molecule still has ten sites of unsaturation, consistent with the addition of an hydroxyl functionality. Detailed analysis of the 'H NMR, COSY, HMQC, and HMBC spectra of 91, summarized in Table 8, indicated that 91 contains the same polyketide and the same A -^methyl iodotyrosine moieties as found in geodiamolide A (77). Analysis of the COSY spectrum identified an isolated spin system with correlations observed from a resonance at 8 7.00 (d, J = 8 Hz: NH10) into a methine resonance at 6 4.43 (m: H10), which in turn correlates to two diastereotopic proton resonances at 5 3.97 (dd, J = 11, 4 Hz: H26) and 5 3.81 (dd, J = 11, 4 Hz: H26'). An HMQC correlation from the proton resonances at 5 3.97 and 8 3.81 (H26 and H26') to a carbon resonance at 8 63.1 (C26), combined with an HMBC correlation from a proton resonance at 8 4.43 (H10) to the carbon resonance at 8 169.8 (C9), confirmed the presence of a serine residue. 104 Table 8. NMR data for geodiamolide J (91V (Recorded in CDC13 at 500 MHz.) Carbon No. 5 1 3 C S X H C O S Y H M B C a 1 175.5 H3', 22 2 42.4 2.26 m H3, 3', 22 H22 3 42.3 2.17 m H2, 3' H22, 23 3' 1.98 dd7=3, 14 Hz H2,3 4 132.9 H3, 3', 23 5 131.3 4.87d7=9Hz H6, 23 H23, 24 6 29.3 2.18m H5,24 H5, 7, 24 7 43.7 1.63 ddd 7= 14, 8, 6 Hz H8 H24, 25 7' 1.39 ddd 7= 14, 8, 4 Hz H8 8 71.9 4.94 m H7, 7', 25 H7, 25 9 169.8 H10 10 53.3 4.43 dq J = 7, 7 Hz NH-10, H26, 26' NH(10) 7.00 d J= 8 Hz H10 11 168.9 H12 12 56.9 5.26 ddJ=9, 7 Hz H15, 15' H15, 15' 13 174.6 H27, 28 14 45.8 4.71 dq 7=9, 4 Hz HN-14, H28 H28 NH(14) 6.44d7=6Hz H14 15 32.5 3.19 ddJ= 15, 7 Hz H12, 15' H12 15' 2.89 ddJ= 15, 9 Hz H12, 15 16 130.3 H15, 15' 17 138.2 7.46dJ=2Hz H21 H15, 15*, 21 18 85.0 H20 19 154.0 H17, 20 20 115.1 6.86 d J = 8 Hz H21 21 130.4 7.03 ddJ=8,2Hz H20 H15, 15', 20 22 19.0 1.12(17= 7 Hz H2 23 17.8 1.45 s H5 H5 24 20.5 0.86dJ=7Hz H6 25 20.8 1.24dJ=6Hz H8 H7 26 63.1 3.97 ddJ= 11,4 Hz H10, 26' 26' 3.81 ddJ= 11,4Hz H10, 26 27 30.7 2.97 s H12 28 18.5 1.08dJ=7Hz H14 'Proton resonances that are correlated to the carbon resonance in the 5 C column. An isolated spin system in the COSY spectrum, containing correlations from a resonance at 5 6.44 (d, J = 6 Hz: NH14) to a methine resonance at 8 4.71 (dq, J = 9, 4 Hz: H14), and from 8 4.71 (HI4) to a methyl resonance at 8 1.08 (d, J = 7 Hz: Me28), established the NH14/C14/C28 connectivity. The combination of these COSY correlations, along with an 105 HMBC correlation from the methyl resonance at 5 1.08 (Me28) to the amide carbonyl resonance at 5174.6 (C13), identified an alanine residue. It was apparent, therefore, that 91 is simply a serine analogue of geodiamolide A (77), and what remained to be determined was which of the two alanine residues in 77 had been replaced by serine. HMBC correlations between the resonance identified as the iodotyrosine JV-Me (6 2.97, s: Me27) and the amide carbonyl resonance at 8 174.6 (C13), and between the resonance at 5 6.44 (d, J = 6 Hz: NH14) and the polyketide amide carbonyl resonance at 8 175.5 (Cl) positioned the alanine residue between the 7V-Me-iodotyrosine and polyketide moieties (Figure 53). The serine moiety, therefore, has to be the link between the amide carbonyl carbon C l 1 (8 168.9) of the iodotyrosine and the terminal hydroxy end of the polyketide residue, closing the ring with an ester linkage at C8 (8 71.9). This connectivity was established using electrospray mass spectral data. Figure 53. Selected HMBC correlations for the alanine unit in 91. The MS/MS electrospray mass spectrum of the molecular ion of geodiamolide J (91) contains two major fragmentations (Figure 54). The ion observed at 409 Da in the electrospray mass spectrum was attributed to the dipeptide fragment 105, which results from the cleavage C13-N amide bond and the cleavage of the C8-0 bond via a McLafferty rearrangement. The observation of fragment 105 indicated that the serine residue must be attached to the amide carbonyl carbon (Cl 1) of the methyl iodotyrosine through an amide bond. Also, in order for the 106 McLafferty rearrangement to be possible, the hydroxyl function of the polyketide fragment had to be attached to the ester carbonyl carbon (C9). Further fragmentation of the ion measuring 409 Da led to the base peak at 276 Da, which arises from cleavage of C l 1-N tyrosine/serine amide bond yielding the methylated iodo-typtamine moiety 106. . O H 13 14 91 9. OH 105 observed: m/z 409 Da calculated: m/z 409 Da ( C 1 3 H 1 8 I N 2 0 5 ) 106 observed: m/z 276 Da calculated: m/z 276 Da ( C 9 H n I N O ) Figure 54. Fragmentation of geodiamolide J (91) in the electrospray MS/MS spectrum. The configuration of the serine residue of geodiamolide J (91) was determined by chemical degradation followed by Marfey's reagent analysis. Hydrolysis of the parent compound with 6 N HC1 followed by reaction with Marfey's reagent15 and HPLC analysis of the derivatized amino acids determined that the serine residue has the L configuration. The similarity between the ! H and 1 3 C NMR shifts for 91 and geodiamolide A (77)5 implied that the 107 chiral centers in the iodotyrosine, alanine, and polypropionate fragments have the same relative configurations in both molecules. 3.2.4 Geodiamolide K (92) 92 Geodiamolide K (92) was isolated as a clear, colourless glass. The HREIMS of 92 revealed that it has a molecular formula of C28H39N307Br (m/z 611.20294/609.20496, AM -1.3/-0.4 ppm, respectively, for the Br and Br isotopes). From the Ff NMR spectrum of 92 (Figure 55), it was clear that the serine-polyketide-alanine part of the molecule is identical to that of geodiamolide J (91), as the resonances for these fragments are superimposable (Table 9). The only difference in the ! H NMR spectra of the two molecules is the chemical shifts of the resonances assigned to the tyrosine moiety. In particular, the resonance assigned HI7 shifts from 8 7.46 in the *H NMR spectrum of geodiamolide J (91) to 8 7.27 in the *H NMR spectrum of 92. This change in chemical shift of HI 7 is consistent with the replacement of the tyrosine iodine 108 Table 9. *H NMR data for geodiamolide J (91), geodiamolide K (92), and geodiamolide L (93) (Recorded in CDC13 at 500 MHz.) geodiamolide J (91) Geodiamolide K (92) geodiamolide L (93) Carbon No. 8*H 8 ^ 8*H 1 2 2.26 m 2.28 m 2.27 m 3 2.17m 2.17m 2.16m 3' A 1.98 ddJ= 14, 3 Hz 1.98 ddJ =  14, 3 Hz 1.98 ddJ = 12, 2 Hz 4 5 4.87dJ=9Hz 4.87 d J = 9 Hz 4.87 dJ = 9 Hz 6 2.18 m 2.18 m 2.16 m 7 1.63 dddJ= 14, 8, 6 Hz 1.64 m nsa 7* 1.39 ddd J= 14, 8, 4 Hz 1.37m 1.37 m 8 Q 4.94 m 4.94 m 4.93 m y 10 4.43 dqJ=7, 7 Hz 4.43 m 4.43 m NH(10) 11 7.00 d J = 8 Hz 6.99 dJ = 8 Hz 6.98 dJ = 7 Hz 11 12 11 5.26 ddJ=9, 7 Hz 5.27 ddJ =  9, 7 Hz 5.28 ddJ-= 9, 7 Hz 10 14 4.71 dqJ=9,4Hz 4.71 m 4.72 m NH(14) 6.44dJ=6Hz 6.43 d J = 6 Hz 6.44 dJ = 6 Hz 15 3.19 ddJ= 15, 7 Hz 3.22 ddJ =  15, 7 Hz 3.22 ddJ =  15, 7 Hz 15' 2.89 ddJ= 15, 9 Hz 2.91 ddJ =  15, 9 Hz 2.93 ddJ-= 15, 9 Hz 16 17 7.46dJ=2Hz 7.27 dJ = 2 Hz 7.13 dJ = 2 Hz 18 19 20 6.86 d J= 8 Hz 6.91 d J = 8 Hz 6.91 dJ = 8 Hz 21 7.03 ddJ=8,2Hz 7.03 ddJ-= 8, 2 Hz 7.00 ddJ =  8, 2 Hz 22 1.12 d J = 7 H z 1.12dJ = 7 Hz 1.13 dJ = 7 Hz 23 1.45 s 1.45 s 1.45 s 24 0.86dJ=7Hz 0.86 d J = 7 Hz 0.87 d J = 7 Hz 25 1.24dJ=6Hz 1.23 dJ = 6 Hz 1.23 dJ = 6 Hz 26 3.97 ddJ= 11, 4 Hz 3.97 ddJ =  4, 11 Hz 3.97 ddJ-= 11, 4 Hz 26' 3.81 ddJ= 11,4Hz 3.82 dm J = 11 Hz 3.82 dm J = 11 Hz 27 2.97 s 2.97 s 2.97 s 28 1.08dJ=7Hz 1.09 dJ = 7 Hz 1.09dJ = 7 Hz "Signal not seen due to sample size and/or interference with H 2 0 peak. in geodiamolide J (91) with bromine in 92. Therefore, it was apparent that 92 is simply the CI 8 bromo analog of geodiamolide J (91). 110 3.2.5 Geodiamolide L (93) 93 Geodiamolide L (93) was isolated as a clear, colourless glass. The HREIMS of 93 revealed that it has the molecular formula C28H39N3O7CI (m/z 567.25250/565.25549 AM 1.2/2.6 ppm respectively for the Cl and Cl isotopes). As was the case with geodiamolide K (92), the *H NMR spectrum of 93 (Figure 56) differed from the ! H NMR spectrum geodiamolide J (91) only in the chemical shift of the resonance assigned to HI 7 (Table 9). The shift from 8 7.46 in ! H NMR spectrum of geodiamolide J (91) to 8 7.13 in ! H NMR spectrum of 93 is consistent with replacement of the tyrosine iodine in geodiamolide J (91) with chlorine in geodiamolide L (93). Thus, 91 was established to be simply the C18 chloro analog of geodiamolide J (93). I l l 112 3.2.6 Geodiamolide M (94) Geodiamolide M (94) was isolated as a clear, colourless glass, which exhibits a parent ion in the HREIMS at m/z 658.19880, appropriate for a molecular formula of C28H40O7N3I (AM -0.19 ppm). The molecular formula of geodiamolide M (94) is identical with the molecular formula of geodiamolide J (91) and their *H NMR spectra (Figure 57 and 52, respectively) are very similar. Analysis of the ! H NMR, COSY, HMQC, and HMBC spectra of 94 showed that it contains the same polyketide fragment and the same substituted iV-methyl iodotyrosine moiety as found in geodiamolide A (77), but again differs in one of the alanine residues (Table 10). Based on these similarities, it was assumed that the C13/C14 alanine residue of geodiamolide A (77) is replaced by a serine residue in geodiamolide M (94). Analysis of the COSY spectrum identified an isolated spin system starting with a correlation between a resonance at 8 6.67 (br s: NH14) and a methine resonance at 8 4.71 (m, H14), and ending with a correlation from H14 (8 4.71) to a broad singlet resonance at 8 3.54 (H28 and H28'). The singlet resonance at 8 3.54 (H28) correlates to the carbon resonance at 8 65.2 (C28) in the HMQC spectrum, thus establishing the presence of the serine residue. 114 Table 10. NMR data for geodiamolide M (941 (Recorded in CDC13 at 500 MHz.) Carbon No. 1 2 3 3' 4 5 6 7 7' 8 9 10 NH(10) 11 12 13 14 NH(14) 15 15' 16 17 18 19 20 21 22 23 24 25 26 27 28 5 1 3 C a 177.6 42.2 43.2 147 131.2 28.7 43.9 70.6 170.4 48.8 168.4 57.1 171.3 52.8 32.2 130.1 138 85.3 154.1 114.9 130.1 18.8 17.7 20.4 20.5 18.1 30.9 65.2 8*H 2.40 m 2.14 m 2.04 m 4.93 d J = 9 Hz 2.16m 1.56 m 1.37 m 4.87 m 4.5 dqj=8, 7 Hz 6.45 d .7=7 Hz 5.16ddJ=9, 8 Hz 4.71 m 6.67 br s 3.17 ddJ= 15, 8 Hz 2.88 ddJ= 15, 9 Hz 7.46 d J = 2 H z 6.88dJ=8Hz 7.05 ddJ=8, 2 Hz 1.16dJ=7Hz 1.49 s 0.88dJ=7Hz 1.23dJ=6Hz 1.32dJ=7Hz 3.03 s 3.54 brs C O S Y H3, 3', 22 H2, 3' H2, 3 H6, 23 H5,24 H7' H7 H25 HN10, H26 H10 H15, 15' HN14, 28 H14 H12, 15' 12, 15 H21 H21 H20 H2 H5 H6 H8 H10 H14 H M B C H3', 22 H22 H23 H23 H3', 23, 24 H24 H24, 25 H25 H10, 26 H12 H15, 15', 27 H27 H12, 17 H15, 15', 17, 20 H15, 15', 21 H17, 20 H17,21 H17, 15, 15', 20 H5 H12 "Obtained from H M Q C and H M B C spectra only. bProton resonances that are correlated to the carbon resonance in the 8 1 3 C column. Another isolated spin system in the COSY spectrum, containing a correlation from an NH resonance at 8 6.45 (d, J = 7 Hz: NH10) to a methine resonance at 8 4.45 (dq, J - 8, 7 Hz: H10), and a correlation from H10 (8 4.45) to a methyl resonance at 8 1.32 (Me26), established 115 the NH10/C10/C26 connectivity. An HMBC correlation from Me26 (8 1.32) to an ester carbonyl resonance at 8 170.4 (C 13) confirmed the presence of the alanine residue. No connectivity information indicating the position of the serine or the alanine residues was obtained from the HMBC spectrum. Fortunately, the iodotyrosine/alanine connectivity was established using the electrospray mass spectral data. Two major fragmentations are observed in the MS/MS electrospray mass spectrum of the molecular ion of geodiamolide M (91) (Figure 58). The ion observed at 393 Da was attributed to the dipeptide fragment 109, which results from the cleavage C13-N amide bond and the cleavage of the C8-0 bond via a McLafferty rearrangement. This indicated that the alanine residue is attached to the amide carbonyl carbon (Cl 1) of the methyl iodotyrosine through an amide bond. Furthermore, in order for the McLafferty rearrangement to be possible, the hydroxyl function of the polyketide fragment has to be attached to the ester carbonyl carbon (C9). Further fragmentation of the ion measuring 393 Da led to the base peak at 276 Da, which arises from the cleavage of the C l 1-N tyrosine/alanine amide bond yielding the methylated iodo-typtamine moiety (106). The configuration of the serine residue of geodiamolide M (94) was determined by chemical degradation followed by Marfey's reagent analysis. Hydrolysis of the parent compound with 6 N HC1 followed by Marfey's amino acid analysis15 determined that the serine residue has the L configuration. Again, the similarity in the H and C NMR shifts between metabolite 94 and geodiamolide A (77)5 implied that the chiral centers in the iodo-tyrosine, alanine and polypropionate fragments have the same relative configurations in both molecules. 116 Figure 58. Fragmentation of geodiamolide M (94) in the electrospray MS/MS spectrum. 117 3.2.7 Geodiamolide N (95) 95 Geodiamolide N (95) was isolated as a clear, colourless glass. The HREIMS of 95 revealed that it has a molecular formula of C28H39N307Br (m/z 611.20374/609.20515, AM -1.3/-0.3 ppm, respectively, for the 8 1 Br and 7 9 Br isotopes). The ! H NMR data obtained for 95 (Table 11) is nearly identical with that recorded for geodiamolide M (94). The change in chemical shift of the resonance assigned to HI 7 from 8 7.46 in geodiamolide M (94) to 8 7.27 in 95, indicates that 95 is simply the C18 bromo analogue of geodiamolide M (94). Although no chloro analogue of geodiamolide M (94) was found, it is likely to be a natural product, and therefore the name geodiamolide O was assigned to the hypothetical structure 96, in anticipation of its future discovery or synthesis. 96 Table 11. *H NMR data for geodiamolide M (94) and geodiamolide N (95) (Recorded in CDC13 at 500 MHz.) Carbon No. geodiamolide M (94) geodiamolide N (95) 1 2 3 3' 4 5 6 7 7' 8 9 10 NH(10) 11 12 13 14 NH(14) 15 15' 16 17 18 19 20 21 22 23 24 25 26 27 28 2.40 m 2.14m 2.04 m 4.93 d J = 9 H z 2.16 m 1.56 m 1.37m 4.87 m 4.45 dqJ=8, 7 Hz 6.45dJ=7Hz 5.16ddJ=9,8Hz 4.71 m 6.67 br s 3.17 ddJ = 15, 8 Hz 2.88 ddJ= 15, 9 Hz 7.46dJ = 2Hz 6.88dJ=8Hz 7.05 ddJ= 8, 2 Hz 1.16dJ=7Hz 1.49 s 0.88dJ=7Hz 1.23dJ=6Hz 1.32dJ=7Hz 3.03 s 3.54 brs 2.41 m 2.10 m 2.04 m 4.93 d J = 9 Hz 2.15m 1.56m 1.37 m 4.87 m 4.45 m 6.46 d J= 8 Hz 5.17 dd./= 10, 7 Hz 4.79 m 6.68 br s 3.18 ddJ= 15, 7Hz 2.91 dd J= 15, 10 Hz 7.27 d J = 2 H z 6.92dJ=8Hz 7.03 ddJ= 8, 2 Hz 1.16dJ=7Hz 1.49 s 0.86dJ=6Hz 1.23dJ=6Hz 1.32dJ=7Hz 3.03 s 3.54 brs 120 3.2.8 Geodiamolide P (97) Geodiamolide P (98) was isolated as a colourless glass, which exhibits a parent ion in the HREIMS at m/z 643.17539, corresponding to the molecular formula of C 2 7 H 3 8 0 7 N 3 l (AM 0.1 ppm). Analysis of the *H NMR spectrum of 97 (Figure 60) indicated that it is very similar to geodiamolide J (91), but lacks one methyl group. A comparison between the *H NMR spectroscopic data for geodiamolide P (97) and geodiamolide D (80) (Table 12) suggested that 97 is simply the serine analogue of geodiamolide D (80). The only significant difference in the data is the absence of the alanine methyl doublet, observed at 8 1.25 (Me26) in the 'H NMR spectrum geodiamolide D (80), and the appearance of a pair of methylene resonances at 8 3.91 (dm, J = 8 Hz: H26) and 8 3.75 (m; H26') in ! H NMR spectrum of 97. The deshielded chemical shifts of the resonances assigned to H26 and H26' are appropriate for the P protons of serine. Evidence for the connectivity of 97 was obtained from the MS/MS electrospray mass spectral data (Figure 61). The ion observed at 409 Da was attributed to the dipeptide fragment 105, which results from the cleavage of the C13-N amide bond and the C8-0 bond. This establishes the serine-iodotyrosine connectivity. Further fragmentation of the ion measuring 409 Da led to the base peak at 276 Da, which could arise from the cleavage of the C l 1-N amide bond 121 122 Table 12. 'H NMR data for geodiamolide D (80) and geodiamolide P (97). (Recorded in CDC13 at 500 MHz.) geodiamolide D (80) geodiamolide P (97) Carbon No. 8*H 8*H 2 2.42 m 2.39 m 3 2.10m 2.10 m 3' 2.10m 2.05 m 5 4.99dJ=8Hz 4.97dJ=9Hz 6 2.21 m 2.19m 7 1.69 m 1.70 m 7' 1.40 m 1.41 m 8 4.86 m 4.88 m 10 4.51 dq7=8, 7 Hz 4.53 dqJ=8, 7 Hz NH(10) 6.6 d J= 8 Hz 6.87 d7 = 8 Hz 12 5.08 ddJ=9, 7 Hz 5.11 dd7=9, 7 Hz 14 4.16 dd J = 18, 1 Hz 4.15 ddJ= 18, 4 Hz 14' 3.77 dd J= 18, 4 Hz 3.78 ddJ= 18, 3 Hz NH(14) 6.45 dd 7 = 4, 1 Hz 6.47 br s 15 3.25 ddJ= 14, 9 Hz 3.24 ddJ= 14, 9 Hz 15' 2.81 ddJ= 14, 7 Hz 2.81 ddJ = 14, 7 Hz 17 7.51 d J = 2 H z 7.50dJ=2Hz 20 6.90dJ=8Hz 6.89dJ=8Hz 21 7.08 dd 7=8, 2 Hz 7.07 ddJ= 8, 2 Hz 22 1.16dJ=7Hz 1.14dJ=7Hz 23 1.53 dJ = 1 Hz 1.49 s 24 0.90dJ=7HZ 0.89dJ=7Hz 25 1.25dJ=6Hz 1.26dJ=6Hz 26 1.30dJ=7Hz 3.91 dmJ=8Hz 26' 3.75 m 27 2.94 s 2.93 s to yield the methylated iodotyptamine moiety 106. The similarity between the H NMR spectrum of geodiamolide D (80) and the *H NMR spectrum of 97, combined with the mass spectral evidence for the two partial structures 105 and 106, allowed the structure of 97 to be determined. No additional stereochemical information was obtained for 97, but it is likely to have the same relative stereochemistry as all of the other geodiamolides. Figure 61. Fragmentation of geodiamolide P ( 9 7 ) in the electrospray MS/MS spectrum. 3.3 CONCLUSIONS 124 During the course of re-isolating geodiamolides A to G (77-83) from Cymbastela sp. for further biological testing, geodiamolides H to N (89-95) and geodiamolide P (97) were isolated. The structures of geodiamolides H (89) and I (90) were determined by comparing their 'H NMR and mass spectral data with the literature values for geodiamolide G (83).6 Geodiamolide H (89) and geodiamolide I (90), along with geodiamolide G (83), complete the iodo, bromo, and chloro series of the enone geodiamolides. These are the only three cyclic depsipeptides in the jasparnide/geodiamolide family that contain a modified polypropionate unit. The structures of geodiamolides J to N (91-95) were determined by interpretation of their spectroscopic data and the structure of geodiamolide P (97) was determined by a comparison between its ] H NMR data and the 'H NMR data for geodiamolide D (80), combined with analysis of its fragmentation pattern in the electrospray mass spectrum. Geodiamolides J to N (91-95) and geodiamolide P (97) represent the first examples in this class of compounds where a serine residue has been incorporated into the structure. No biological testing has been carried out on the new geodiamolides. Including these new compounds, there are now 17 representatives of the geodiamolide family, which were isolated from four species of sponge (Figure 62). 4 , 5 , 7' 9 Variations have been observed in all three amino acid positions and also in the polyketide portion of the molecule. Amino acid A (see Figure 62) varies between alanine (77-78, 94, 95), serine (91-93, 97), and valine (85, 87). Amino acid B varies only in the C18 halogen atom, with iodine (77, 80, 83, 85, 87, 91, 94, 97), bromine (78, 81, 89, 92, 95), and chlorine (79, 82, 90, 93) all being observed. Amino acid C has been alanine (77-79, 85, 91-93), glycine (80-82, 87, 97), and serine (94, 95). 125 Compound R i X 77 geodiamolide Me Me I 78 geodiamolide B a , b Me Me Br 79 geodiamolide C b Me Me CI 80 geodiamolide D b Me H I 81 geodiamolide E b Me H Br 82 geodiamolide F b Me H CI 91 geodiamolide Jb C H 2 O H Me I 92 geodiamolide K b C H 2 O H Me Br 93 geodiamolide L b C H 2 O H Me CI 94 geodiamolide M b Me C H 2 O H I 95 geodiamolide N b Me C H 2 O H Br 97 geodiamolide P b C H 2 O H H I 85 geodiamolide TA° /-Pr Me I 87 neosiphoniamolide A d /-Pr H I sponge source: Geodia sp. b sponge source: Cymbastela sp.5 c sponge source: Hemiasterella minor* d sponge source: Neosiphonia superstes9 Figure 62. Compounds in the geodiamolide family. I H Compound 83 geodiamolide G b I 89 geodiamolide H b Br 90 geodiamolide I b CI Comparison of the cytotoxicities of the previously reported geodiamolides A to F (77-82), and TA (85) showed that significant variation in the three amino acid residues causes only minor changes in the levels of cytotoxicity exhibited by the compounds. In contrast, geodiamolide G (83) (in vitro human glioblastoma/astrocytoma U373, ED50 7.7 pg/mL; in vitro human ovarian carcinoma HEY, ED50 8.6 pg/mL), with its modified polyketide fragment, is significantly less cytotoxic than the analogous geodiamolide A (77) (in vitro human glioblastoma/astrocytoma U373, ED50 0.016 pg/mL; in vitro human ovarian carcinoma HEY, ED50 0.043 pg/mL). No data are available to determine the effect of the serine residue on the biological activity. 126 The jaspamide/geodiamolide family of metabolites occurs across a taxonomically distant group of sponge species. To account for this observation, it has been suggested that one or two species of microorganisms associated with the sponges produces these metabolites.7 The chondramides (88-91), recently isolated from cultures of various strains of Chondromyces myxobacteria,12a'b represent the first microbial metabolites in the jaspamide/geodiamolide family. Attempts to isolate the marine microorganisms that are thought to produce the metabolites in the jaspamide/geodiamolide family are currently underway. 3.4 EXPERIMENTAL 127 3.4.1 Materials. Specimens of Cymbastela sp. were collected by hand using SCUBA on outer reefs at -10 to -25 meters near Motupore and Madang in Papua New Guinea. Freshly collected sponge was frozen on site and transported to Vancouver packed in dry ice. Bio-assay guided fractionation of the Cymbastela sp. sponge was performed as follows: freeze-dried sponge (160 g dry weight) was extracted exhaustively with a 1:1 solution of MeOH/CH2Ci2 (4 x 4L). Combined organic extracts were filtered and concentrated in vacuo to yield 14 g of a dark yellow/orange oil. The oil was suspended in 250 mL of a 9:1 solution of MeOH/H20 and extracted with hexanes (3 x 150 mL). The MeOH/H 20 layer was diluted with Ff20 to a 4:1 MeOH/H 20 solution and extracted with CCL (3 x 150 mL). Further dilution of the MeOH/H 20 layer with H 2 0 resulted in a 2:3 MeOH/H 20 solution which was extracted with CHC13 (3 x 150 mL). Al l traces of MeOH were removed in vacuo and the resulting aqueous solution was extracted with EtOAc (3 x 150 mL). Biological testing of the five fractions revealed that the majority of the biological activity was concentrated in the CCI4 and CHCI3 soluble portions which were combined and fractionated further. Separation of the combined CCI4/CHCI3 fraction on Sephadex LH-20 chromatography, eluting with MeOH, yielded two active fractions, one containing a mixture of geodiamolides and the other containing a mixture of peptides as identified by TLC and ! H NMR. Sequential application of the geodiamolide fraction to silica-gel flash column chromatography using a step gradient from 1:1 to 2:1 EtOAc/hexanes and reversed phase HPLC, eluting with 2:3 H20/MeOH, yielded pure samples of geodiamolides H to N (89-95), geodiamolide P (97), and the known metabolites, geodiamolides A to G (77-83). 128 3.4.1.1 Geodiamolide A (77). 77 Compound 77 was isolated as a clear colourless glass (55 mg, 0.02 % dry wt); 'H NMR (500 MHz, CDCI3, 8): 0.88 (d, 3H, J=7Hz), 1.09 (d, 3H, J = 7 Hz), 1.15 (d, 3H, J = 7 Hz), 1.24 (d, 3H, J=6 Hz), 1.35 (d, 3H J = 7 Hz), 1.36 (m, 1H), 1.50 (d, 3U,J= 1 Hz), 1.61 (m, 1H), 2.04 (dd, 1H, J= 14, 4 Hz), 2.16 (d, 1H, J= 16 Hz), 2.16 (m, 1H), 2.32 (m, 1H), 2.91 (dd, 1H, J = 15, 9 Hz), 2.97 (s,3H),3.16(dd, 1H,7 = 15, 8 Hz), 4.48 (dq, 1H, J = 8, 7 Hz), 4.75 (dq, 1H,J = 7, 7 Hz), 4.88 (m, 1H), 4.93 (d, 1H, J = 9 Hz), 5.20 (dd, 1H, J = 9, 8 Hz), 5.44 (s, OH, 1H), 6.47 (d, J = 7 Hz, NH, 1H), 6.54 (d, J = 8 Hz, NH, 1H), 6.89 (d, 1H, J = 8 Hz), 7.07 (dd, 1H, J = 8, 2 Hz), 7.48 (d, 1H, 7 = 2 Hz); exact mass calculated for C 28H4oN 30 6I (M+): 641.1962; found: 641.1965. 3.4.1.2 Geodiamolide B (78). 78 Compound 78 was isolated as a clear colourless glass (36 mg, 0.013 % dry wt); 'H NMR (, 500 MHzCDCl 3 ,8): 0.88 (d, 3H, J= 7 Hz), 1.08 (d, 3H, J = 1 Hz), 1.15 (d, 3H, J = 7 Hz), 1.24 (d, 3H,y=6Hz), 1.35 (d, 3HJ=7Hz) , 1.36 (m, 1H), 1.51 (d,3H,J= 1 Hz), 1.60 (m, 1H), 2.04 129 (dd, IH, J= 14, 4 Hz), 2.16 (d, IH, J= 16 Hz), 2.16 (m, IH), 2.33 (m, 1H),2.92 (dd, IH, J = 15, 9 Hz), 2.97 (s, 3H),3.17(dd, IH, J= 15, 8 Hz), 4.48 (dq, lH,7=8,7Hz) , 4.75 (dq, 1H,J = 7, 7 Hz), 4.88 (m, IH), 4.93 (d, IH, J = 9 Hz), 5.21 (dd, IH, J = 9, 8 Hz), 5.30 (s, OH, IH), 6.47 (d, J = 7Hz, NH, IH), 6.57 (d, J = 8Hz, NH, IH), 6.92 (d, IH, J = 8 Hz), 7.05 (dd, IH, J = 8, 2 Hz), 7.29 (d, IH, J = 2 Hz); exact mass calculated for C^oNsOs^Br/CsgH^NsOe^Br (M+): 595.2080/593.2100; found: 595.2084/593.2091. 3.4.1.3 Geodiamolide C (79). Compound 79 was isolated as a clear colourless glass (12 mg, 0.004 % dry wt); 'H NMR (CDC13, 500 MHz, 8): 0.90 (d, 3H, J = 7 Hz), 1.09 (d, 3H, 7=7 Hz), 1.15 (d, 3H, J = 1 Hz), 1.24 (d, 3H, 7= 6 Hz), 1.35 (d, 3H,7=7Hz), 1.36 (m, IH), 1.51 (d, 3H,J = 1 Hz), 1.60 (m, 1H),2.03 (dd, 1H,7 = 4, 14 Hz), 2.16 (d, 1H,7= 16 Hz), 2.16 (m, lH),2.32(m, 1H),2.92 (dd, 1H,7= 15, 9 Hz), 2.97 (s, 3H), 3.17 (dd, IH, J = 15, 8 Hz), 4.48 (dq, IH, J= 8, 7 Hz,), 4.75 (dq, IH, 7= 7, 7 Hz,), 4.88 (m, IH), 4.93 (d, IH, J= 9 Hz), 5.21 (dd, IH, J= 9, 8 Hz), 5.50 (s, OH, IH), 6.46 (d, J = 6 Hz, NH, IH), 6.56 (d, J = 8Hz, NH, IH), 6.93 (d, IH, J = 8 Hz), 7.01 (dd, IH, J = 8, 2 Hz), 7.15 (d, IH, J = 2Hz); exact mass calculated for C2 8H4oN306 3 7Cl/C28H4oN3063 5Cl (M+): 551.2576/549.2605; found: 551.2562/549.2603. H 79 130 3.4.1.4 Geodiamolide D (80). H 80 Compound 80 was isolated as a clear colourless glass (10 mg, 0.004 % dry wt); see Table 12 for 'H NMR data; exact mass calculated for C 27H38N 30 6I (M+): 627.1805; found: 627.1798. 3.4.1.5 Geodiamolide E (81). 81 Compound 81 was isolated as a clear colourless glass (11 mg, 0.004 % dry wt); *H NMR (CDC13, 500 MHz, 5): 0.90 (d, 3H, J = 7 Hz), 1.16 (d, 3H, J = 7 Hz), 1.25 (d, 3H, J = 6 Hz), 1.30 (d, 3H,J=7Hz), 1.40 (m, IH), 1.54 (d, 3H,7 = 1 Hz), 1.69 (m, IH), 2.10 (m, 2H),2.21 (m, IH), 2.42 (m, IH), 2.83 (dd, IH, J = 14, 7 Hz), 2.97 (s, 3H), 3.26 (dd, IH, J = 14, 9 Hz), 3.77 (dd, IH, J= 18, 4 Hz), 4.16 (dd, 1H,7= 18, 1 Hz), 4.52 (dq, IH, 7=8, 7 Hz), 4.86 (m, IH), 4.99 (d, IH, J = 8 Hz), 5.09 (dd, IH, J =9,1 Hz), 5.46 (s, OH, IH), 6.44 (dd, J = 4, 1 Hz, NH, IH), 6.60 (d, J = 8 Hz, NH, IH), 6.94 (d, IH, J = 8 Hz), 7.06 (dd, IH, J = 8, 2 Hz), 7.32 (d, IH, J = 2Hz) ; exact mass calculated for C27H38N30681Br/C27H38N30679Br (M+): 581.1923/579.1943; found: 581.1909/579.1955. 131 3.4.1.6 Geodiamolide F (82). H 82 Compound 82 was isolated as a clear, colourless glass (2 mg, 0.0007 % dry wt); 'H NMR (CDCI3, 500 MHz, 5): 0.91 (d, 3H, J = 7 Hz), 1.16 (d, 3H, J = 1 Hz), 1.25 (d, 3H, J = 6 Hz), 1.30 (d, 3H,7=7 Hz), 1.40 (m, 1H), 1.50 (d, 3H, J= 1 Hz), 1.69 (m, lH),2.10(m, 2H),2.21 (m, 1H), 2.40 (m, 1H), 2.84 (dd, 1H, 7= 14, 7 Hz), 2.93 (s, 3H), 3.26 (dd, 1H, J = 14, 9 Hz), 3.77 (dd, 1H, J= 18, 4 Hz), 4.16 (dd, 1H, J= 18, 1 Hz), 4.52 (dq, 1H, J=8, 7 Hz), 4.86 (m, 1H), 4.99 (d, 1H, J = 8 Hz), 5.11 (dd, 1H, J=9,7 Hz), 5.30 (s, OH, 1H), 6.44 (dd, J = 4, 1 Hz, NH, 1H), 6.60 (d, J = 8 Hz, NH, 1H), 6.93 (d, 1H,J= 8 Hz), 7.02 (dd, 1H,7 = 8, 2 Hz), 7.18 (d, 1H, J = 2 Hz); exact mass calculated for CiTHagNsO^Cl/Ciy^gNsOe^Cl (M+): 537.2419/535.2449; found: 535.2449/537.2442. 3.4.1.7 Geodiamolide G (83). 83 83 was isolated as a clear colorless glass (2 mg, 0.0007 % dry wt); for lH NMR data, see Table 7; exact mass calculated for C28H38N3O7I (M+): 655.17548; found: 655.17492. 132 3.4.1.8 Geodiamolide H (89). H 89 Compound 89 was isolated as a clear, colourless glass (0.3 mg, 0.00012 % dry wt.); see Table 7 for ! H NMR data; exact mass calculated for CagHsg^OT^Br^grfeNsOT^Br (M+H)+: 610.19509/608.19713; found: 610.19431/608.19508. 3.4.1.9 Geodiamolide I (90). H 90 Compound 90 was isolated as a clear, colourless glass (0.3 mg, 0.00012 % dry wt.); see Table 7 for ! H N M R data; exact mass calculated for C28H38N30737CyC28H38N3C>735Cl (M + ): 565.23688/563.23981; found: 526.23956/563.24011. 133 3.4.1.10 Geodiamolide J (91). 91 Compound 91 was isolated as a clear, colourless glass (2 mg, 0.0008 % dry wt.); see Table 8 for NMR data; exact mass calculated for C28H41N3O7I (M+H)+: 658.19892; found: 658.19767. 3.4.1.11 Geodiamolide K (92). H 92 Compound 92 was isolated as a clear, colourless glass (1.5 mg, 0.0006 % dry wt.); see Table 9 for *H NMR data; exact mass calculated for Cig^oNsCV'Br/CagFLtoNsOT^Br (M+): 611.20294/609.20496; found: 611.20371/609.20519. 134 3.4.1.12 Geodiamolide L (93). H 93 Compound 93 was isolated as a clear, colourless glass (0.8 mg, 0.00032 % dry wt.); see Table 9 for ' H N M R exact mass calculated for C28H4oN30737Cl/C28H4oN30735Cl (M + ): 567.25250/565.25549; found: 567.25184/565.25403. 3.4.1.13 Geodiamolide M (94). OH H 94 Compound 94 was isolated as a clear, colourless glass (1.5 mg, 0.0006 % dry wt.); see Table 10 for NMR data; exact mass calculated for C 2 8 H 4 iN 3 07l (M+H)+: 658.19892; found: 658.19880. 135 3.4.1.14 Geodiamolide N (95). OH H 95 Compound 95 was isolated as a clear, colourless glass (1.5 mg, 0.0006 % dry wt.); see Table 11 for *H NMR data; exact mass calculated for C28H4oN30781Br/C28H4oN30779Br (M+): 611.20294/609.20496; found: 611.20374/609.20515. 3.4.1.15 Geodiamolide P (97). 97 Compound 97 was isolated as a clear, colourless glass (1 mg, 0.0004 % dry wt.); see Table 12 for 'H NMR data; exact mass calculated for C27H38N3O7I (M+): 643.17548; found: 643.17539. 3.4.2 Acid Hydrolysis of Geodiamolides. Pure individual geodiamolides (-0.25 mg each) were dissolved in 1 mL of freshly distilled constant boiling HC1 and the resulting solution was heated at 108° C with stirring for 16 hours in 136 a threaded Pyrex® tube sealed with a Teflon® screw cap. The cooled reaction mixture was evaporated to dryness and traces of HC1 were removed from the residual hydrolyzate by repeated evaporation from H2O (3x3 mL). 3.4.2.1 Derivatization of Amino Acids with Marfey's Reagent and H P L C Analysis.1 5 To a 1 mL vial containing 2 mmol of pure amino acid standard in 80 pL of H 2 O was added 2.8 mmol of iV-a-(2,4-dinitro-5-fluorophenyl)-Z-alaninamide (FDAA) in 170 pL of acetone followed by 20 pL of 1 N NaHCC»3. The mixture was heated for 1 hour at 40° C. After cooling to room temperature, 10 pL of 2 N HC1 were added and the resulting solution was filtered through a 4.5 pm filter and stored in the dark until HPLC analysis. To prepare FDAA derivatives of the amino acids in the hydrolyzate of the ozonolysis product of the geodiamolides, a 90 uL aliquot containing 0.9 mg of amino acid mixture was reacted with 2.86 mmol of FDAA in 172 uL of acetone as described above. A 10 pL aliquot of the resulting mixture of FDAA derivatives was analyzed by reversed-phase HPLC. A linear gradient of (A) 9:1 triethylammonium phosphate (50 mM, pH 3.0)/MeCN and (B) MeCN, with 0% B at the start to 40% B over 60 minutes (flow rate 1 mL/min) was used to separate the FDAA derivatives, which were detected by UV at 340 nm. Each peak in the chromatographic trace was identified by comparing its retention time with that of the FDAA derivative of the pure amino acid standard and by co-injection. In all cases a peak was observed at -35.7 minutes which was attributed to excess FDAA. The HPLC retention times are tabulated below: Table 13. HPLC retention times for serine Marfey's derivatives in minutes. Z-ser DIL-stx FDAA standards 26.0 L25.9 D26.6 35.8 geodiamolide J (91) 25.9 35.7 geodiamolide M (94) 25.9 35.8 137 1.1. ENDNOTES: CHAPTER 3: GEODIAMOLIDES FROM CYMBASTELA SP. 1) Fusetani, N. ; Matsunaga, S. Chem. Rev. 1993, 93, 1793. 2) Crews, P.; Manes, L. V.; Boehler, M . Tetrahedron Lett. 1986, 27, 2797. 3) Zabriskie, M . T.; Klocke, J. A.; Ireland, C. M. ; Marcus, A. H.; Molinski, T. F.; Faulkner, D. J.; Xu, C ; Clardy, J. C. J. A. Chem. Soc. 1986,108, 3123. 4) Chan, W. R.; Tinto, W. F.; Manchand, P. S.; Todaro, L. J. J. Org. Chem. 1987, 52, 3091. 5) de Silva, E. D.; Andersen, R. J.; Allen, T. M . Tetrahedron Lett. 1990, 31, 489. 6) Coleman, J. E.; de Silva, E. D.; Kong, F.; Andersen, R. J.; Allen, T. M . Tetrahedron 1995, 51, 10653. 7) Crews, P.; Farias, J., J.; Emrich, R.; Keifer, P., A. J. Org. Chem. 1994, 59, 2932. 8) Talpir, R.; Benayahu, Y.; Kashman, Y.; Pannell, L.; Schleyer, M . Tetrahedron Lett. 1994, 55,4453. 9) D'Auria, M . V.; Paloma, L. G.; Minale, L.; Zampella, A. J. Nat. Prod. 1995, 55, 121. 10) Senderowicz, A. M . J.; Kaur, G.; Sainz, E.; Laing, C ; Inman, W. D.; Rodriguez, J.; Crews, P.; Malspeis, L.; Grever, M . R.; Sausville, E. A.; Duncan, K. L. K. J. Natl. Cancer Inst. 1995, 87, 46. 11) de Vos, L.; Rutzler, K.; Boury-Esnault, N . ; Donadey, C ; Vacelet, J. Atlas of Sponge Morphology; Smithsonian Institution Press: Washington, 1991, p 57 . 12a) Kunze, B.; Jansen, R.; Sasse, F.; Hofle, G.; Reichenbach, H. J. Anitbiotics 1995, 48, 1262. 12b) Jansen, R.; Kunze, B.; Reichenbach, H.; Hofle, G. Liebigs Ann. 1996, 285. 13) Stingl, J.; Andersen, R. J.; Emerman, J. T. Cancer Chemother. Pharmacol. 1992, 30, 401. 14) Identified by Dr. R. W. M . van Soest, University of Amsterdam. A voucher sample has been deposited at the Zoological Museum of Amsterdam. 15) Marfey, P. Carlsberg. Res. Commun 1984, 49, 591. 138 4 P E P T I D E S F R O M T H E M A R I N E SPONGE CYMBASTELA SP. 4.1 INTRODUCTION Cancer is a dreaded disease. In fact, cancer, which may arise from any type of cell and in any body tissue, is not one disease, but a large number of diseases classified according to the tissue and type of cell of origin.1 A cancerous growth, or neoplasm, is clonal; that is, all its cells are descendants of a single 'transformed' cell. 2 a Because these transformed cells have escaped the control of the normal forces regulating cellular growth, they are allowed to reproduce endlessly. As they multiply, these cells may form a mass called a tumor, which enlarges and continues to grow without regard to the function of the tissue of origin.23 Almost all cancers form tumors, but not all tumors are cancerous, or malignant. The greatest number of tumors are benign and are characterized by entirely localized growth and are usually separated from neighboring tissue by a surrounding capsule.1 Malignant tumors, on the other hand, are able to spread beyond their site of origin 2 3 Cancer may invade neighboring tissues by direct infiltration or it may disseminate to distant sites, forming secondary growths known as metastases.23 Effective treatments for malignant cancers must control the invasion and spread of the disease. If this can be accomplished, patient survival rates are dramatically increased.23 Surgery, radiotherapy, and chemotherapy are the three forms of cancer treatment generally employed. Surgery and radiotherapy can sometimes eradicate primary or localized disease, but often fail to stop cancer that has metastasized to other parts of the body.23 Chemotherapy, in conjunction 139 with surgery, radiotherapy, or both, may be able to control metastatic growth and prolong life.2 a Chemotherapy is described as the use of drugs in the treatment of cancer. Because a drug is distributed throughout a patient's body by the blood stream, chemotherapy plays a useful role in the treatment of tumors that have spread beyond the area accessible by surgery and radiotherapy.23 There is a growing number of different types of anticancer drugs available, varying greatly in structure and mechanism of action. Most of these drugs share one common characteristic: they are effective because they interfere with an aspect of the cell cycle.23 For example, antimetabolites such as Ara-C (4) and fluorouracil (111) interfere with the production of nucleotides, thus inhibiting DNA synthesis.2b Covalent DNA-binding drugs, such as cisplatin (112), bind to the DNA and decrease its ability to act as a template for DNA synthesis.20 Antimitotics, such as vincristine (113) and paclitaxel (Taxol®, 114), exert their cytotoxic effects by interfering with microtubule formation during mitosis.2d Cancerous tissues, in general, have a larger proportion of dividing cells than normal tissues, and thus these rapidly dividing cells are more sensitive to chemotherapy than normal cells. Unfortunately, some of the body's normal cells also have a very high proliferation rate, including bone marrow and the lining cells of the gastrointestinal tract.23 These two sites are generally the most sensitive to chemotherapeutic side-effects and constitute the sites of toxicity that limit the tolerable dose of most drugs. Anticancer drugs are most useful in the treatment of high growth factor leukemias and lymphomas, since these cancers have very high cell-proliferation rates.23 Unfortunately, most solid tumors (e.g. colon, rectum, lung, breast) have a lower proportion of dividing cells, and are therefore less susceptible to treatment by drugs alone.23 Despite this fact, preliminary drug treatment may control metastatic tumor spread, shrink the local tumor, or slow the spread of the disease while waiting to remove or eradicate the primary source. 140 O 114 A major problem in the use of anticancer drugs is acquired drug resistance. This is the most common reason for the failure of drug treatment in cancer patients with initially sensitive tumors.26 As a result, drugs are now usually given in combination therapy. A lower proportion of tumor cells will be resistant to therapy if two drugs with different mechanisms of action are administered together.26 If the cancer cell is able to mutate at a resistance rate of 10"6 for one drug, and at a resistance rate of 10"6 for the second drug, then its rate of resistance to both will be very small (10" ). As a cell builds up a resistance to the first drug, it can be killed by the second drug, and vice versa. 141 4 .1 .1 Mitosis The cycle of a rapidly dividing human cell is about 24 hours in duration, and can be divided between interphase and mitosis.3 Interphase accounts for the majority of the cycle time and consists of the DNA synthesis phase, called S-phase, which is preceded by a gap called GI, and followed by a gap named G2 (Figure 63). In a typical animal cell cycle, GI lasts for 12 hours, S-phase for 6 hours, G2 for 6 hours, and mitosis occurs over approximately 30 minutes.3 Mitosis is the division of a cell to form two daughter cells having identical genetic information. The stages of mitosis are outlined in Figure 64. During interphase, the cell duplicates its single mitotic organizational center (MTOC) (also called spindle poles or asters) and microtubules begin to radiate from each MTOC. As the cell develops towards prophase, the chromosomes condense into distinguishable sister chromatids and the spindle poles migrate to opposite ends of the cell. The end of prophase is signified by the breakdown of the nuclear envelope. In metaphase, the microtubule network becomes organized into the mitotic spindle. Sister chromatids are aligned midway between the spindle poles with each of the chromatids' kinetochores, the point of linkage between two daughter chromatids, attached to microtubules originating from opposite poles. After a brief pause at the end of metaphase, anaphase begins with the kinetochores dissolving and each separate chromosome moving along the microtubules to opposite poles of the spindle. As the chromosomes near the poles, cytokinesis, the process of cell division, begins. As cytokinesis continues, the cells get pinched into two daughter cells, the chromosomes decondense, a new nuclear envelope is formed in each, and the two new daughter cells enter into interphase. 142 Mitosis Interphase Figure 63. Stages of the cell cycle.3 Almost all antimitotic agents have been found to alter microtubule assembly reactions by interacting with the protein tubulin during mitosis.4 Microtubules are a ubiquitous component of the cytoskeleton of eucaryotic cells involved in a variety of cellular processes including mitosis, intracellular transport, and cellular mobility.5 Microtubules are polymers of tubulin, a dimer consisting of two -50 kDa subunits (a- and P-tubulin) with two tightly bound molecules of guanosine nucleotide. The bound nucleotide of P-tubulin is readily exchangeable, whereas the bound nucleotide of a-tubulin is not.4 Tubulin dimers are asymmetric molecules, and when they are polymerized to form microtubules, the resultant structures are polar, with distinct plus and minus ends. In the spindle, the microtubules are nucleated with the minus ends of tubulin at the centrosome and polymerization/depolymerization occurs at the plus ends.3 Each microtubule consists of 13 individual strings of tubulin dimer, which are aligned side by side to form a 25 nm diameter tube.6 143 Interphase Figure 64. Stages of mitosis. Microtubules are continuously undergoing rapid assembly and disassembly, a behavior called dynamic instability. The hydrolysis of bound guanosine triphosphate (GTP) to guanosine diphosphate (GDP) controls this process. GTP-tubulin adds to the plus end of the polymer, where it can later be hydrolyzed to GDP. A GDP-tubulin unit located within a microtubule stays 'locked' in position, but one situated on an exposed end dissociates from the polymer.6 Hence, microtubules become longer when the rate of addition of GTP-tubulin is faster than the rate of hydrolysis of bound GTP. 144 organizing center microtubule growing polymer shrinking polymer free tubulin dimers Figure 65. Addition of tubulin dimers to a microtubule. 4.1.2 A n t i m i t o t i c N a t u r a l P r o d u c t s A summary of antimitotic agents would be incomplete without first describing the mechanism of action of colchicine (115), the vinca alkaloids vincristine (113) and vinblastine (116), and paclitaxel (114). All four of these compounds have been well studied and are often used for comparison purposes when new compounds with potential as antimitotic agents are being investigated. 4.1.2.1 Colchicine (115) MeO MeO 1 'NHCOCH 3 OMe 115 Colchicine (115) is an alkaloid obtained from the terrestrial plant Colchicum autumnaleJ 145 Colchine's antimitotic activity has been known since the end of the 19th century, and it is now the most studied of all antimotic agents.8 Colchicine (115) reacts negligibly with tubulin at 0°C, but binds extremely tightly to tubulin at physiological temperatures.9 Upon binding, colchicine (115) induces hydrolysis of GTP bound in the exchangeable site, resulting in complete inhibition of tubulin assembly.9 This inhibition is observed at substoichiometric concentrations.9 4.1.2.2 Vinca alkaloids (113,116) R HO COOCH 3 113 R = C H 3 116 R = CHO Vincristine (113) and vinblastine (116) are two closely related alkaloids isolated from the periwinkle plant Catharanthus rosea (also called Vinca rosea).10 Vincristine (113) is a widely used microtubule antagonist. Some of the cancers currently treated with 113 include acute leukemias, Hodgkin's and non-Hodgkin's lymphoma, small-cell lung cancer, breast cancer, and Wilm's tumor.2d Vinblastine (116), although having found less applicability than 113 in chemotherapy, is used for treatment of Hodgkin's and non-Hodgkin's lymphoma, testicular cancer, and ovarian cancer.2d Both of these vinca alkaloids (113,116) effect their cytotoxicity by binding to free 146 tubulin dimers and disrupting the sensitive balance between microtubule polymerization and depolymerization by decreasing the pool of free tubulin dimers available for microtubule assembly.2d This results in net dissolution of microtubules, leading to destruction of the mitotic spindle and arrest of the cells in the metaphase of mitosis. When the vinca alkaloids (113,116) are reacted with an equimolar amount of tubulin, they cause the formation of paracrystalline aggregates, shifting the equilibrium even further towards disassembly (Figure 66).2d shrinking free vinca-bound paracrystalline polymers tubulin tubulin dimers aggregates dimers Figure 66. Disruption of tubulin/microtubule equilibrium by vinca alkaloids.2*1 Vinblastine (116) strongly inhibits tubulin-dependent GTP hydrolysis and weakly inhibits binding of GDP and GTP at the exchangeable site, but does not displace nucleotides bound in the exchangeable site.9 Competitive inhibition experiments using radiolabeled vinblastine (116) indicated that colchicine (115) binds in an uncompetitive manner to tubulin with respect to vinblastine (116).9 The vinca alkaloids and colchicine (115) both inhibit tubulin polymerization by binding to tubulin, but do not share the same binding site.2d 147 4.1.2.3 Paclitaxel (114) 114 Paclitaxel (Taxol®, 114), a complex, highly functionalized diterpene, was first isolated 1 1 1 9 1 ^ 1 ^ from the Pacific yew tree Taxus brevifolia. ' Paclitaxel (114) has now been synthesized and has also been isolated from the common garden yew T. baccata.16 Although its antitumor 19 activity has been known since 1971, its mechanism of action was not known until 1979, when Horwitz determined that it enhances microtubule stability and assembly rather than inhibiting it. Experiments using radio labeled paclitaxel (114) demonstrated that the drug binds reversibly to tubulin in polymerized microtubules, but not to free tubulin.9 Binding is inhibited by other antimitotics, such as vincristine (113), but the results are difficult to interpret since paclitaxel (114) binds to polymerized tubulin and vincristine (113) binds to free tubulin.9 Up until 1984, all antimitotic agents were derived from plants.9 Since then, many new agents have been isolated from a variety of organisms including marine invertebrates. The following is a brief summary of known marine antimitotic natural products. 148 4.1.2.4 Curacin A (117) 117 Curacin A (117) was isolated from the marine cyanobacterium Lynbya majuscula.18'19 It binds rapidly and tightly with tubulin and induces GTP hydrolysis.9 Curacin A (117) has been shown to competitively inhibit the binding of colchicine (115) to tubulin 9 4.1.2.5 Dolastatin 10 (118) Several unusual cytotoxic peptides have been isolated from the sea hare Dolabella auricularia, of which dolastatin 10 (118) is the most potent.20 Dolastatin 10 (118) binds reversibly to tubulin causing mitotic arrest by strongly inhibiting tubulin polymerization.4 It non-competitively inhibits the binding of vincristine (113) to tubulin, stabilizes colchicine (115) binding to tubulin, inhibits nucleotide exchange, and tubulin-dependent GTP hydrolysis.4 Dolastatin 10 (118) also inhibits the binding of radiolabeled phomopsin A (119), a terrestrial peptide antimitotic, to tubulin 4 4.1.2.6 Halichondrin B (120) C H 3 C H 3 120 The polyether macrolide halichondrin B (120) is one of several related compounds that have been isolated from a number of sponge species, the first being Halichondria okadai. Halichondrin B (120) is highly cytotoxic and inhibits tubulin assembly, tubulin-dependent GTP hydrolysis, and nucleotide exchange. It is a noncompetitive inhibitor of the binding of radiolabeled vinblastine (116) to tubulin, but does not interfere with the binding of colchicine (115).22 4.1.2.7 Spongistatins (121-129) 150 Hii.. >OH Hi., .OH 0CH3 121 spongistatin 1 R = Cl, Rj = R 2 = C O C H 3 125 spongistatin 5 R = Cl, R} = H 122 spongistatin 2 R = H, R, = R 2 = C O C H 3 127 spongistatin 7 R = H, R, = H 123 spongistatin 3 R = Cl, R\ = H, R 2 = C O C H 3 128 spongistatin 8 R = H, Rx = C O C H 3 124 spongistatin 4 R = Cl, R, = C O C H 3 , R 2 = H 129 spongistatin 9 R = Cl , Rj = C O C H 3 126 spongistatin 6 R = Cl, R, = C O C H 3 , R 2 = C O C H 3 Nine highly cytotoxic polyether macrolides, called spongistatins 1-9 (121-129), have been isolated from the marine sponges Spongia sp. 2 3 and Spirastrella spinispirulifera.24'26 Spongistatin 1 (121) is the most abundant and has been studied in the greatest detail. Hamel has described 121 as the most potent cytotoxic antimitotic drug ever tested in his laboratory, having an IC50 value of 20 p M against L1210 murine leukemia cells. 9 Spongistatin 1 (121) inhibits microtubule assembly, as well as nucleotide exchange, and in competitive binding studies 121 has been shown to inhibit both the binding of radiolabeled vinblastine (116) and dolastatin 10 (118) in a noncompetitive manner 27 151 4.1.2.8 Cryptophycins (130-133) 130 cryptophycin-1: R = C1, Z-O-mefhyltyrosine 131 cryptophycin-2: R = H, D-O-methyltyrosine 132 cryptophycin-3: R = C1, Z-O-methyltyrosine 133 cryptophycin-4: R = H, Z)-0-methyltyrosine The cyclic depsipeptide cryptophycin-1 (130) (originally called cryptophycin) was first isolated from the blue-green alga (cyanobacterium) Nostoc sp. In addition to 130, 24 other analogues have been isolated from the alga (cryptophycins-1 to 4 (131-133) shown).28"30 Cryptophycin-1 (130) has demonstrated potent cytotoxicity, both in vitro and in vivo.2 8 Initial work on cryptophycin-1 (130) indicates that it inhibits tubulin polymerization substoichiometrically and inhibits vinblastine (116) binding to tubulin.31 Hamel has also demonstrated that cryptophycin-1 (130) inhibits the binding of vincristine (113) and dolastatin 10 (118) to tubulin.9 152 4.1.2.9 Discodermolide (134) Gunasakera et al. isolated the lactone discodermolide (134) from the marine sponge Discodermia dissoluta. The mechanism by which discodermolide (134) effects its cytotoxicity is by stabilizing microtubule formation, which is similar to paclitaxel (114).33 Discodermolide (134) was found to be 7 times more active in stabilizing tubulin than paclitaxel (114).9 4.1.2.10 Eleutherobin (135) OH 135 The most recently identified marine antimitotic is eleutherobin (135), a diterpene glycoside isolated from the marine soft-coral Eleutherobia sp.34 Eleutherobin (135) was shown to be a potent cytotoxin and has been found to stabilize microtubules by competing for the paclitaxel (114) binding site on the microtubule polymer.34 4.1.3 The vinca domain 153 As can be seen from the compounds described above, there are a wide range of structures that interfere with microtubule dynamics by binding to tubulin. Hamel has proposed a model for the area on the (3-unit of tubulin were these drugs are thought to bind, called the "Vinca Domain" (Figure 67).4 The model rationalizes the observed inhibitory data and proposes that there are at least three distinct drug binding regions within the vinca domain: the vinca region, the peptide groove, and the polyether region.4'9 To account for dolastatin 10's (118) noncompetitive inhibition of the vinca alkaloid's binding to tubulin and inhibition of nucleotide exchange, an area called the peptide groove has been placed in close proximity to both the vinca binding region and the exchangeable GTP site 4 The noncompetitive inhibition of vinblastine (116), dolastatin 10 (118), and the strong inhibition of nucleotide exchange by spongistatin 1 (121) implied there is a third 154 distinct binding site within the vinca domain, termed the polyether site.9 The model places the binding regions close enough to each other that a drug bound in one region can sterically interfere with the binding of a drug to an adjacent region or with nucleotide exchange.4 4.1.4 Cytotoxic peptides Our investigation of the hemiasterlins and criamides began serendipitously. During the re-isolation of geodiamolides A to G (77-83) from the Cymbastela sp. sponge for further biological evaluation, a second potently cytotoxic fraction was found. Bio-assay guided fractionation of this cytotoxic fraction led to the isolation of the tripeptides hemiasterlin (86), hemiasterlin A (98), and hemiasterlin B (99), and the tetrapeptides criamide A (100) and criamide B (101). One of these peptides, hemiasterlin (86), had been previously isolated from the marine sponge Hemiasterella minor collected in South Africa,35 and a related compound, milnamide A (84), was reported from the sponge Auletta cf. constricta collected in Papua New Guinea. The original structure determinations of 84 and 86 did not give any stereochemical information about these compounds. The following section describes the structure determination of the new compounds hemiasterlin A (98), hemiasterlin B (99), criamide A (100), and criamide B (101), as well as the determination of the absolute stereochemistry of hemiasterlin (86). The biological activity of the series is discussed. 155 4.2 RESULTS AND DISCUSSION Specimens of Cymbastela sp. were collected by hand using SCUBA on outer reefs at -10 to -25 meters near Motupore and Madang in Papua New Guinea (see Collection Map, Chapter 3, Figure 49). Freshly collected sponge was frozen on site and transported to Vancouver packed in dry ice. Bio-assay guided fractionation of the Cymbastela sp. sponge was performed as outlined below. Freeze dried sponge (160 g dry weight) was extracted exhaustively with a 1:1 solution of CH 2Cl 2/MeOH (4 x 4L). The combined organic extracts were filtered and concentrated in vacuo to yield 14 g of a dark yellow/orange oil which was suspended in 250 mL of a 9:1 solution of MeOH/H 20 and extracted with hexanes (3 x 150 mL). The MeOH/H 20 layer was diluted with H 2 0 to a 4:1 MeOH/H 20 solution and extracted with CC14 (3x150 mL). Further dilution of the MeOH/H 20 layer with H 2 0 resulted in a 2:3 MeOH/H 20 solution which was extracted with CHCI3 (3 x 150 mL). Al l traces of MeOH were removed in vacuo and the resulting aqueous solution was extracted with EtOAc (3 x 150 mL). Biological testing of the five fractions revealed that the majority of the biological activity was concentrated in the CCI4 and C H C I 3 soluble portions, which were combined and fractionated further. Separation of the combined CCI4 and CHCI3 fractions by Sephadex LH-20 chromatography, eluting with MeOH, yielded two active fractions, one containing a mixture of geodiamolides* and the other containing mixture of peptides, as identified by TLC and ! H NMR spectroscopy. Further purification of the peptide fraction by gradient reversed phase flash chromatography (H 20 to H 20/MeOH 30:70) followed by reversed phase isocratic HPLC (0.05% TFA/H 20:MeOH 1:1) afforded pure hemiasterlin (86) (40 mg, 0.015 % dry wt), hemiasterlin A (98) (32 mg, 0.012 % dry wt), hemiasterlin B (99) (1 mg, 0.0004 % dry wt), criamide A (100) (3.3 mg, 0.0013 % dry wt) and criamide B (101) (3.3 mg, 0.0013 % dry weight). The structure of hemiasterlin was determined by a comparison * see Chapter 3 for a discussion of the geodiamolides 156 of its spectroscopic data with the literature values, and its absolute configuration was determined through a combination of degradative studies and single crystal X-ray diffraction analysis. The structures of hemiasterlin A and B (98, 99) and criamide A and B (100,101) were determined using spectroscopic methods, particularly one and two dimensional NMR and mass spectrometry. Copies of l 3 C NMR spectra and two dimensional NMR spectra (COSY, HMBC, and HMQC) for compounds 98 to 101 not included in the main body of the thesis can be found in Appendix A. 157 4.2.1 The hemiasterlins 4.2.1.1 Hemiasterlin (86) 23 86 Hemiasterlin (86) was isolated as an amorphous white solid which gave an (M+H)+ ion in the HRFABMS at m/z 527.3594, appropriate for a molecular formula of C30H46N4O4, requiring 10 sites of unsaturation. Comparison of the NMR data obtained for hemiasterlin (86) isolated from Cymbastela sp. with the literature values for hemiasterlin (86) isolated from H. minor35 indicated that the two compounds have identical chemical constitutions (Table 14). The discrepancy in the chemical shift values of all of the carbons, except C l 1, C12, C17, and C29, is due to the use of two different NMR solvents (CDCI3 3 5 and DMSO-d 6 3 7)- The differences in the chemical shifts values for the resonances assigned to C l 1, C12, C17, and C29 are too large to result from different NMR solvents, but may be explained by the different isolation procedures used in each case. Hemiasterlin (86), isolated from H. minor, was purified by reversed phase HPLC (MeOH/H20, 90:10),35 yielding 86 as an 'uncharged' peptide. While hemiasterlin (86) obtained from Cymbastela sp. was also purified by reversed phase HPLC, the solvent system used was 0.05% TFA/H 20/MeOH 1:1, resulting in the isolation of 86 as the TFA salt (Figure 70). Due to the amphoteric character of amino acids, their chemical shifts are strongly pH-dependant. It has been observed that the transition from an amine to a protonated amine causes Table 14. ' H and 1 3 C NMR data for hemiasterlin (86V (Recorded at 500 MHz.) hemiasterlin (86) (DMSO-d6)a hemiasterlin (86) (CDCl 3) b Carbon no. 5 1 3 C 5 'H 5 1 3 C 5 'H 2 128.7 7.16 s 127.5 6.86 s 3 116.5 121.1 4 125.0 127.0 5 120.6 8.09 d 7= 8 Hz 121.2 7.90 d J = 8 Hz 6 118.4 7.07tJ=8Hz 119.1 7.08tJ=8Hz 7 121.1 7.20t7=8Hz 122.3 7.22 t J= 8 Hz 8 110.0 7.44dJ=8Hz 109.5 7.29dJ=8Hz 9 137.7 139.2 10 37.5 38.4 11 67.5 4.44dJ=6Hz 73.2 3.60 s 12 166.0 173.1 13 32.4 3.75 s 33.1 3.75 s 14 27.0 1.41 s 28.2 1.60 s 15 22.5 1.38 s 23.5 1.44 s 16-N 7.38 br s;8.85 br s 17 33.4 2.24 s 36.0 2.00 s 18-N 8.87 s 7.90dJ=8Hz 19 56.2 4.84d7=8Hz 55.2 4.88 d J = 8 Hz 20 170.1 172.5 21 34.6 35.5 22 26.3 0.99 s 27.2 1.00 s 23 26.3 0.99 s 27.2 1.00 s 24 26.3 0.99 s 27.2 1.00 s 26 55.6 4.93 1 7 = 10 Hz 56.3 5.111 J= 10Hz 27 138.3 6.66 d 7 = 10 Hz 140.1 6.73 d J = 10 Hz 28 131.6 131.1 29 168.5 172.1 30 31.1 3.03 s 31.4 3.06 s 31 28.7 2.01 m 30.3 1.86 m 32 19.3 0.80 d7= 7 Hz 19.5 0.79d7=6Hz 33 18.9 0.78d7=7Hz 19.0 0.86dJ=6Hz 34 13.5 1.80 s 14.6 1.90 s a N M R data from hemiasterlin sample isolated from Cymbastela sp. b N M R data from hemiasterlin sample isolated from Hemiasterlla minor: 1 "X an increase in the shielding of the a-carbon and the carboxyl C nuclei. Similarly, the transition from the acid to the carboxylate ion causes a reduction in the shielding of these two 1 3 C nuclei.38 For example, the effects that pH have on the chemical shifts of the carbon atoms in glycine are summarized in Table 15. 161 86 from Cymbastela sp. 86 from H. minor' Figure 70. Chemical shift differences in hemiasterlin (86). Table 15. Effect of pH on the C chemical shifts glycine. Compound pD sec1) 5(C2) H 2N-CH 2-COOH 0.45a 171.2 41.5 12.05a 182.7 46.0 Solvent: D 2 0 Hemiasterlin (86) isolated from H. minor was reported to have an [OC]D = -95°, while the sample obtained from Cymbastela sp. was found to have an [a]o25 = -76°. 3 7 Since the specific rotations for the two samples of hemiasterlin (86) have the same sign, and their NMR data indicate that the two compounds have identical constitutions and relative configurations, we have assumed that the two samples have the same absolute configuration. The observed differences in the magnitude of the specific rotations are assumed to be the result of measurement errors. Hemiasterlin contains three non-protein amino acids: terr-leucine (pseudoisoleucine) (136), 7vT-methyl-4-amino-2,5-dimethylhex-2-enoic acid (137), and N.N'.fifi-tetramethyltryptophan (138). tert-Leucine (136) was first discovered in the peptide bottromycin (139), the active component of an extract of the terrestrial Streptomyces bottropensis bacteria.39 It was subsequently discovered in the marine environment in the cyclic depsipeptides discodermins A to C (140-142), isolated from the sponge Discodermia kiiensis,40'42 in the cyclic 162 depsipeptide polydiscamide A (143), isolated from the Caribbean sponge Discodermia sp., and in milnamide A (84), a tripeptide isolated from the sponge Auletta cf. constricta?6 The y-amino acid /V-methyl 4-amino-2,5-dimethylhex-2-enoic acid (137) is a constituent of the tripeptide milnamide A (84), and was given the trivial name of TV-methyl homo vinylogous valine (MHVV). 3 6 A^/j^Tetramethyltryptophan (138) is unique to hemiasterlin, although milnamide A (84) incorporates a closely related N,N',/?,/?-tetramethylated P-carboline.36 Hemiasterlin has recently been isolated from several other sponge species.44 The original structure elucidation of hemiasterlin (86) did not provide any information concerning the absolute configuration of the component amino acids.35 To be thorough, and in anticipation of undertaking a total synthesis of hemiasterlin (86) and related analogs, the absolute stereochemistry of hemiasterlin (86) was determined by a combination of chemical degradation and single crystal X-ray diffraction. 136 137 138 139 Ozonolysis followed by oxidative work-up using hydrogen peroxide converted the 163 MHVV residue in 8 6 to an A -^methyl valine residue. Hydrolysis of the ozonolysis product with 6 N HC1 at 108°C for 16 hours afforded the free aliphatic amino acids (Scheme 5 ) . Unfortunately, the indole ring of the tryptophan moiety was destroyed during the acid hydrolysis. Derivatization of the hydrozylate and of the amino acid standards with Marfey's reagent,45 followed by HPLC analysis, showed that the terMeucine and the Af-mefhyl valine residues in 8 6 both have L configurations. o NH, . S03 R2 ° O ^ N - " O N U I "NH 140 discodermin A Rj = R 2 = Me 141 discodermin B Rj = H R 2 = Me 142 discodermin C Rj = Me R 2 = H ~H OH NH, o NH, U = H U U = H U = H U = H [ | 'NH a-O N J o H 2 N ^ N H NH2 143 164 1. 0 3 , MeOH,-78°C 2. 30% H 2 0 2 rt, 40 min Scheme 5. Chemical degradation of hemiasterlin (86). Attempts to determine the configuration of the A^A /^2/2-tetramethylated tryptophan residue in hemiasterlin (86) by means of CD analysis using models of the tryptophan moiety, proved to be inconclusive. Fortunately, hemiasterlin methyl ester (144), formed by treating 86 with diazomethane, gave crystals suitable for single crystal X-ray diffraction analysis. The results of the X-ray diffraction analysis, based on the knowledge that both the fert-leucine and MHVV residues in hemiasterlin (86) have the L configuration, indicated that the tetramethylated tryptophan residue also has the L configuration. The structure was shown to be stabilized by one intramolecular hydrogen bond, between NH18 and N16, and one intermolecular hydrogen bond, between NH16 and C=0 (C34), which links adjacent molecules together in a head-to-tail chain. 165 C32 C23 Figure 71. Computer generated ORTEP drawing of hemiasterlin methyl ester (144). Talpir et al. reported that hemiasterlin (86), isolated from H. minor, showed only moderate cytotoxicity in vitro against murine leukemia P388 cells (ED5o = 0.01 u.g/mL).35 It was suggested that the observed activity was the result of small traces of jaspamide (76), a highly cytotoxic co-occurring metabolite, in the test sample.35 However, hemiasterlin (86) obtained from Cymbastela sp. was found to be extremely cytotoxic in vitro against murine leukemia P388 cells (ED50 = 4.57 x 10"5 u.g/mL).37 A detailed discussion of the biological activity of the hemiasterlins can be found in Section 4.2.3, entitled: Biological activity of the hemiasterlins and criamides. 166 4.2.1.2 Hemiasterlin A (98) 6 23 22 | 24 o 32 33 2 9 OH 98 Hemiasterlin A (98) was isolated as an amorphous white solid which exihibits an (M+H)+ ion in the HRFABMS at m/z 513.3471, appropriate for a molecular formula of C29H44N4O4, requiring 10 degrees of unsaturation. Table 16 summarizes the NMR data obtained for 98. The 1 3 C NMR signals of 98 are well dispersed and the HMQC data revealed that 40 of the 44 hydrogens in the molecule are attached to carbon (10 CH, 10 CH3). Three 1 3 C NMR resonances at 8 166.0, 170.3, and 168.7 were assigned to carbonyl carbons and ten deshielded resonances at 8 138.4, 137.7, 131.9, 124.8, 124.4, 121.1, 120.4, 118.4, 117.2, and 112.1 were identified as sp1 hybridized carbons. No additional unsaturated functionalities were apparent in the molecule; therefore the remaining two sites of unsaturation must be rings. A number of features of the NMR data obtained for 98 suggested that it is very closely related to the co-occurring tripeptide hemiasterlin (86). An indole ring was identified by eight characteristic indole 1 3 C NMR resonances (8 124.4 (C2), 8 117.2 (C3), 8 124.8 (C4); 120.4 (C5), 118.4 (C6), 121.1 (C7), 112.1 (C8), and 8 137.7 (C9)) and the XH NMR resonances assigned to the five attached protons ((8 7.15 (d, J = 2 Hz: H2), 8.08 (d, J = 8 Hz: H5), 7.02 (t, J = 8 Hz: H6), 7.11 (t, J = 8 Hz: H7), and 7.40 (d, J = 8 Hz: H8)). An isolated spin system in the COSY spectrum, which continued uninterrupted from H5 (8 8.08) through H6 and H7 (8 7.02, 7.11) and ended at H8 (8 7.40), confirmed the C5/C6/C7/C8 connectivity of the indole ring. 167 168 Table 16. NMR data for hemiasterlin A (98). (Recorded in 1 DMSO-d6 at 500MHz.) Carbon no. 8 ° C 8*H C O S Y H M B C 8 1-N 11.20 s 2 124.4 7.15dJ=2Hz 3 117.2 HI, 2, 14 4 124.8 HI, 2, 6,8 5 120.4 8.08dJ=8Hz H6 H7 6 118.4 7.02tJ=8Hz H5, 7 H8 7 121.1 7.111 J= 8Hz H6, 8 H5, 8 8 112.1 7.40 d J = 8 Hz H7 H6 9 137.7 HI, 2, 5, 7 10 37.7 H l l , 14, 15 11 67.6 4.48 d 7 = 10 Hz H16 H14 12 166.0 H l l , 18, 19 14 27.2 1.42 s H15 15 22.6 1.40 s H l l , 14 16-N 7.38 br s H16', 11, 17 16-N' 8.80 br s H16, 17 17 33.6 2.23t7=5Hz H16, 16' H l l 18-N 8.90 br d J = 9 Hz H19 19 55.7 4.85d7=9Hz H18 H22, 23, 24 20 170.3 H19, 26,30 21 34.8 H22, 23, 24 22 26.5 1.00 s H23, 24 23 26.5 1.00 s H22, 24 24 26.5 1.00 s H22, 23 26 56.5 4.9417 = 10 Hz H27, 31 H32, 33 27 138.4 6.67 d 7 = 10 Hz H26 H26, 34 28 131.9 H26, 34 29 168.7 H27, 34 30 31.3 3.03 s H26 31 28.9 2.01 m H26, 32, 33 H32, 33 32 19.4 0.81 d J = 7 H z H31 H33 33 19.0 0.79dJ=7Hz H31 H32 34 13.6 1.80 s H27 Proton resonances that are correlated to the carbon resonance in the 8 1 3 C column. The main difference between the 'H NMR spectrum of hemiasterlin (86) (Figure 68) and the lH NMR spectrum of 98 (Figure 72) is the absence of the indole JV-methyl substituent in 98. Thus, the resonance at 8 3.75 (Mel3) in the 'H NMR spectrum of hemiasterlin (86), assigned to the indole JV-methyl resonance, is replaced in the lH NMR spectrum of 98 by a resonance at 8 11.20 (s, NH1), assigned to the indole NH. HMBC correlations between the resonance at 169 8 11.20 (NH1) and the carbon resonances assigned to C3 (8 117.2), C4 (8 124.8), and C9 (8 137.7) of the indole ring, confirmed its assignment. Other HMBC correlations consistent with the indole ring were observed between the proton resonance at 8 7.15 (H2) and the carbons resonances assigned to C3 (8 117.2), C4 (8 124.8), and C9 (8 137.7), between the proton resonance at 8 8.08 (H5) and the carbon resonance assigned to C9 (8 137.7), between the proton resonance at 8 7.02 (H6) and the carbon resonances assigned to C4 (8 124.8) and C8 (8 112.1), between the proton resonance at 8 7.11 (H7) and the carbon resonance assigned to C9 (8 137.7), and finally, between the proton resonance at 8 7.40 (H8) and the carbon resonances assigned to C4 (8 124.8) and C7 (8 121.1) (Figure 73). An isolated spin system in the COSY spectrum, containing correlations between a methyl resonance at 8 2.23 (t, J = 5 Hz: Mel7) and two NH resonances 8 7.38 (br s: NH16) and 8.80 (br s: NH16'), of which NH16 (8 7.38) is further correlated to a methine resonance at 8 4.48 (d, J = 10 Hz: HI 1), established the C17/N16/C11 connectivity. The methine resonance at 8 4.48 (Hll) is correlated to a deshielded carbon resonance at 8 67.6 (CI 1) in the HMQC spectrum. The chemical shift of CI 1 (8 67.7) is consistent with the presence of a positive charge on the adjacent amino group. HMBC correlations between a pair of methyl resonances at 8 1.42 (s: Mel4) and 1.40 (s: Mel5) and the carbons assigned to C3 (8 117.2) and C10 (8 37.7), between the Mel4 (8 1.42) resonance and the a-methine carbon resonance at 8 67.7 (CI 1), and between the HI 1 (8 4.48) resonance and the carbon resonances assigned to C10 (8 37.7) and C12 (8 166.0), established the C3/C10/C11 connectivity and completed the structure of the unusual amino acid TV' /?,/^trimethyltryptophan (A) (Figure 73). Two nitrogens atoms, one oxygen atom, and the last two sites of unsaturation are now accounted for; therefore an acyclic structure is required for the remainder of the molecule. 170 A Figure 73. Selected HMBC correlations for the Af'/?,/^trimethyltryptophan fragment (A). A COSY correlation from an amide resonance at 5 8.90 (br d J = 9 Hz: NH18) to a deshielded methine resonance at 4.85 (d J = 9 Hz: HI9), attached to a carbon at 8 55.7 (C19), identified an amino acid a-methine. HMBC correlations between the large singlet resonance at 8 1.00 (Me22, Me23, Me24), which integrates for 9 protons, and the carbon resonances assigned to C21 (8 34.8) and C19 (8 55.7), combined with an HMBC correlation between the a-methine resonance HI9 (8 4.85) and the carbon resonance assigned to C20 (8 170.3), identified the amino acid as terMeucine (B) (Figure 74). Figure 74. Selected HMBC correlations for the terMeucine residue (B). The remaining C9H16NO2 portion of the molecule contains four methyl groups, a carbon-carbon double bond, and a carboxylic acid. The methyl resonance at 8 3.03 (s: Me30), correlated 1.00 B 171 to a carbon resonance at 8 31.3 (C30) in the HMQC experiment, has a chemical shift appropriate for an /V-mefhyl group. The methyl resonance at 8 1.80 (s: Me34), correlated to a carbon resonance at 8 13.6 (C34) in the HMQC experiment, was characteristic of an olefinic methyl group. An isopropyl group was identified by two methyl doublets at 8 0.81 (d, J = 6 Hz: H32) and 0.79 (d, J = 6 Hz: H33), which are both correlated to a methine resonance at 8 2.01 (m: H31) in the COSY spectrum. The resonance at 8 2.01 (H31) is also coupled to a methine resonance at 8 4.94 (t, J = 10 Hz: H26), which is further coupled to an olefinic methine resonance at 6.67 (d, J = 10 Hz: H27). HMBC correlations between the resonance at 8 4.94 (H26) and the carbon resonances assigned to C30 (8 31.3), C27 (8 138.4), and C28 (8 131.9), between the olefinic methine resonance H26 (8 6.67) and the carbonyl carbon resonance at 8 168.7 (C29), and between Me34 (8 1.80) and carbon resonances assigned to C27 (8 138.4), C28 (8 131.9), and C29 (8 168.7), completed the assignment of the TV-methyl homo-vinylogous valine (C) (Figure 75). C Figure 75. Selected HMBC correlations for ./V-mefhyl homo vinylogous valine residue (C). The connectivity of fragments A, B, and C was determined by several HMBC correlations (Figure 76). Correlations between the resonances at 8 8.90 (NH18) and 4.85 (H19) and the carbonyl carbon resonance at 8 166.0 (C12) established the amide linkage between the 172 Af',/i,/J-trimethyltryptophan (A) and tert-leucine (B) moieties. Correlations between the resonances at 5 4.94 (H26) and 3.03 (Me30) and the carbonyl resonance at 5 170.3 (C20) established the amide linkage between the tert-leucine (B) and MHVV (C) moieties. 98 Figure 76. HMBC correlations establishing the connectivity of hemiasterlin A (98). Determination of the absolute stereochemistry of hemiasterlin A (98) was accomplished through a combination of chemical degradation and CD analysis. As with hemiasterlin (86), chemical degradation and subsequent Marfey's amino acid analysis45 of hemiasterlin A (98) determined that the tert-leucine and MHVV moieties both have the L configuration. A comparison between the CD spectra of hemiasterlin (86) and hemiasterlin A (98) (Figure 77) revealed that both curves contain a negative Cotton effect at approximately 245 nm and a positive Cotton effect at approximately 220 nm. Since terMeucine and MHVV have the L configuration in both molecules, and the N,N\/?,/i-tetramethyltryptophan in hemiasterlin (86) was determined to have the L configuration by X-ray diffraction analysis, the N',B,B-trimethyltryptophan moiety of hemiasterlin A (98) must also have the L configuration in order for the two CD spectra to be so similar. 173 a . o o o E + o i CD Cmdeg] - a . o o o E + o i 1 . 0 0 0 E + 0 3 HT tv: O.OOOE+OO 1 I 1 • ' 1 I • 1 1 ' I 1 1 1 1 I 1 1 1 1 I 1 11 I I I 1 1 1 1 1 • I • 11 1 1 • I I • I I • • 11111 I 11 1 I ' 1 1 ' I • ' 1 ' I 1 ' ' • I • • • • I 1 • ' • I ' I 11 • 1 1 1 1 1 1 1 1 1 1 1 1 1 I • 1 1 1 1 1 1 • 11 I I 11 I • • • 11 B l . O O O E + O l CD [ffldeg] - 1 . B O O E + O l 1.OOOE+03 1 I 1 ' ' ' 1 1 ' ' ' I 1 1 ' ' I ' ' ' ' I ' H T M O.OOOE+OO I i • t i I • i i i I i I i i i i I i i i i I i WL [nm] Figure 77. CD spectra of hemiasterlin (86, spectrum A) and hemiasterlin A (98, spectrum B) 174 4.2.1.3 Hemiasterlin B (99) 99 Hemiasterlin B (99) was isolated as an amorphous white solid which exhibits an (M+H)+ ion in the HRFABMS at m/z 499.3319, appropriate for a molecular formula of C28H42N4O4, requiring 10 sites of unsaturation. Examination of the ! H NMR, 1 3 C NMR, COSY, HMQC, and HMBC spectral data for hemiasterlin B (13) (Table 17) revealed the molecule contains the same Af',/?,/frrimethyltryptophan and the same iV-methyl homo vinylogous valine (MHVV) 3 6 residues present in hemiasterlin A (98). The main differences between the *H NMR spectrum of hemiasterlin A (98) (Figure 72) and ! H NMR spectrum of hemiasterlin B (99) (Figure 78) is the absence of the large tert-b\x\y\ singlet resonance at 8 1.00 and the appearance of two methyl doublet resonances at 5 0.88 (Me22; d, J = 7 Hz) and 0.92 (Me23; d,J = 7 Hz) in the lH NMR spectrum of 99. This change is consistent with the ter/-leucine residue in hemiasterlin A (98) being replaced with a valine residue in 99. Evidence in support of this was found in the COSY spectrum of 99. An isolated spin system in the COSY spectrum starts from the methyl resonances at 5 0.88 (Me22) and 0.92 (Me23), which both correlate to an aliphatic methine resonance at 5 2.08 (m: H21), and continues through an amino acid a-methine proton resonance at 8 4.62 (H19; t, J = 9 Hz), before terminating at an amide NH resonance 5 9.25 (d, J = 8 Hz: NH18) (Figure 79). The HMBC spectrum contains correlations between both of the methyl resonances Me22 (8 0.88) and Me23 (8 0.92) and the carbon resonances assigned to C19 (8 54.8) and C21 (830.0), and also between 175 176 Table 17. NMR data for hemiasterlin B (99). (Recorded in DMSO-J 6 at 500MHz.) Carbon no. 6 1 3 C 6lH C O S Y H M B C 3 1-N 11.14s 2 124.1 7.14dJ= 1 Hz 3 117.0 HI, 2, 14, 15 4 124.5 HI, 2, 6,8 5 120.0 8.08dJ=8Hz H6 6 118.3 7.03tJ=8Hz H5,7 7 121.0 7.12 t J= 8 Hz H6, 8 H8 8 112.0 7.41 d J = 8 H z H7 H6 9 137.5 HI, 2, 5, 7 10 37.4 H l l , 14, 15 11 67.5 4.31 d J = 10Hz H16 H14, 15 12 165.4 H l l , 18, 19 14 27.0 1.42 s H15 15 22.0 1.38 s H l l , 14 16-N 7.35 brs H16', 11, 17 16-N' 8.81 brs H16, 17 17 33.4 2.21 XJ= 5 Hz H16, 16' H l l 18-N 9.25dJ=8Hz H19 19 54.8 4.62 t J = 9 H z H18.21 H22, 23 20 170.8 H19, 26,30 21 30.0 2.08 m H19, 22, 23 H19, 22, 23 22 18.4 0.88dJ=7Hz H21 H23 23 18.5 0.92dJ=7Hz H21 H22 26 56.5 4.91 tJ= 10 Hz H32, 33 27 137.9 6.66dJ=9Hz H27, 31 H26, 34 28 131.7 H26 H34 29 168.5 H27, 34 30 30.5 3.04 s H26 31 28.8 2.01 m H32, 33 32 19.2 0.81 d J = 6 H z H26, 32, 33 H33 33 18.8 0.80d7=6Hz H31 H32 34 13.3 1.79 s H31 H27 Proton resonances that are correlated to the carbon resonance in the 8 1 3 C column. HI 9 (8 4.62) and the carbon assigned to C20 (5 170.8), thus confirming the presence of a valine residue (Figure 79). HMBC correlations established the connectivity of the three amino acid residues in 99 (Figure 80). Correlations between the proton resonances HI 1 (5 4.31), NH18 (5 9.25), and HI9 177 Figure 79. COSY and HMBC correlations for the valine residue in hemiasterlin B (99). (8 4.62) and the carbon resonance assigned to C12 (8 165.4) confirmed the amide linkage between the TV'/?/^trimethy tryptophan and valine residues. Correlations between the proton resonances H19 (8 4.62), H26 (8 4.91), and H30 (8 3.04) and the carbon resonance assigned to C20 (8 170.8) confirmed the amide linkage between the valine and MHVV residues, thus establishing the structure of hemiasterlin B (99). Figure 80. HMBC correlations establishing the connectivity of hemiasterlin B (99). As in hemiasterlin A (98), determination of the absolute stereochemistry of hemiasterlin B (99) was accomplished through a combination of chemical degradation and CD analysis. Chemical degradation and subsequent Marfey's amino acid analysis45 of hemiasterlin B (99) determined that the valine and MHVV residues both have the L configuration. A comparison between the CD spectra of hemiasterlin (86) and hemiasterlin B (99) revealed that the two curves are very similar, each having a negative Cotton effect at approximately 245 nm and a positive 178 Cotton effect at approximately 220 nm (Figure 81). Since both the valine/MHVV dipeptide portion of 99 and the terMeucine/MHVV dipeptide portion of 86 were determined to have the LL configurations, the two molecules must have the same configuration in their trypotophan moieties in order to account for the similarity of their CD spectra. Since the N,N',B,B-tetramethyltryptophan in hemiasterlin (86) was determined to have the L configuration by X-ray diffraction analysis, the trimethyltryptophan moiety of hemiasterlin A (99) must also have the L configuration. - 2 . oooe+oI I . i i i i i i . i i i i i i i i . i . i . . . . i . . . . I . . . . i . . . . i . . i i i . . . . i . i , i i 1 . O O O E + 0 3 , . . . M . . . . i . i . M . . , , i . . . . . . . . . . . . . . i M . . . . i . . . . i O . O O O E + O O • • • • • I • • • • I • • • • I i • • • I • • • • I • • • i I i • • • I i • • • I • i • • I • • • • I i • • i I • • • • I S O O . O Wl_ [ n m ] 3 2 0 . 0 Figure 81. CD spectra of hemiasterlin (86, spectrum A) and hemiasterlin B (99, spectrum B) 4.2.2 The criamides 179 4.2.2.1 Criamide A (100) 100 Criamide A (100) was isolated as an amorphous white solid which exhibits an (M+H)+ ion in the HRFABMS at m/z 669.4454, appropriate for a molecular formula of C35H56N8O5, requiring 12 sites of unsaturation. A comparison between the NMR data obtained for criamide A (100) (Table 18) with the NMR data obtained for hemiasterlin A (98) (Table 16) revealed that 100 contained a tripeptide substructure (Nl to C34) that is identical to the tripeptide structure of hemiasterlin A (98). Analysis of the 'H NMR spectrum of criamide A (100) (Figure 82) revealed the appearance of a new a-methine resonance at 5 4.27 (m: H36), indicating the presence of an additional amino acid. The remaining unidentified atoms in criamide A (C6H13N4O2) could can be attributed to an arginine residue. Consistent with this assignment was the observation of a resonance at 5 156.7 in the C NMR spectrum of criamide A (100) which can be assigned to C42 of the guanidine functionality. COSY correlations define a spin system extending from the arginine a-amino proton (5 8.02, d, J = 8 Hz: NH35) through the a-methine H36 (5 4.27) and three contiguous methylenes (8 1.81, m: H38; 1.67, m: H38'; 1.51, m: H39; 3.10, dd, J = 7, Table 18. NMR data for criamide A (100). (Recorded in DMSO-ck at 500MHz.) Carbon no. 5 U C 5*H COSY HMBC" 1-N 11.14, s 2 124.0 7.16dJ = 2Hz 3 117.2 HI, 2, 14, 15 4 124.5 HI, 2, 6,8 5 120.0 8.08dJ=8Hz H6 H7 6 118.1 7.03tJ=8Hz H5,7 H8 7 120.9 7.12tJ=8Hz H6, 8 H5 8 111.7 7.41 d J = 8Hz H7 H6 9 137.3 H2, 5, 7 10 37.4 H14, 15 11 67.5 4.45 d J=9Hz H16 H14, 15 12 165.8 H19 13 14 26.9 1.42 s H15 15 22.4 1.38 s H14 16-N 7.30 brs H16', 11, 17 16-N' 8.77 br s H16, 17 17 33.4 2.23tJ=5Hz H16, 16' 18-N 8.86dJ = 9Hz H19 19 55.3 4.88dJ=8Hz H18 H22, 23, 24 20 169.8 H19, 26,30 21 34.6 H22, 23, 24 22 26.5 1.00 s H19, 23, 24 23 26.5 1.00 s HI9, 22, 24 24 26.5 1.00 s H19, 22, 23 26 55.6 4.98 t J = 10 Hz H27, 31 H30, 32, 33 27 132.4 6.30dJ = 9Hz H26 H26, 34 28 134.8 H26, 34 29 168.5 H27, 34, NH35 30 30.6 3.04 s H26 31 28.9 2.01 m H26, 32, 33 H32, 33 32 19.2 0.83 d J=7Hz H31 H33 33 18.7 0.78dJ = 7Hz H31 H32 34 13.8 1.82 s H27 35-N 8.02dJ=8Hz H36 36 51.7 4.27 m H35,38, 38* 37 b 38 27.7 1.81 m H36, 38' 38' 1.67 m H38, 39 39 25.2 1.51 m H38', 40 40 40.2 3.10 dd .7=7, 6 Hz H39,41 41-N 7.53tJ=6Hz H40 42 156.7 a Proton resonances that are correlated to the carbon resonance in the 8 1 3 C column. b Not observed due to small sample size. 182 6 Hz: H40) before terminating at the NH resonance of the guanidine functionality (87.53, t, J = 6Hz:NH41)(Figure83). Figure 83. COSY correlations for the arginine residue in criamide A (100). Confirmation of the presence of the arginine residue was obtained by acid hydrolysis of criamide A (100), followed by reaction of the hydrolyzate with 2,4-pentanedione to convert the arginine into its dimethylpyrimidine derivative. The arginine dimethylpyrimidine derivative was further treated with pentafluoropropionyl anhydride and isopropyl alcohol to give a volatile derivative for GC analysis. Comparison of the profile of the derivatized criamide A hydrolyzate on a chiral GC column with the GC retention times for the corresponding derivatives of authentic L and D arginine confirmed the presence of arginine in criamide A (100), and showed that it has the L configuration. HMBC correlations observed between the NH35 (8 8.02), H27 (8 6.30) and H34 (8 1.82) resonances and the carbon resonance assigned to C29 (8 168.5) established the peptide linkage between the arginine a-amino group and the carboxyl group of the MHVV residue. The confirmation of the arginine-MHVV linkage, combined with HMBC correlations consistent with criamide A (100) containing a tripeptide substructure (NI to C34) that is identical to the tripeptide structure of hemiasterlin A (98) (Figure 84), established the structure of criamide A as 183 that shown in 100. We have assumed that the N'/j/J-trimethylated tryptophan, terr-leucine, and MHVV residues in 100 all have the L configuration consistent with hemiasterlin A (98). 100 Figure 84. HMBC correlations establishing the connectivity of criamide A (100). 101 Criamide B (101) was isolated as an amorphous white solid which shows an (M+H)+ ion in the HRFABMS at m/z 683.4597, appropriate for a molecular formula of C36H58N8O5. A comparison of the 'H NMR and COSY data obtained for criamide B (101) with that recorded for criamide A (100) (Table 19) showed that 101 differs from criamide A (100) only by the presence of an indole TV-methyl substituent (8 3.75, s: Mel 3). Consistent with this assignment is the gain of 14 mass units in the mass spectrum and the loss of the indole N-H resonance in the ! H NMR spectrum of 101 (Figure 85). A comparison of the 'H NMR data of 101 (Table 19) with that of hemiasterlin (86) (Table 14) indicated that the tripeptide substructure (Nl to C34) of 101 is identical to the tripeptide structure of 86. The presence of the arginine residue was confirmed by COSY correlations that defined the nine-proton spin system extending from the arginine a-amino proton (8 8.03, d J = 8 Hz: NH35) to the NH41 (8 7.59, d J = 5 Hz) proton. No stereochemical information was obtained for criamide B (101) but we have assumed that the N,N',/?,/^tetramemylatedtryptophan, terr-leucine and MHVV residues in 101 all have the L configurations, as found in hemiasterlin (86), and that the arginine residue has the L configuration, as found criamide A (100). Table 19. NMR data for the criamides. (Recorded in ] DMSO-</6 at 500MHz.) Criamide A (100) Criamide B (101) Carbon no. 6 ' H 6 ' H C O S Y 1-N 11.14s 2 3 7.16dJ=2Hz 7.17 s 4 5 8.08d7=8Hz 8.10d7= 8Hz H6 6 7.03tJ=8Hz 7.08tJ=8Hz H5, 7 7 7.12tJ=8Hz 7.21 t 7=8Hz H6, 8 8 Q 7.41 d J = 8 H z 7.46d7=8Hz H7 y 10 11 4.45d7=9Hz 4.44dJ=9Hz H16 12 13 3.75 s 14 1.42 s 1.42 s 15 1.38 s 1.38 s 16-N 7.30 br s 7.35 br s HI 6', 11, 17 16-N' 8.77 br s 8.84 br s H16, 17 17 2.23t7=5Hz 2.24 br s H16, 16' 18-N 8.86d7=9Hz 8.87d7=6Hz H19 19 4.88dJ=8Hz 4.88d7=9Hz H18.21 20 21 22 1.00 s 1.00 s 23 1.00 s 1.00 s 24 1.00 s 1.00 s 26 4.98 17 = 10 Hz 4.98 \J = 10 Hz H27, 31 27 6.30dJ=9Hz 6.30d7=9Hz H26 28 29 30 3.04 s 3.01 s 31 2.01 m 1.94 m H26 32 33 32 0.83d7=7Hz 0.83dJ=8Hz H31 33 0.78d7=7Hz 0.77dJ=8Hz H31 34 1.82 s 1.82 s 35-N 8.02 d7= 8 Hz 8.03d7=8Hz H36 36 4.27 m 4.26 dd7=9,4Hz H35,38,38' 37 38 1.81 m 1.79 m H36, 38' 38' 1.67 m 1.69 m H38,39 39 1.51 m 1.49 m H38', 40 40 3.10 dd 7=7, 6 Hz 3.09 dd7=7, 6 Hz H39, 41 41-N 7.53tJ=6Hz 7.59 XJ= 5 Hz H40 4.2.3 Biological activity of the hemiasterlins and criamides 187 Results of the biological testing of the hemiasterlins and criamides were obtained through collaborative efforts. The in vitro cytotoxicity assays against murine leukemia P388, human mammary carcinoma MCF-7, human glioblastoma/astrocytoma U373, human ovarian carcinoma HEY, and the in vivo cytotoxicity assay against P388 cells in mice were performed in Dr. Theresa Allen's laboratory, in Department of Pharmacology at the University of Alberta.37 The study of the affect of the hemiasterlins on cell survival and cell cycle progression in MCF-7 cells was performed by Dr. Michel Roberge and Dr. Hilary Anderson, in Department of Biochemistry at the University of British Columbia.46 The hemiasterlins and criamides were found to be potent in vitro cytotoxins, showing activity against murine leukemia P388, human mammary carcinoma MCF-7, human glioblastoma/astrocytoma U373, and human ovarian carcinoma HEY (results summarized in Table 20). Hemiasterlin (86) also showed promising in vivo cytotoxicity against P388 cells in mice. Mice injected with 1 x 106 P388 leukemia cells and treated with hemiasterlin (86), 5 daily doses of 0.45 pg 86 begun 24 hours after cancer implantation, were observed to have a 123.4% increase in their life span. Preliminary data indicated that the effective in vivo dose against P388 in mice is at least two orders of magnitude lower than the toxic dose.47 The promising in vitro and in vivo cytotoxicity results prompted an investigation into the mechanism of action of the hemiasterlins. In collaboration with Dr. Michel Roberge, a study of the effect of the hemiasterlins on cell survival and cell cycle progression in MCF-7 cells was initiated. Results for hemiasterlin (86), hemiasterlin A (98), and hemiasterlin B (99) suggest that they exert their cytotoxic effect by binding to tubulin and inhibiting spindle microtubule dynamics. 188 Table 20. in vitro Cytotoxicities for the hemiasterlins and criamides.37 I C 5 0 Values (nM) compound P388 U373 HEY MCF-7 mitotic arrest hemiasterlin (86) 0.09 (1.9 x 104)a 23 3 0.3 0.3 milnamide A (84) N H ^ (1.4 x 106)b hemiasterlin A (98) 0.003 3 15 2 2 hemiasterlin B (99) yJ ^N.h H o 1 H 14 32 7 20 criamide B (101) H ft NH, f Y O "^i^ 0 ^ N H ^ ^ ^ N i ^ J I ^ ^ N ^ ^ ^ J I ^ ^ O H 11 390 280 10 x 103 a hemiasterlin (86) sponge source Hemiasterella minor b milnamide A (84) sponge source Auletta cf. Constricta To ensure that all of the cell cycle progression experiments were performed using the correct concentration of drugs, the cytotoxicity of the hemiasterlins towards MCF-7 human mammary carcinoma cells was determined. Hemiasterlin (86) killed MCF-7 cells half-maximally at 0.3 nM, while hemiasterlin A (98) and hemiasterlin B (99) showed similar cytotoxicity at concentrations of 2 nM and 7 nM, respectively (Figure 86). To ensure that the 189 cytotoxic effects observed for the hemiasterlins were not due to contamination by the co-occurring cytotoxins geodiamolide A to G (77-83), a representative of this family, namely geodiamolide B (78), was also tested and was found to be considerably less cytotoxic, killing cells half maximally at 80 nM. 4 6 hemiasterlin (86) hemiasterlin A (98) hemiasterlin B (99) paclitaxel (114) nocodazole (145) vinblastine (116) geodiamolide B (78) 0.01 0.1 1 10 100 1000 [drug] nM Figure 86. Cytotoxicity of the hemiasterlins, microtubule inhibitors, and geodiamolide B (78). Mean and standard deviation of quadruplicate measurements are shown for each point.46 Many antitumor drugs affect the transition between particular phases of the cell cycle. As a starting point in determing the mechanism of action of the hemiasterlins, their effect on mitosis was examined. Cycling MCF-7 cells treated with the hemiasterlins induced a concentration-dependent accumulation of cells in mitotic metaphase, with hemiasterlin (86) having a half-maximal effect at 0.5 nM, while hemiasterlin A (98) and hemiasterlin B (99) showed the same effect at 2.5 nM and 28 nM concentrations, respectively (Figure 87). Al l three peptides caused cytotoxicity and mitotic arrest in the same concentration range, indicating that mitotic arrest is an important determinant of their cytotoxic activity.46 190 O N N N O C H 3 H O 145 Hamel has made the general observation that almost all antimitotic agents found to date alter microtubule assembly reactions by interacting with the protein tubulin during mitosis.4 This, combined with our observation that the hemiasterlins cause mitotic arrest in cells, prompted us to compare the cytotoxic and antimitotic effects of the hemiasterlins to the known microtubule inhibitors paclitaxel (114), vinblastine (116), and nocodazole (145) (Figure 86 and 87). Against human breast cancer MCF-7 cells, the hemiasterlins proved to be potent cytotoxins, having half-maximal effects of 0.3 nM, 2 nM, and 7 nM for hemiasterlin (86), hemiasterlin A (98), and hemiasterlin B (99), respectively. Both hemiasterlin (86) and hemiasterlin A (98) proved to be more cytotoxic than the known microtubule inhibitors which have half-maximal effects of 2 nM for paclitaxel (114), 6 nM for vinblastine (116), and 15 nM for nocodazole (145)46 This trend continued when the mitotic blocking ability of the hemiasterlins was compared to the known microtubule inhibitors (Figure 87). Hemiasterlin (86) and hemiasterlin A (98) have half-maximal effects at 0.3 and 2 nM, respectively, compared with half-maximal effects of 7 nM for paclitaxel (114), 15 nM for vinblastine (116), and 40 nM for nocodazole (145)46 Geodiamolide B (78) has no effect on cells in mitosis, even at 300 nM 4 6 191 80 i hemiasterlin (86) hemiasterlin A (98) hemiasterlin B (99) paclitaxel (114) nocodazole (145) vinblastine (116) geodiamolide B (78) 0.01 0.1 1 10 100 1000 [drug] nM Figure 87. Antimitotic activity of the hemiasterlins, microtubule inhibitors and geodiamolide B Next, MCF-7 cells were treated with concentrations of hemiasterlin A (98), paclitaxel (114), vinblastine (116), and nocodazole (145) that produced half-maximal or maximal mitotic arrest. The cells were examined on the basis of the morphology of their mitotic spindles by indirect immunofluorescence using a monoclonal antibody to P-tubulin, and the distribution of their chromosomes using the fluorescent DNA dye bisbenzimide. The results are shown in Figure 88. In untreated cells, the mitotic spindle can be seen radiating from two asters (Figure 88 A) and the condensed chromosomes are aligning between the two asters (Figure 88 B). Cells treated with hemiasterlin A (98) at 2 nM have no normal spindles. Some cells show relatively minor abnormalities in which a bipolar spindle is present, but the astral microtubules are considerably longer than normal and the chromosomes are not completely confined to the metaphase plate. Most commonly cells have multiple asters, and the chromosomes are distributed in a spherical mass (Figure 88 C, D). Half-maximal concentrations of paclitaxel (114), vinblastine (116), and nocodazole (145) produce the same types of abnormal spindle (not shown).46 Hemiasterlin A (98) at 10 nM, the lowest concentration causing maximal mitotic (78). 46 192 chromosomes are again distributed in a spherical mass (Figure 88 E, F ) 4 6 The same effect is observed for high concentrations of vinblastine (116) and nocodazole (145) (not shown). At even higher concentration, vinblastine (116) assembles tubulin dimers into nearly crystalline arrays called vinblastine paracrystals. Paclitaxel (114), which stabilized microtubules at high concentration has quite a different effect, causing bundling of cytoplasmic microtubules in interphase cells and very dense multiple asters in mitotic cells (not shown).46 A B , m * • D E F Figure 88. Microtubule and chromosome distribution after 20 h incubation with no drug (A, B ) . 2nM hemiasterlin A (C, D ) or 10 nm hemiasterlin A (E, F). A, C, E antitubulin immunofluorescence; B , D , F DNA staining with the dye bisbenzimide. Bar 26 4.3 CONCLUSIONS 193 Hemiasterlin A (98), hemiasterlin B (99), criamide A (100), and criamide B (101) are four new members of the hemiasterlin family of peptides. The structures of 98-101 were determined using spectroscopic techniques. Degradative studies and X-ray diffraction analysis were used to determine the absolute configuration of hemiasterlin (86). terr-Leucine, MHVV, and TV, TV'/?,/?-tetramethytryptophan were all shown to have the L configuration. By a combination of degradative studies and comparative CD analysis, the absolute configurations of hemiasterlin A (98) and hemiasterlin B (99) were determined, and all the amino acids were shown to have the L configuration. Cymbastela sp. is the third sponge, along with Auletta c f 36 35 constricta and Hemiasterella minor, that has been found to contain members of both the jaspamide/geodiamolide and hemiasterlin/criamide families of cytotoxic peptides. As was the case with the jaspamide/geodiamolide family of metabolites, the hemiasterlin family of metabolites occurs across a taxonomically distant group of sponge species. It is likely that one or two species of microorganisms associated with the sponges produce these metabolites. Attempts to isolate the marine microorganisms that are thought to produce the metabolites in the hemiasterlin family are currently underway. The hemiasterlins are extremely cytotoxic peptides that show promise as antitumor drugs. They exhibit potent in vitro cytotoxicity against a variety of human solid tumor cell lines, and hemiasterlin (86) has shown equally potent cytotoxicity against murine leukemia P388 both in vitro and in vivo. Preliminary data indicate that the effective in vivo dose against P388 is at least two orders of magnitude lower than the toxic dose. We have shown that hemiasterlin A (98) is cytotoxic to MCF-7 cells at subnanomolar concentrations and causes mitotic arrest in the same low nanomolar range. 194 A broad array of drugs can arrest cells in mitosis, producing similar mitotic spindle abnormalities. Some of these, such as the microtubule inhibitors paclitaxel (114), vinblastine (116) and nocodazole (145), directly bind tubulin. When MCF-7 breast cancer cells were treated with these drugs at their mitotic arrest concentrations, the cells' microtubule dynamics were affected, resulting in abnormal spindle structure and function. The same effects were observed in cells that were treated with hemiasterlin A (98). At higher concentrations, vinblastine (116) and nocodazole (145) caused the depolymerization of microtubules, as did hemiasterlin A (98). At very high concentrations of vinblastine (116), tubulin dimers were assembled into nearly crystalline arrays called vinblastine paracrystals. In contrast, at high concentration paclitaxel (114) enhanced microtubule assembly and produced thick bundles of microtubules. At high concentration, hemiasterlin A (98) did not produce paracrystals or microtubule bundles. 195 4.4 EXPERIMENTAL 4.4.1 Materials. Specimens of Cymbastela sp. were collected by hand using SCUBA on outer reefs at -10 to -25 meters near Motupore and Madang in Papua New Guinea. Freshly collected sponge was frozen on site and transported to Vancouver packed in dry ice. Bio-assay guided fractionation of the Cymbastela sp. sponge was performed as outlined below. Freeze-dried sponge (160 g dry weight) was extracted exhaustively with a 1:1 solution of CH 2Cl 2/MeOH (4 x 4L). The combined organic extracts were filtered and concentrated in vacuo to yield 14 g of a dark yellow/orange oil which was suspended in 250 mL of a 9:1 solution of MeOH/H 20 and extracted with hexanes (3 x 150 mL). The MeOH/H 20 layer was diluted with H 2 0 to a 4:1 MeOH/H 20 solution and extracted with CCL (3 x 150 mL). Further dilution of the MeOH/H 20 layer with H 2 0 resulted in a 2:3 MeOH/H 20 solution which was extracted with CHC13 (3 x 150 mL). Al l traces of MeOH were removed in vacuo and the resulting aqueous solution was extracted with EtOAc (3 x 150 mL). Biological testing of the five fractions revealed that the majority of the biological activity was concentrated in the CCI4 and CHCI3 soluble portions which were combined and fractionated further. Separation of the combined CCI4/CHCI3 fraction by Sephadex LH-20 chromatography, eluting with MeOH, yielded two active fractions, one containing a mixture of geodiamolides and the other containing mixture of peptides as identified by TLC and 'H NMR. Further purification of the peptide fraction by gradient reversed phase flash chromatography (H 20 to H 20/MeOH 30:70) followed by reversed-phase isocratic HPLC (0.05% TFA/H 20:MeOH 1:1) afforded pure hemiasterlin (86), hemiasterlin A (98), hemiasterlin B (99), criamide A (100) and criamide B (101). 196 4.4.1.1 Hemiasterlin (86). 86 Compound 86 was isolated as an amorphous white solid (40 mg, 0.015 % dry wt); [CC]D = -76° (c 0.7, MeOH); UV (MeOH) X m a x (s) 216 (15,400), 273 nm (1,600); IR (neat): 3412, 2962, 1650, 1635 cm"1; for NMR data, see Table 14; exact mass calculated for C30H47N4O4 (M+H)+: 527.35973; found: 527.35937. 4.4.1.2 Hemiasterlin A (98). H 98 Compound 98 was isolated as an amorphous white solid (32 mg, 0.012 % dry wt); [CC]D = -45° (c 0.25, MeOH); UV (MeOH) X m a x (e) 218 (23,400), 280 nm (2,800); IR (neat): 3418, 2966, 1689, 1680, 1643 cm"1; for NMR data, see Table 16; exact mass calculated for C29H45N4O4 (M+H)+: 513.34408; found: 513.34711. 4.4.1.3 Hemiasterlin B (99). H 99 197 Compound 99 was isolated as an amorphous white solid (1 mg, 0.0004 % dry wt); for NMR data, see Table 17; exact mass calculated for C 2 8 H 4 3 N 4 O 4 (M+H)+: 499.32843; found: 499.33187. 4.4.1.4 Criamide A (100). 100 Compound 100 was isolated as an amorphous white solid (3.3 mg, 0.0013 % dry weight); [a ] D 2 5 = +97° (c 0.2, MeOH); for NMR data, see Table 18; exact mass calculated for C 3 5H 5 7 N 8 05 (M+H)+: 669.44519; found: 669.44593. 4.4.1.5 Criamide B (101). H 101 Compound 101 was isolated as an amorphous white solid (3.3 mg, 0.0013 % dry weight); for NMR data, see Table 19; exact mass calculated for C 36H59N 80 5 (M+H)+: 683.46084; found: 683.45970. 198 4.4.1.6 Hemiasterlin methyl ester (141). 141 To a stirred suspension of hemiasterlin (86) (5.0 mg, 0.009 mmol) in ether (1 mL) at room temperature was added an ethereal solution of diazomethane dropwise until the yellow color of the diazomethane persisted in the reaction mixture, and TLC analysis showed complete consumption of starting material. Excess diazomethane was removed under a stream of argon and the remaining solvent removed in vacuo to afford ester 141 as a crystalline solid (5 mg, "100%"). Recrystallization of 141 from 1:3 acetone/hexanes solution afforded clear colourless rod-like crystals suitable for X-ray diffraction analysis. *H NMR (CDC13, 500 MHz, 5): 0.78 (d, J=7Hz , 3H), 0.86 (d, J=7Hz , 3H), 0.96 (s, 9H), 1.45 (s, 3H), 1.57 (s, 3H), 1.89 (s, 3H), 1.98-2.05 (m, 1H), 3.03 (s, 3H), 3.59 (s, 1H), 3.73 (s, 3H), 3.74 (s, 3H), 4.84 (m, 1H), 5.08 (t, J= 9.5 Hz, 1H), 6.64 (d, J = 9.5 Hz, 1H), 7.07 (t, J = 8 Hz, 1H), 7.12 (t, J = 8 Hz, 1H), 7.20 (d, J = 8 Hz, 1H), 7.29 (s, 1H), 7.86 (m, 1H); exact mass calculated for C31H49N4O4 (M+H)+: 541.37538; found: 541.37553. 4.4.1.6.1 Single crystal X-ray diffraction data for hemiasterlin methyl ester (141). formula C 3 1 H 4 8 N 4 0 4 radiation C u K a formula weigh 540.74 temperature (K) 294 crystal system Orthorhombic 26 max (deg) 155 space group P2i2i2i No. reflections 2125 [7>3a(7)] a (A) 12.2102 (8) parameters 370 b(A) 31.464(2) R 0.0391 c(A) 8.368 (2) Rw 0.0375 Z 4 goodness to fit 2.207 Dcaic (g cm"3) 1.117 residual peak -0.10 to+0.13 199 density (e A"3) Single crystal X-ray diffraction analysis performed on a Rigaku AFC-6S diffractometer. The structure was solved by direct methods using SIR92 and expanded using Fourier techniques. Al l calculations were performed using TEXSAN and MSC/AFC Diffractometer Control Software was used for data collection and cell refinement. 4.4.1.7 Ozonolysis of Hemiasterlins. Pure individual hemiasterlins (~ 1 mg) were dissolved in 6 mL of MeOH and cooled to -78° C. A stream of ozone was bubbled through the solution until a deep blue color persisted. A stream of nitrogen was then bubbled through the solution to remove excess ozone and the reaction mixture was allowed to warm to room temperature of 30% H2O2 (3 mL) was added and the resulting solution was stirred for 40 min. Excess reagents were removed under high vacuum to yield the crude ozonolysis product. 4.4.1.8 Hydrolysis of the Ozonolysis Product of Hemiasterlins. The crude reaction product from the ozonolysis of hemiasterlins (see above) was dissolved in 1 mL of freshly distilled HC1 and the resulting solution was heated at 108°C with stirring for 16 hours in a threaded Pyrex® tube sealed with a Teflon® screw cap. The cooled reaction mixture was evaporated to dryness and traces of HC1 were removed from the residual hydrolyzate by repeated evaporation from H 2 0 (3x3 mL). 4.4.1.9 Derivatization of amino acids with Marfey's Reagent and HPLC analysis. To a 1 mL vial containing 2 mmol of pure amino acid standard in 80 pL of H2O was added 2.8 mmol of 7V-a-(2,4-dinitro-5-fluorophenyl)-Z-alaninamide (FDAA) in 170 pL of acetone 200 followed by 20 u L of I N NaHCC»3. The mixture was heated for 1 hour at 40°C. After cooling to room temperature, 10 p L o f 2 N HC1 were added and the resulting solution was filtered through a 4.5 pm filter and stored in the dark until H P L C analysis. To prepare F D A A derivatives of the amino acids in the hydrolyzate of the ozonolysis product of the hemiasterlins, a 90 p L aliquot containing 0.9 mg of amino acid mixture was reacted with 2.86 mmol of F D A A in 172 p L of acetone as described above. A 10 uL aliquot of the resulting mixture of F D A A derivatives was analyzed by reversed-phase H P L C . A linear gradient of (A) 9:1 triethylammonium phosphate (50 m M , p H 3.0) /MeCN and (B) M e C N , with 0% B at the start to 40% B over 60 minutes (flow rate 1 mL/min) was used to separate the F D A A derivatives, which were detected by U V at 340 nm. Each peak in the chromatographic trace was identified by comparing its retention time with that of the F D A A derivative of the pure amino acid standard and by co-injection. In all cases a peak was observed at -35.7 min which was attributed to excess F D A A . The H P L C retention times are tabulated below: Table 21. H P L C retention times for Marfey's amino acid derivatives in minutes. D-t-Le L-t-Leu D/L-Val I -Va l D-JV-Me-Val I -N-Me-Val standards 61.30 55.66 149.19 D 55.06 49.40 59.10 55.81 hemiasterlin (86) 55.75 56.01 hemiasterlin A (98) 55.54 56.21 hemiasterlin B (99) 48.57 55.99 4.4.1.10 Gas Chromatography (GC). Capillary G C analyses were carried out on a Hewlett Packard 5880A gas chromatograph using a Chirasil V a l column (0.25 mm x 50 m, film thickness: 0.16 mm, All tech Associates, Deerfield, IL) and helium as a carrier gas (flow rate: 30 ml/min). The program rate for the analysis of amino acid derivatives was 90°C (4 min) to 220°C at 4°C/min, then 220°C for 30 min. The injector temperature was at 250°C and the detection temperature was set at 275°C. L and D arginine standards were converted to their dimethylpyrimidine derivatives prior to G C analysis. 201 The arginines (500 ug) were treated with a mixture of H2O (25 pL), EtOH (50 pL), triethylamine (2-5 pL), and acetylacetone (50 uL) at 110°C for 4 hours in a threaded Pyrex® tube sealed with a Teflon® screw cap. The reaction mixture was evaporated to dryness under a stream of N 2 . The arginine dimethylpyrimidine derivatives (500 pg) were treated with 250 pL of isopropyl acetate at 110°C for 75 minutes in a screw cap vial. The reaction mixture was evaporated under a stream of N 2 , CH 2C1 2 (250 pL) and pentafluoropropionic anhydride (PFPA) (100 pL) was added, and heated at 110°C for 15 minutes and then evaporated under N 2 . The residual material was dissolved in CH 2C1 2 (200 pL) for GC analysis. Retention times: D 42.71 min, L 43.02 min. 4.4.1.11 Acid Hydrolysis of Criamide A (100) and Derivatization of the Hydrolysate. Criamide A (100) (~ 0.5 mg) was dissolved in 1 mL of freshly distilled HC1 and the resulting solution was heated at 108°C with stirring for 16 hours in a threaded Pyrex® tube sealed with a Teflon® screw cap. The cooled reaction mixture was evaporated to dryness and traces of HC1 were removed from the residual hydrolyzate by repeated evaporation from H 2 0 (3x3 mL). The residue was converted to the dimethylpyrimidine derivative and then treated with isopropyl acetate and PFPA prior to GC analysis as described above. 202 4.5 ENDNOTES: CHAPTER 4: PEPTIDES FROM CYMBASTELA SP. I) Crowley, L. V. Intorduction to Human Disease; 4 ed.; Jones and Bartlett Publishers: Sudbury, Massachuesetts, 1988, pp 225-223. 2a) Pratt, W. B.; Ruddon, R. W.; Ensminger, W. D.; Maybaum, J. The Anticancer Drugs; 2 n d ed.; Oxford University Press: New York, 1994, pp 3-24. 2b) ibid, pp 69-100. 2c) ibid, pp 108-144. 2d) ibid, pp 188-193. 2e) ibid, pp 50-64. 3) Murray, A.; Hunt, T. The Cell Cycle: an introduction; Oxford University Press: New York, New York, 1993, pp 3-15. 4) Hamel, E. Pharmac. Ther. 1 9 9 2 , 55,31. 5) Baker, J. J. W.; Allen, G. E. The study of biology; 3 r d ed.; Addison-Wesly Publishing Company: Reading, Massachusetts, 1978, 145-149. 6) Stryer, L. Biochemistry; 3 r d ed.; W. H. Freeman and Company: New York, New York, 1988, pp 938-945. 7) Nakamura, T. Chem. Pharm. Bull. 1962 ,10 , 299. 8) Uppuluri, S.; Knipling, L.; Sackett, D. L.; Wolfe, J. Proc. Natl. Acad. Sci. U.S.A. 1 9 9 3 , 90, 11598. 9) Hamel, E. Med. Res. Rev. 1996 ,16 , 207. 10) Neuss, N. ; Gorman, M. ; Hargrove, W.; Cone, N . J.; Biemann, K.; Bilchi, G.; Manning, R. E. J. Am. Chem. Soc 1 9 6 4 , 86, 1440. II) Wall, M . E.; Wani, M. C. 153rd National Meeting, American Chemical Society: Paper M -006, 1967. 12) Wani, M . C ; Taylor, H. L.; Wall, W. E.; Coggin, P.; McPhail, A. T. J. Am. Chem. Soc. 1 9 7 1 , 93, 2325. 13) Nicolaou, K. C ; Yang, Z.; Liu, J. J.; Ueno, H.; Niantermet, P. G.; Guy, R. K.; Claiborne, C. F.; Renaud, J.; Couladouros, E. A.; Paulvannan, K.; Sorensen, E. J. Nature 1 9 9 4 , 367, 630-634. 203 14) Holton, R. A. J. Am. Chem. Soc. 1994,116, 1597. 15) Holton, R. A. J. Am. Chem. Soc. 1994,116, 1599. 16) Senilh, V.; Blechert, S.; Colin, M. ; Guenard, D.; Picot, F.; Potier, P.; Varenne, P. J. Nat. Prod. 1984,47, 131. 17) Schiff, P. B.; Fant, J.; Horwitz, S. B. Nature 1979, 277, 665. 18) Nagle, W. H.; Geralds, R. S.; Yoo, H.; Gerwick, W. H.; Kim, T.; Nambu, M. ; White, J. D. Tetrahedron 1995, 36, 1189. 19) Gerwick, W. H.; Proteau, P. J.; Nagel, D. G.; Hamel, E.; Blokhin, A.; Slate, D. L. J. Org. Chem. 1994, 59, 1243. 20) Pettit, G. R.; Kamano, Y.; Herald, C. L.; Tuinman, A. A.; Boettner, F. E.; Kizu, H.; Schmidt, J. M. ; Baczynskyj, L.; Tomer, K. B.; Bontems, R. J. J. Am. Chem. Soc. 1987,109, 6883. 21) Hirata, Y.; Uemura, D. Pure Appl. Chem. 1986, 58, 701. 22) Bai, R.; Paull, K. D.; Herald, C. L.; Malspeis, L.; Pettit, G. R.; Hamel, E. J. Bio. Chem. 1991,266, 15882. 23) Pettit, G R; Cichacz, Z. A.; Gao, F.; Herald, C. L.; Boyd, M . R. J. Chem. Soc, Chem. Commun. 1993, 1166. 24) Pettit, G R; Herald, C. L.; Cichacz, Z. A.; Gao, F.; Boyd, M . R.; Schmidt, J. M. ; Boettner, F. E.J. Chem. Soc, Chem. Commun. 1993, 1805. 25) Pettit, G R; Herald, C. L.; Cichacz, Z. A.; Gao, F.; Boyd, M . R.; Christie, N . D.; Schmidt, J. M. Nat. Prod. Lett. 1993, 3, 239. 26) Pettit, G R; Cichacz, Z. A.; Herald, C. L.; Gao, F.; Boyd, M . R.; Schmidt, J. M. ; Hamel, E.; Bai, R.J. Chem. Soc, Chem. Commun. 1994, 1605. 27) Bai, R.; Taylor, G. F.; Cichacz, Z. A.; Herald, C. L.; Kepler, J. A.; Pettit, G. R.; Hamel, E. Biochemistry 1995, 34, 9714. 28) Trimurtulu, R. E.; Ohtani, I.; Patterson, G. M . L.; Moore, R. E.; Corbett, T. H.; Valeriote, F. A.; Demchik, L. J. Am. Chem. Soc. 1994,116, 4729. 29) Golakoti, T; Ogino, J; Heltzel, C. E.; Husebo, T. L; Jensen, C. M. ; Larsen, L. K.; Patterson, G. M . L.; Moore, R. E.; Mooberry, S, L.; Corbett, T. H.; Valeriote, F. A. J. Am. Chem. Soc. 1995,777,12 030.. 30) Subbaraju, G. V.; Trimurtulu, R. E.; Patterson, G. M . L.; Moore, R. E. J. Nat. Prod. 1997, 60, 302. 204 31) Kerksiek, K.; Mejillano, M . R.; Schwartz, R. E.; Georg, G. I.; Himes, R. H. FEBS Letters 1995, 377, 56. 32) Gunasekera, S. P.; Gunasekera, M. ; Longley, R. E. J. Org. Chem. 1994, 56, 4912. 33) ter Haar, E.; Kowalski, R. J.; Hamel, E.; Lin, C. M. ; Longley, R. E.; Gunasekera, S. P.; Rosenkranz, H. S.; Day, B. W. Biochemistry 1996, 35, 243. 34) Lindel, T.; Jensen, P. R.; Fenical, W.; Long, B. H.; Casazza, A. M. ; Carboni, J.; Fairchild, C. R. J. Am. Chem. Soc. 1997,119, 8744. 35) Talpir, R.; Benayahu, Y.; Kashman, Y.; Pannell, L.; Schleyer, M . Tetrahedron Letters 1994, 35, 4453. 36) Crews, P.; Farias, J., J.; Emrich, R.; Keifer, P., A. J. Org. Chem. 1994, 59, 2932. 37) Coleman, J. E.; de Silva, E. D.; Kong, F.; Andersen, R. J.; Allen, T. M . Tetrahedron 1995, 51, 10653. 38) Friebolin, H. Basic One- and Two-Dimensional NMR Spectroscopy; second ed.; VCH Publishers: Weinheim (Feberal Repuplic of Germany), 1993. 39) Schipper, D. J. Antibiot. 1983, 36, 1076. 40) Matsunaga, S.; Fusetani, N. ; Konosu, S. Tetrahedron Lett. 1984, 25, 5165. 41) Matsunaga, S.; Fusetani, N. ; Konosu, S. J. Nat. Prod. 1985, 48, 236. 42) Matsunaga, S.; Fusetani, N. ; Konosu, S. Tetrahedron Lett. 1985, 26, 855. 43) Gulavita, N . K.; Gunasekers, S. P.; Pomponi, S. A.; Robinson, E. V. J. Org. Chem. 1992, 57, 1767. 44) Cardelina, J. personal communication, 1997. 45) Marfey, P. Carlsberg. Res. Commun 1984, 49, 591. 46) Anderson, H., J.; Coleman, J. E.; Andersen, R., J.; Roberge, M . Cancer Chem. Pharm. 1996, 39, 223. 47) Allen, T. personal communication, 1996. 205 5 SYNTHESIS O F H E M I A S T E R L I N (86) A N D S T R U C T U R E A C T I V I T Y STUDY. 5.1 INTRODUCTION The naturally occurring hemiasterlins and criamides (86, 98-101) show potent in vitro cytotoxicity against a variety of cell lines1 and have been shown to be antimitotic agents2 (see Chapter 4, Section 4.2.3). Against human breast cancer MCF7 cells, hemiasterlin (86) and hemiasterlin A (98) were found to be more potent cytotoxins and mitotic blockers than vincristine (112), paclitaxel (114), and nocodazole (145). 86 Rj = R 2 = Me 98 Rj = H, R 2 = Me 99 R!=R 2 = H 100 R = H 101 R = Me The promising biological activity of the hemiasterlins has caused an increased demand for these peptides, resulting in our supply being almost exhausted. Isolating more of these 206 compounds from Cymbastela sp. is not practical, since Cymbastela sp. is very rare. Another possibility, based on the belief that the hemiasterlins and criamides are microbial metabolites of associated microorganisms harbored within the sponge,4 is to identify and culture the microbial source of the peptides. Efforts in this area are ongoing, but thus far have been unsuccessful. Therefore, in order to provide the quantities of the hemiasterlins and criamides required for complete evaluation of their potential as anticancer drugs, a general synthetic route to this family of peptides was developed.5 This synthetic methodology will also allow for the synthesis of analogues so that a complete structure activity study can be performed. 146 TV-Methyl amino acids are becoming common constituents of peptides and depsipeptides isolated from plants, microbes, and marine organisms.6 Some of these peptides, such as jaspamide (76), the geodiamolides (77-83),7,8 the hemiasterlins (86, 98, 99),2 dolastatine 10 (115),9 and cyclopsporine (146),10 have therapeutically proven or promising biological activity. 207 The addition of TV-methylated amino acids to a peptide has been observed to affect the peptide's conformation and activity.11 Peptide syntheses involving coupling reactions with hindered TV-methylated amino acids and amino acids with bulky side chains are difficult, and standard reagents are often inefficient 19 or completely ineffective. The decreased nucleophilicity of an TV-methyl nitrogen towards acylation is due to its increased steric bulk and is responsible for the poor yields observed in coupling reactions.12 Good results for these couplings have been obtained with Dpp-Cl (147)13 and BOP-C1 (148),14 but both of these reagents are difficult to use due to their high reactivity with water and amines.11 Coste et al. have shown that the reagent BroP (149),15'16 in which the oxybenzotriazole (-OBt, 150) moiety of BOP (151) has been replaced by bromine,17 allows for efficient coupling of TV-methylated amino acids, and the results are comparable to those obtained with Dpp-Cl (147) and BOP-C1 (148).18 However, the manufacture and use of BOP coupling agents involve the use or the formation of hexamethylphosphoric triamide (HMPA, 152), a powerful carcinogen.18 Fortunately BOP (151), BOP-C1 (148) and BroP (149) can be replaced by their pyrrolidino homologues, PyBOP® (153), PyCloP (154), and PyBroP (155), reagents which are converted to tripyrrolidino phosphine oxide (TPPA, 156), a non-carcinogenic 1 $1 by-product. (Of 147 148 X = C1 (BOP-C1) 154 X = C1 (PyCloP) 149 X = Br (BroP) 155 X = Br (PyBroP) 151 X = OBt (BOP) 153 X = OBt (PyBOP®) 208 The proposed mechanism for the amide coupling involving PyBroP (153) is outlined in Scheme 6.16 Formation of the transient acyloxyphosphonium salt 158 results from the reaction of the TV-protected amino acid 157 with 153. Trapping of 158 by the C-protected amino acid 159 affords the dipeptide 160. It is also possible that other reaction intermediates, such as the symmetric anhydride 161 or the acyl halide 162, are formed during the reaction sequence.16 H O R 2 160 Scheme 6. Proposed mechanism of PyBroP-mediated peptide coupling. 209 When coupling certain Boc protected amino acids with PyBOP®, PyBroP, and DCC, a side-reaction has been observed in which an TV-methyl-TV-carboxyanhydride (166) is formed.19 The activation of TV-Boc-TV-Me-val-OH with PyBroP/DIEA leads immediately to a 1:1 mixture of the symmetrical anhydride 164 and 166, which accounts for the poor yield of the desired peptide. Loss of a r-butyl cation from the oxazolonium ion 165 is thought to account for the formation of 166 (Scheme 7).19 This, however, does not appear to be the case with all TV-Boc protected amino acids. For example, when 1.5 equivalents of TV-Boc-7V-Me-leu-OH were coupled with one equivalent of yV-Me-val-OMe, an 88% yield of the desired dipeptide was obtained, despite the rapid formation of an TV-methyl-TV-carboxyanhydride during the pre-activation reaction.19 V * > * JX 169 168 167 166 Scheme 7. Formation of 166 from TV-Boc-TV-Me-Val-OH (163). Our strategy for the elaboration of the hemiasterlins and structural analogs was a convergent approach centering around the synthesis of the two terminal amino acids (Scheme 8). Following Schreiber's methodology for the formation of vinylogous amino acids, a general route to TV-methyl-vinylogous amino acids (ex. 173) was devised. A general route to TV-Boc-methylated tryptophan amino acids, such as 170, was developed by Dr. D. Wallace, a researcher in Dr. E. Piers' laboratory at the University of British Columbia. Scheme 9 briefly summarizes 210 the experimental details for Dr. Wallace's synthesis of enantiomerically pure N'-Boc-N,N',/3,j3-tetramethyltryptophan (170) (for more information on this synthesis please see: Tetrahedron Letters, 38, 317, (1997)). tert-Leucine (133) is commercially available in the Boc (171) and Fmoc (172) protected forms, in both the D and L configurations. Described herein are the syntheses of N-Boc-MHVV-OEt (173), (-)-hemiasterlin (86), and several analogs. 212 182 R = N 3 181 183 R = N H 2 184 R = N(H)Boc 1 (95%) T 185 186 R = H 170 R = Me Scheme 9. Synthesis of TV-Boc-TV./V'^ /i-tetramethyltryptophan (170). a: KN(SiMe 3 ) 2 (3 equiv), THF, -78 to 0°C, 3h; Mel , -78 to 0°C, 2h. b: KN(SiMe 3 ) 2 (1.5 equiv), then as in a. c: i-B u 2 A l H , E t 2 0. d: TPAP, N M O , D C M , 4A molecular sieves, e: Ph 3P=CHOMe, THF, r.t. f: p-TsOH, H 2 0 , dioxane, 60°C, 16 h. g: NaC10 2, N a H 2 P 0 4 , 2-methylbut-2-ene, /-BuOH, H 2 0 , 0°C. h: M e 3 C C O C l , E t 3 N, THF, -78°C; 180. i: KN(SiMe 3 ) 2 (1.1 equiv), THF, -78°C; 2,4,6-triisopropylbenzenesulfonyl azide, THF, -78°C, 1 min; HOAc, 30-40°C, 1 h. j: SnCl 2 , dioxane, H 2 0 , r.t., 36 h. k: (Me 3 CO) 2 0, dioxane, H 2 0 , r.t, 16 h. 1: L i O H , H 2 0 2 , THF, H 2 0 , r.t., 16 h; citric acid, H 2 0 . m: NaH (~6 equiv), D M F ; Mel , r.t. 16 h. n: L i O H , MeOH, H 2 0 , 60°C, 24 h; citric acid, H 2 0 . 5.2 RESULTS AND DISCUSSION. 213 5.2.1 TV-Methyl Homo Vinylogous Valine (MHVV) (173) A general route to the synthesis of vinylogous amino acids has been reported by Schreiber,20 where the TV-Boc protected amino acids are converted to the corresponding aldehydes (189) via the Weinreb amides (188) (Scheme 10). Scheme 10. Schreiber's route to vinylogous amino acids. Thus, the structurally novel amino acid salt 193 was prepared using the modified Schreiber procedure as summarized in Scheme 11. Conversion of TV-Boc amino acids to their TV'-methoxy-TV'-methyl-TV-Boc-amides (Weinreb methoxamides) via DCC or BOP-mediated coupling with the commercially available TV,0-dimethyl-hydroxylamine hydrochloride has been reported to proceed efficiently with yields of 78-99%.20"21 Attempts to form the Weinreb methoxamide derivative of (S)-TV-Boc-TV-214 Boc Boc 163 PyBop, DIEA HCl-NMe(OMe) CH 2 C1 2 , 0°C to rt 1 hour 85% Boc. (Ph)3F OEt o CH 2 C1 2 , rt, 4 hours 173 75% TFA/CH 2 C1 2 Ar, rt, 0.5 hours "100%" OEt 193 Scheme 11. Synthesis of M H V V - T F A (193). OMe L i A l H 4 THF -78°C to 0°C 92% Boc 177 methylvaline (191) using DCC methodology resulted in a yield of only 60%. Alternatively, the PyBOP-mediated coupling provided an 85% yield of 191. Reduction of the Weinreb amide 191 with excess L i A l H 4 2 1 , 2 2 afforded TV-Boc-TV-methylvalinal (177) in 92% yield. Reaction of 177 with [(l-ethoxycarbonyl)ethylidene]triphenylphosphorane in CH2CI2 afforded, stereoselectively, a 75%) yield of the 2s-2-alkenoate 173. Trifluoroacetic acid mediated cleavage of the /V-Boc 215 group led to the quantitative formation of the required ammonium trifuoroacetate derivative 193. The overall yield of 193 from TV-Boc-TV-methylvaline (163) was 58%. 5.2.2 ( - ) -Hemiaster l in (86) The synthesis of (-)-hemiasterlin (86) from the three protected constituent amino acids 170,171, and 173 is outlined in Scheme 12. The coupling of (S)-TV-Boc-terMeucine (171) with the amino acid salt 193 is a sterically demanding reaction and, as a result, is very difficult to perform. DCC-Mediated coupling of 171 and 193 gave only recovered starting materials, while coupling with PyBOP 2 3 and PyBroP2 4 resulted in poor yields of 194 (9 and 22%, respectively). These low yields are likely the result of a combination of steric factors and the formation of the jV-Me-NCA. Success was finally achieved using 2,2-dimethylpropanoyl chloride (pivoyl chloride) to pre-activate (S)-YV-Boc-ter/-leucine (171). Thus, reaction of 171 with pivoyl chloride in the presence of one equivalent of DIEA gave the corresponding mixed anhydride, which, upon treatment with amino acid salt 193 and two equivalents of DIEA, afforded the protected dipeptide 194 in 62% yield. TFA mediated cleavage of the 7V-Boc group converted 194 to its TFA salt, which immediately underwent a PyBrop-mediated coupling to yV'-Boc-7V,7V',/7,/i-tetramethyltryptophan (170) in the presence of 2.6 equivalents of DIEA and 0.6 equivalents of DMAP, affording TV-Boc protected hemiasterlin ethyl ester (195) in 53% yield. Base hydrolysis of the ester function of 195, followed by TFA promoted removal of the iV-Boc group and purification of the resultant crude product by reversed phase HPLC (0.05% 216 195 1. LiOH, H 20/MeOH, rt, 16 h; citric acid 2. TFA/CH 2C1 2 1:1, 0.5 h 83% T 86 Scheme 12. Synthesis of (-)-hemiasterlin (86). TFA/H20:MeOH 1:1), provided an 83% yield of (-)-hemiasterlin (86) as an amorphous white solid. The synthetic (-)-hemiasterlin (86) had spectroscopic data (*H NMR, 1 3 C NMR, 217 HRFABMS), and in vitro cytotoxic and antimitotic ICso's against human breast cancer MCF7 cells identical with the those of the natural (-)-hemiasterlin (86) isolated from Cymbastela sp. Although this general approach to (-)-hemiasterlin (86) does produce the tripeptide, the overall yield is only 27% from TV-Boc-ferMeucine (171). Some improvement may be possible by using Z or Fmoc protected amino acids in the coupling steps. Initial attempts using TV-Fmoc-te/t-leucine (172) in the PyBroP-mediated coupling for the formation of the dipeptide 194 showed no improvement over the Boc analogue. Replacement of the Boc protecting group on Af,/V',/?,/^ tetramethyltryptophan (170) by either a Z or Fmoc introduces some potential problems. For example, the final step in the synthesis of 170 involves the 7Y-methylation of the protected amine 186 with Mel and NaH, and because Fmoc is base sensitive, it probably would not survive this procedure. Z, on the other hand, would work well for the Af-mefhylation, but could be problematic in the deprotection step. Removal of Z is usually performed under strongly acidic conditions,25 in which the tryptophan ring may decompose, or by catalytic hydrogenation,25 which would also reduce the double bond of MFfVV. The synthesis of 170 is currently being streamlined by Dr. J. Nieman of Dr. Piers' laboratory, and attempts to use different protecting groups are being made. 218 5.3 STRUCTURE ACTIVITY STUDY. In the hope of being able to synthesize a simpler, equally active analogue of hemiasterlin (86), and in order to determine what aspects of its structure are important for its biological activity, a structure activity study of the hemiasterlins has been undertaken. Since the hemiasterlins (86, 98, 99) caused mitotic arrest and cytotoxicity towards MCF-7 cells in the same concentration range, for the simplification of this discussion, any reference to activity refers only to the cytotoxic effects of the peptides. The varying activities observed for the three naturally occurring peptides of the hemiasterlin family give some clues as to what aspects of the tripeptide structure are required for activity. Removal of the indole yV-Me from hemiasterlin (86) to form hemiasterlin A (98) results in a decrease in activity by a factor of 7, while removal of the indole /V-Me and replacement of ter/-leucine by valine in 86 to form hemiasterlin B (99) results in a decrease in activity by a factor of 23. These observations imply that the central amino acid must be tert-leucine and the indole must be /V-methylated for the peptide to show high activity. Four derivatives of hemiasterlin (86) were synthesized by making simple modifications to its structure. Esterification of 86 with diazomethane afforded hemiasterlin methyl ester (144) in quantitative yield. The formation of the methyl ester had little affect on the activity (IC50 0.5 nM). Reduction of the MHVV double bond in 86 using H 2 and Pd/charcoal afforded 27,28-dihydrohemiasterlin (196) in 95 % yield. Reduction of the double bond resulted in the activity being reduced by a factor of 8 (IC50 2.5 nM). Acetylation of the amino nitrogen with acetic anhydride and pyridine gave /V'-acetyl-hemiasterlin (197) in 80 % yield. Acetylation of the amine had a dramatic affect on the activity, reducing it by a factor of 1000 (IC50 300 nM). Methylation of 86 with potassium carbonate and methyl iodide resulted in the formation of the 219 potassium salt of amino-7Y', '^-dimethyl-hemiasterlin-methyl ester (198) in 92 % yield. 198 was completely inactive (IC5o<1000 nM). From the activity results from these four simple derivatives of hemiasterlin (86), several observations can be made. It appears that esterification of the acid function on 86 has little or no effect on its activity, implying that either the acid function is not involved at the active site or that there is a lipase enzyme present that is able to hydrolyze the ester 196 back to hemiasterlin (86). Reduction of the double bond causes a significant decrease in the activity, which suggests that there may be a change in the conformation of the peptide, since 196 is not be as rigid as hemiasterlin (86), and this decreases the activity. Addition of substituents to the amino nitrogen drastically reduces the activity, indicating that the amino portion of the molecule is very important for high activity. Using the general approach employed for the synthesis of (-)-hemiasterlin (86), several synthetic analogues of 86 have been synthesized and tested for cytotoxicity and the ability to stop cell mitosis against the human breast cancer cell line MCF7. The structures of the analogues prepared and their associated biological activities are summarized in Table 20. When the A'.A '^^ yS-tetramethyltryptophan of 86 is replaced by A^/?,/krimethyltryptophan (compound 199), the activity decreases by a factor of 160 (IC50 50 nM) compared to 86. Synthesis of 201, using A/^V-dimethyltryptophan, resulted in a further decrease in the activity (IC50 60 nM). Replacing A /^V'.^ yS-tetramethyltryptophan with tryptophan (compound 201) also resulted in a decrease in activity relative to 86, but only by a factor of 130 (IC50 50 nM). This final result was surprising, since it was expected that the activity would drop as the number of methyl substituents on the tryptophan moiety decreased from three (199) to two (200) to none (201). Apparently, the relationship between the number of methyl substituents on the tryptophan moiety and the activity is not a simple one. What is clear is that removal of the methyl substituents from 221 the tryptophan moiety results in a large decrease in activity. Replacement of the MHVV moiety in hemiasterlin (86) with A^-methyl-4-amino-butyric acid (compound 202) results in a complete loss of activity. This result indicates the importance of having alkyl groups on the C-terminal amino acid. Other derivatives which contain modifications in the C-terminal amino are currently being synthesized. When the activity results for these peptides are available, some generalizations about what components are required for activity can be made. When terr-leucine is replaced with valine or butyrine, as in the synthesis of 203 and 204, respectively, the activity of the tripeptides was only slightly affected (IC50 50 nM for 203 and 204). It appears that the size of the alkyl group on the central amino acid has little bearing on the activity. Analogs with alanine and glycine as the central amino acid in the tripeptide will be synthesized to complete the series. The results from the preliminary structure activity study of the hemiasterlins suggest that changing the alkyl substituents on the two terminal amino acids drastically effects the activity, but changing the alkyl group of the central amino acid causes little change. Preparation of several other structural analogs is currently underway. 222 5.4 EXPERIMENTAL 5.4.1.1 Na-Boc-Na-methyl-l-valine N-methoxy-N-methylamide (191). To a cold (0°C) solution of Boc-TV-methylvaline (163) (5.0 g, 21.6 mmol), N,0-dimethylhydroxylamine hydrochloride (2.8 g, 28 mmol), and PyBOP (11.2 g, 22 mmol) in CH2CI2 (22 mL) was added DIEA (8.4 pL, 75 mmol). After 1 min., the reaction was warmed to room temperature and stirring continued for 1 hour. If the pH value of the mixture was less than 7, the mixture could be neutralized by adding a few drops DIEA to allow the reaction to go to completion. The mixture was poured into 200 mL of Et20 and washed successively with 3 N hydrochloric acid (3x30 mL), saturated sodium hydrogen carbonate solution (3 x 30 mL), and saturated sodium chloride solution (3x30 mL). The organic layer was dried with magnesium sulfate and the solvent evaporated, followed by chromatography of the crude product (silica gel, 3:1 pet. ether/Et20), to afford 191 (4.46 g, 75%) as a colourless oil. IR (neat): 3550, 3350, 2940, 1693, 1668, 1475, 1392 cm'1. 'H NMR (200 MHz, C D C I 3 , 8): 0.84 (d, J = 6.6 Hz, 4H, (CH3)2), 0.85 (d, J = 6.6 Hz, 2H, (CH3)2), 1.41 (s, 6H, Boc-(CH3)3), 1.44 (s, 3H, Boc-(CH3)3), 2.15-2.30 (m, IH, CH), 2.75 (s, IH, NCH 3), 2.78 (s, 2H, NCH 3), 3.1 (bs, 3H, NCH 3), 3.64 (s, IH, OCH 3), 3.68 (s, 2H, OCH3), 4.66 (d,J= 10 Hz, 0.4H, CH), 4.95 (d, J = 10 Hz, 0.6H, CH). 1 3 C NMR (50 MHz, CDC13, 8): 18.1, 19.0, 19.4, 26.7, 26.9, 28.1, 28.2, 28.8, 29.3, 31.5, 57.7, 59.5,61.3,79.2,79.6, 155.8, 155.3. exact mass calculated for Ci 3H 270 4N2 (M+H)+: 275.19708; found: 275.19710. Elemental analysis for C13H26O4N2: calculated: C 56.91 H 9.55 N 10.21 found: C 56.92 H 9.41 N 10.40 [oc]D25: +128.3 (c 2.9, CHCI3). 223 5.4.1.2 /V-Boc-Af-methyl-l-valinal (177). Lithium aluminum hydride (875 mg, 23 mmol) was added to a solution of 7Va-Boc-7Va-methyl-l-valine N-methoxy-YV-methylamide (191) (2.0 g, 7.7 mmol) in dry THF (8 mL) and the reaction stirred for 20 min. The mixture was poured into a stirring solution of potassium hydrogen sulphate (3.14 g, 23 mmol) in water (100 mL). Ether (75 mL) was added, the layers separated and the aqueous layer extracted with ether (3 x 50 mL). The organic layers were combined, washed with 3 N hydrochloric acid (3x30 mL), saturated sodium hydrogen carbonate solution (3 x 30 mL), and saturated sodium chloride solution (3 x 30 mL), dried with magnesium sulfate and the solvent evaporated to yield the crude aldehyde 177 (1.52 g, 92%). Aldehyde 177 was used without further purification. Note: 177 can be stored under argon for ~ 2 weeks, but when stored in organic solvents at room temperature, undergoes slow decomposition. IR(neat): 3500, 2970, 1736, 1699, 1475, 1386 cm"1. 'H NMR (200 MHz, CDC13, 6): 0.73 (d, J = 6.9 Hz, 3H, CH 3), 0.91 (d, J = 6.9 Hz, 3H, CH 3), 1.27 (s, 9H, Boc-(CH3)3), 2.02-2.15 (m, IH, CH), 2.63 (s, 2H, NCH 3), 2.72 (s, IH, NCH 3), 3.44 (d, J = 9.5 Hz, 0.5H, CH), 3.86(d, 7=9 Hz, 0.5H, CH), 9.45 (s, IH, CH). 1 3 C NMR (50 MHz, CDC13, 8): 18.9, 19.9, 20.3, 26.2, 27.1, 27.9, 32.7, 34.1, 70.1, 71.9, 79.9, 80.4,198.2,198.8. exact mass calculated for C i i H 2 2 N 0 3 (M+H)+: 216.15997; found: 216.15996. Elemental analysis for CnH 2 iN0 3 : calculated: C 61.37 H 9.83 N 6.51 found: C 61.28 H 9.78 N 6.54 [oc]D25: -104.2 (c 5.5, CHC13). 224 5.4.1.3 A/-Boc-A7-methyI-4-amino-2,5-dimethylhex-2-enoic acid (173). To a solution of aldehyde 177 (1.75 g, 8.7 mmol) in dry CH2CI2 (9.0 mL) under an argon atmosphere at room temperature was added [(l-ethoxycarbonyl)ethylidene]triphenylphosphonane (4.19 g, 11.3 mmol) and stirring continued for a four hours. The reaction mixture was diluted with water (100 mL) and extracted with ether (3 x 100 mL). The combined organic extracts were washed with saturated sodium chloride solution (100 mL), dried with magnesium sulfate and concentrated in vacuo. The crude oil was purified by flash chromatography (silica gel, 23:2 pet. ether/ Et20) to afford the required E-2-alkenoate 173 as a colourless oil (2.13 g, 82 %). IR (neat): 3600, 3400, 2972, 1714, 1655, 1467, 1392 cm"1. *H NMR (200 MHz, CDCI3, 5): 0.74 (d, J = 6 Hz, 3H, CH 3), 0.79 (d, J = 6 Hz, 3H, CH 3), 1.17 (t, J= 7 Hz, 3H, CH 3), 1.34 (s, 9H, Boc-(CH3)3), 1.69-1.75 (m, 1H, CH), 1.78 (s, 3H, CH 3), 2.60 (bs, 3H, NCH 3), 4.08 (q, J= 7 Hz, 2H, CH 2), 4.15-4.20 (m, 0.5H, CH), 4.21-4.32 (m, 0.5H, CH), 6.54 (d, J=8Hz , 1H, CH). 1 3 C NMR (50 MHz, CDC13,5): 13.0, 14.0, 18.3, 19.3, 28.2, 28.8, 31.1, 59.4, 79.3,130.8,138.6, 155.5, 167.7. exact mass calculated for C, 6H 3oN0 4 (M+H)+: 300.21750; found: 300.21754. Elemental analysis for C16H29NO4: calculated: C 64.19 H 9.76 N 4.68 found: C 64.40 H 9.96 N 4.75 [oc]D25: +61.1 (c9.1,CHCl 3). 225 5.4.1.4 Trifluoacetic acid mediated cleavage of N-Boc groups: General procedure 1. /V-Boc-amino acid ester (1.0 equiv.) was treated with TFA /CH2CI2 (1 mL/0.1 mmol) at room temperature for 0.5 hours. Removal of the excess solvent in vacuo, followed by repeated rinsing with CH2CI2 (3x5 mL) and evaporation afforded the TFA salt of the amino acid ester in quantitative yield. TFA salts were used without further purification. 5.4.1.5 Trimethylacetyl chloride mediated peptide coupling: General procedure 2. To a cold (-78°C) stirred solution of acid 26 (1.1 equiv.) in dry THF (1 mL/mmol) under an argon atmosphere was added DIEA (1.5 equiv.) and trimethylacetyl chloride (1.2 equiv.), and the resulting mixture was warmed to 0°C for one hour and then re-cooled to -78°C. DIEA (2.2 equiv.) was added via cannula to the reaction flask followed by the addition via cannula of the TFA salt of the amino acid ester (1.0 equiv.) in dry THF (0.5 mL/mmol) at -78°C. Stirring was continued for one hour, water (40 mL) was added, the reaction mixture was allowed to warm to room temperature, and was then extracted with ether (3 x 50 mL). The combined organic extracts were washed with saturated sodium chloride solution (50 mL), dried over magnesium sulfate, and concentrated in vacuo. The crude oil was purified by flash chromatography (silica gel, ether-pet. ether) to afford the desired dipeptide as a colorless oil. 5.4.1.6 A'-Boc-5'-ter^-leucine-A7-methyl-4-amino-2,5-dimethylhex-2-enoic ethyl ester (194). 194 Following general procedure 2, ester 183 was prepared with the following quantities of reagents and solvents: tert-leucine (173), 156 mg (0.52 mmol); trimethylacetyl chloride, 64 pL (0.52 mmol); DIEA, 99 pL (0.57 mmol); JV-Boc-MHVV-OEt (171), 110 mg (0.47 mmol); DIEA, 198 pL (1.14 mmol); THF, 7 mL. Purification of the crude product by flash chromatography (silica gel, 1:5 ether-pet. ether) afforded 121 mg of 194 (62 %). IR(neat): 2966, 1710, 1641, 1493, 1390 cm"1. 226 'H NMR (200 MHz, CDC13, 8): 0.76 (d, J = 6 Hz, 3H, CH 3), 0.80 (d, J = 6 Hz, 3H, CH 3), 0.88 (s, 9H, (CH3)3), 1.22 (t, J = 7 Hz, 3H, CH 3), 1.33 (s, 9H, Boc-(CH3)3), 1.79-1.89 (m, IH, CH), 1.83 (s, 3H, CH 3), 2.91 (s, 3H, NCH 3), 4.12 (q,J=7 Hz, 2H, CH 2), 4.35 (d, J = 10 Hz, IH, CH), 5.03 (t,J= 10 Hz, IH, CH),5.14(d, J= 10 Hz, IH, NH), 6.57 (d,7=8Hz, IH, CH). 1 3 C NMR (50 MHz, CDC13, 8): 13.7, 14.1, 18.5, 19.4, 26.3, 28.2, 30.0, 30.9, 34.8, 55.8, 56.2, 60.3, 79.4, 132.5, 138.5, 156.0, 167.6, 172.2. exact mass calculated for C22H41N2O5 (M+H)+: 413.30154; found: 413.30119. Elemental analysis for C22H40N2O5: calculated: C 64.04 H 9.69 N 6.79 found: C 64.40 H 9.89 N 7.00 [a]D 2 5 = -76.9° (c 2.43, CHC13). 5.4.1.7 PyBroP-mediated peptide coupling: General procedure 3. A solution (or suspension) of the of W-Boc protected amino acid (1.0 equiv.), the TFA salt of the amino acid ester (1.0 equiv.), and PyBroP (1.0 equiv.) in CH2CI2 (1 mL/mmol) was treated with DIEA (2.6 equiv.) and DMAP (0.6 equiv.) at 0°C. The ice bath was removed after 1 min. and stirring continued at room temperature for 16 to 24 hours. The mixture was poured into EtOAc (50 x the volume of DCM) and washed with 3 N hydrochloric acid (3 x 30 mL), saturated sodium hydrogen carbonate solution (3x30 mL), and saturated sodium chloride solution (3 x 30 mL). The organic layer was dried with magnesium sulfate and the solvent evaporated, followed by flash chromatography (silica gel, pet. ether-Et20) of the crude product to afford the peptide as a colourless oil. 5.4.1.8 /V'-Boc-O-hemiasterlin-ethyl ester (195). 195 Following general procedure 3, ester 195 was prepared with the following quantities of reagents and solvents: /V'-Boc-/V)7V',/i,/?-tetramethyl-5'-tryptophan (170), 50 mg (0.14 mmol); TV-Boc-S-terMeucine-Af-methyl-4-amino-2,5-dimethylhex-2-enoic ethyl ester (194), 74 mg (0.18 mmol); PyBrop, 65 mg (0.14 mmol); DIEA, 72 uL (0.41 mmol); DMAP, 10 mg (0.08 mmol); CH 2C1 2, 227 2 mL. Purification of the crude product by flash chromatography (silica gel, 3:7 ether-pet. ether) afforded 48 mg of 195 (53 %). 'H NMR (400 MHz, CDC13, 5): 0.42 (s, 6H, (CH3)3), 0.46 (s, 3H, (CH3)3), 0.82 (t, J = 6.7 Hz, 3H, CH 3), 0.86 (d, J= 6.4 Hz, 3H, (CH3)3), 1.26 (t, J= 7.3 Hz, 3H, CH 3), 1.40 (s, 1H), 1.42 (s, 1H, CH 3), 1.49 (s, 6H, Boc-(CH3)3), 1.51 (s, 2H, CH 3), 1.52 (s, 3H, Boc-(CH3)3)), 1.54 (s, 2H, CH 3), 1.84 (d, J= 1.2 Hz, 3H, CH 3), 1.8-1.90 (m, 1H, CH), 2.92 (s, 3H, CH 3), 2.99 (s, 3H, CH 3), 3.72 (s, 2H, CH 3), 3.73 (s, 1H, CH 3), 4.09 (q, J = 7 Hz, 0.3 H, CH 2), 4.16 (q, J = 7 Hz, 0.7 H, CH 2), 4.37 (d, J = 8.6 Hz, 0.7 H, CH), 4.45 (d, J = 8.8 Hz, 0.3H, CH), 5.02 (t, J = 10.4 Hz, 1H, CH), 5.64 (s, 0.3H, CH), 5.97 (s, 0.7H, CH), 6.12 (d, J = 8.6 Hz, 0.3H, NH), 6.18 (d, J= 8.5 Hz, 0.7H, NH), 6.60 (dd, J = 9, 1.2 Hz, 1H, CH), 7.05 (s, 0.3H, CH), 7.06 (s, 0.7H, CH), 7.16 - 7.30 (m, 3H, 3CH), 7.95 (d, J = 8.2 Hz, 0.3H, CH), 8.27 (d, J = 7.6 Hz, 0.7H, CH). I 3 C NMR (100 MHz, CDC13, 8): 13.8, 14.2, 18.8, 19.4, 24.1, 24.3, 25.9, 27.7, 28.4, 28.6, 29.7, 30.1, 30.3, 31.1, 32.6, 32.7, 33.8, 34.1, 34.4, 34.7, 39.5, 39.6, 54.8, 55.2, 55.9, 60.8, 63.0, 64.2, 77.2, 79.5, 80.6, 109.3, 109.7, 118.8, 119.7, 120.7, 121.4, 121.6, 121.9, 121.9, 122.1, 124.9, 125.0, 126.6, 126.8,132.4, 132.5, 138.0, 138.8, 139.0, 156.3, 157.3,167.7, 169.7, 170.4, 171.3, 171.4. exact mass calculated for C 3 7 H 5 9N 4 06 (M+H)+: ,655.44346; found: 655.44237. Elemental analysis for C37H58N4O6: calculated: C 67.86 H 8.93 N 8.56 found: C 67.98 H 8.70 N 8.34 5.4.1.9 Hydrolysis of esters and removal of iV-Boc protecting groups: General procedure 4. To a stirred solution of the ethyl ester (1 equiv.) in 1:1 MeOH/H 20 (1 mL/ 0.01 mmol) at room temperature was added lithium hydroxide (8 equiv. of a 1.0 M aqueous solution), and the reaction mixture was then warmed to 60°C and stirred for 20 hours. The reaction mixture was acidified by drop wise addition of 1.0 M citric acid and extracted with ethyl acetate (3x10 mL). The combined organic extracts were washed with saturated sodium chloride solution (2x5 mL), dried with sodium sulfate and concentrated in vacuo. Under an argon atmosphere, the crude oil was treated with 1:1 TFA/CH 2C1 2 (1 mL/0.1 mmol) at room temperature for 0.5 hours. Removal of the excess solvents in vacuo, followed by repeated rinsing with CH 2C1 2 (3x5 mL) and evaporation, afforded the TFA salt. HPLC purification of the crude product (0.05% TFA/H20:MeOH) afforded peptides as white powders. 228 5.4.1.10 (-)-hemiasterlin (86). 86 Following general procedure 4, hemiasterlin (86) was prepared with the following quantities of reagents and solvents: W-Boc-hemiasterlin ethyl ester (195), 10 mg (0.015 mmol); lithium hydroxide (1.0 M) 122 pL (0.12 mmol), 1:1 TFA/CH 2C1 2 (1 mL). Purification of the crude product by reversed phase HPLC (1:1 0.05% TFA/H20:MeOH) afforded 6.5 mg of 86 (83 %). IR (neat): 3412, 2962, 1650, 1635 cm"1. 'H NMR (400 MHz, DMSO-J 6, 5): 0.78 (d, J = 7 Hz, 3H, CH 3), 0.80 (d, J = 7 Hz, 3H, CH 3), 0.99 (s, 9H, (CH3)3), 1.37 (s, 3H, CH 3), 1.40 (s, 3H, CH 3), 1.80 (s, 3H, CH 3), 1.99-2.08 (m, IH, CH), 2.22 (bs, 3H, CH 3), 3.02 (s, 3H, CH 3), 3.75 (s, 3H, CH 3), 4.41 (d, J = 10.3 Hz, IH, CH), 4.85 (d, J = 9.4 Hz, IH, CH), 4.93 (t, J = 10.1 Hz, IH, CH), 6.66 (dd, J = 10, 1.1 Hz, IH, CH), 7.08 (t, J = 8 Hz, IH, CH), 7.17 (s, IH, CH), 7.20 (t, J = 8 Hz, IH, CH), 7.34 (bs, IH, NH), 7.45 (d, J = 8 Hz, IH, CH), 8.11 (d, J = 8 Hz, IH, CH), 8.85 (bs, IH, NH), 8.89 (d, J = 8.4 Hz, IH, NH). 1 3 C NMR (100 MHz, DMSO-cfe, 5): 13.5 (C 3 4), 18.9 (C 3 3), 19.3 (C 3 2), 22.6 (Ci5), 26.3 (C 2 2-C 2 4), 27.0 (C 1 4), 28.7 (C 3 1), 31.1 (C 3 0), 32.4 (C 1 3), 33.4 (C 1 7), 34.6 (C 2 1), 37.5 (C 1 0), 55.6 (C 2 6), 56.2 (Ci9), 67.5 (Cn), 110.0(C8), 116.5 (C3), 118.4 (C6), 120.6 (C5), 121.1 (C7), 125.0 (C4), 128.7 (C2), 131.6 (C 2 8), 137.7 (C9), 138.3 (C 2 7), 166.0 (Ci2), 168.5 (C 2 9), 170.1 (C 2 0). exact mass calculated for C 3 0 H 4 7 N 4 O 4 (M+H)+: 527.35973; found: 527.36095. UV (MeOH) Xmm (s) 216 (15,400), 273 nm (1,600). 229 5.4.2 Structure Activity Study 5.4.2.1 26,27-dihydro-hemiasterlin (196). 196 Hemiasterlin (86) (3 mg) was stirred overnight under in a sealed Erlenmeyer flask with a positive pressure of hydrogen with ethanol (2 mL) and Pd/C catalyst (~1.0 mg). Celite® (50 mg) was added and the resulting slurry filtered. Excess solvents were removed in vacuo and the crude material was purified by reversed phase HPLC (1:1 0.05% TFA/H 20:MeOH) to afford 2 mg of 196 (67 %). 'H NMR (MeOD-J6, 500 MHz, 8): 0.80 (d, J = 6.4 Hz, 3H), 1.01 (d, J = 6.4 Hz, 3H), 1.10 (s, 9H), 1.19 (d, J = 7.1 Hz, 3H), 1.25-1.32 (m, 2H), 1.45 (s, 3H), 1.52 (s, 3H), 2.01-2.09 (m, 2H), 2.43 (s, 3H), 3.05 (s, 3H), 3.79 (s, 3H), 4.45 (s, 1H), 5.03 (s, 1H), 5.30-5.36 (m, 1H), 7.12 (s, 1H), 7.13 (t, J= 8.1 Hz, 1H), 7.24 (t, J = 8.1 Hz, 1H), 7.42 (d, J= 8.1 Hz, 1H), 8.07 (d, J = 8.1 Hz, 1H); exact mass calculated for C30H49N4O4 (M+H)+: 529.37538; found: 529.37593. 5.4.2.2 Hemiasterlin-TV'-acetyl (197). 197 Hemiasterlin (86) (5 mg) was stirred overnight under a nitrogen atmosphere with pyridine (0.5 mL) and acetic anhydride (0.5 mL). Excess reagents were removed in vacuo and the crude material was purified by reversed phase HPLC (3:7 0.05% TFA/H 20:MeOH) to afford 4 mg of 197 (74 %). lH NMR (DMSO-J 6, 500 MHz, 8): 0.40 (s, 9H), 0.74 (d, 3H, J = 8.4 Hz), 0.79 (d, 230 3H, J=8Hz) , 1.29 (s, 1.5H), 1.47 (s, 1.5 H), 1.57 (s, 1H), 1.61 (s, 2H), 1.73 (s, 2 H), 1.78 (s, 1H), 1.1-1.97 (m, 1H), 2.08 (s, 3 H), 2.87 (s, 2H), 2.89 (s, 1H), 2.97 (s, 1H), 3.04 (s, 2H), 3.70 (s, 1H),3.73 (s, 2H), 4.33 (d, 1H, J= 10.6 Hz), 4.75 (d, 0.5H, J= 11.4 Hz), 4.86-4.94 (m, 1.5H), 6.26 (s, 0.4H), 6.29 (d, 0.6H, J = 11.1 Hz), 6.60 (dd, 0.6H, J = 11.8, 1.5 Hz), 6.64 (d, 0.4H, J = 11.4 Hz), 6.98-7.06 (m, 1H), 7.10-7.19 (m, 1H), 7.23 (s, 0.6H), 7.35 (d, 0.4H, J= 10.3 Hz), 7.42 (d, 0.6H, J = 10.7 Hz), 7.96 (d, 0.4H, J = 10.3 Hz), 8.15 (d, 0.6H, J = 10.3 Hz), 8.48 (d, 0.4H, J = 11.1 Hz); exact mass calculated for C32H49N4O5 (M+H)+: 569.37029; found: 569.36751. 5.4.2.3 N -Permethyl-hemiasterlin methyl ester (198). Hemiasterlin (86) (5 mg, 0.009 mmol) was stirred overnight under a nitrogen atmosphere with potassium hydrogen carbonate (~2 mg) and methyl iodide (50 pL, 0.3 mmol) in DMF (2 mL). Excess reagents were removed in vacuo and purification of the crude product by reversed phase HPLC (2:1 0.05% TFA/H 20:MeOH) afforded 3.5 mg of 197 (64 %). 'H NMR (DMSO-J 6, 500 MHz, 5): 0.77 (d, 3H, J = 6.7 Hz), 0.78 (d, 3H, J = 6.7 Hz), 1.00 (s, 9H), 1.37 (s, 3H), 1.72 (s, 3H), 1.83 (s, 3H), 2.01-2.06 (m, 1H), 2.90 (s, 9H), 3.04 (s, 3H), 3.66 (s, 3H), 3.76 (s, 3H), 4.74 (d, 1H, J = 7.6 Hz), 4.87 (s, 1H), 4.92 (t, 1H, J = 10 Hz), 6.70 (dd, 1H, J= 1.5, 9.7 Hz), 7.14 (t, 1H, J= 7.6 Hz), 7.22 (t, 1H, J = 7.6 Hz), 7.39 (s, 1H), 7.46 (d, 1H, J = 7.6 Hz), 8.05 (d, 1H, J = 7.6 Hz), 8.83 (d, 1H, J = 7.3 Hz); exact mass calculated for C 3 3 H 5 3 N 4 O 4 (M+): 569.40668; found: 569.40687. 198 231 5.4.2.4 ^-Boc-5'-valine-^-methyl-4-amino-2,5-dimethylhex-2-enoic ethyl ester (199). Following the general procedure 2, ester 199 was prepared with the following quantities of reagents and solvents: YV-Boc-valine (163), 100 mg (0.46 mmol); TV-Boc-MHVV-OEt (171), 126 mg (0.42 mmol); trimethylacetyl chloride, 56 pL (0.46 mmol); DIEA part A 109 pL (0.63 mmol), part B, 160 pX (0.92 mmol); THF 7 mL. Purification of the crude product by flash chromatography (silica gel, 3:7 ether-pet. ether) afforded 121 mg of 199 (69%). IR (CHC13): 2970, 1705, 1637, 1410 o n 1 . *H NMR (200 MHz, CDCI3, 5): 0.82 (d, 3H, J = 6.8 Hz), 0.85 (d, 3H, J = 6.8 Hz), 0.86 (d, 6H, J = 6.8 Hz), 1.25 (t, 3H, J = 6.8 Hz), 1.38 (s, 9H), 1.82 - 1.94 (m, 2H), 1.84 (d, 3H, 7=1.5 Hz), 2.90 (s, 3H), 4.16 (q, 2H, 7=7.3 Hz), 4.31 (dd, IH, 7= 9.5, 6.8 Hz), 5.00 (dd, 1H,7= 10.4,9.5 Hz), 5.16 (d, IH, 7=9.3 Hz), 6.60 (dd, 1H,7 = 9.3, 1.5 Hz). 1 3 C NMR (50 MHz, CDCI3, 5): 11.5, 13.6, 14.2, 17.5, 18.7, 19.4, 19.5,28.6,29.9,30.3,30.9, 55.4, 56.3, 60.8, 79.4, 132.7, 138.3, 156.0, 167.7, 172.6. Exact mass calculated for C22H41N2O5 (M+H)+: 413.30154; found: 413.30167. 5.4.2.5 A'-Boc-^-buterine-A^-methyl^-ainino^jS-dimethylhex^-enoic ethyl ester (200). Following the general procedure 2, ester 200 was prepared with the following quantities of reagents and solvents: 7V-Boc-2-amino-butyric acid, 93.5 mg (0.46 mmol); /V-Boc-MHVV-OEt (171) 126 mg (0.42 mmol); trimethylacetyl chloride, 60 pL (0.50 mmol); DIEA part A, 109 uL (0.63 mmol), part B, 160 pL (0.92 mmol); THF 7 mL. Purification of the crude product by flash chromatography (silica gel, 2:3 ether-pet. ether) afforded 105 mg of 200 (63 %). 199 200 232 IR(CHC13): 2974, 1705, 1639, 1496 cm"1. 1 HNMR(200MHz,CDCl 3 ,5) : 0.82-0.91 (m, 9 H), 1.27 (t, 3H, J = 7.1 Hz), 1.39 (s,9H), 1.81 -1.89(m, 1H), 1.82 (d, 3H,J= 1.5 Hz), 1.91-1.95 (m, 1H), 2.88 (s, 3H), 4.17 (q, 2H, J=7.1 Hz), 4.44 - 4.52 (m, 1H), 4.99 (dd, 1H, J = 10.6, 9.3 Hz), 5.27 (d, 1H, J = 9 Hz), 6.61 (dd, 1H, J = 9, 1.5 Hz). 1 3 C NMR (50 MHz, CDC13, 5): 9.6, 11.5, 13.5, 14.1, 18.7, 19.4, 25.9, 28.7, 29.9, 51.6, 56.5, 60.8, 79.4, 132.8, 138.2,167.7,172.5. Exact mass calculated for C2iH 3 9N 205 (M+H)+: 399.28589; found: 399.28556. 5.4.2.6 iV-Boc-5-ter/-leucine-A7-methyl-4-amino-methylbutyrate (201). H O 201 Following the general procedure 2, ester 201 was prepared with the following quantities of reagents and solvents: ./V-Boc-terMeucine (173), 106 mg (0.42 mmol); 7Y-methyl-4-amimo-methyl-butyrate hydrochloride, 71 mg (0.42 mmol); trimethylacetyl chloride, 60 pL (0.50 mmol); DIEA part A, 109 pL (0.63 mmol), part B, 160 pL (0.92 mmol); THF 7 mL. Purification of the crude product by flash chromatography (silica gel, 1:4 ether-pet. ether) afforded 98 mg of 201 (68 %). *H NMR (200 MHz, CDC13, 5): 0.89 (s, 3 H), 0.93 (s, 6H), 1.36 (s, 9H), 1.83 - 1.89 (m, 2H), 2.25 (t, 2H, J = 7.6 Hz), 2.86 (s, 1H), 3.06 (s, 2H), 3.11 - 3.50 (m, 2H), 4.42 (dd, 1H, J = 9.8, 7.5 Hz), 5.27 (d, 1H, J = 9.8 Hz). 1 3 C NMR (50 MHz, CDC13, 8): 22.2, 26.3, 28.2, 35.4, 36.3, 47.2, 49.7, 51.5, 55.6, 79.3, 171.9, 173.3. Exact mass calculated for C i 7 H 3 3 N 2 0 5 (M+H)+: 345.23895; found: 345.23866. 233 5.4.2.7 A^Boc-AyV'-dimethyl-tryptophan (202). 202 (S)-A '^-Boc-tryptophan (600 mg, 1.9 mmol) and methyl iodide (618 uL, 9.85 mmol) were dissolved in dry THF (4 mL) and the solution was cooled to 0°C under an argon atmosphere. A 60% dispersion of sodium hydride in oil (236 mg, 5.9 mmol) was added cautiously with gentle stirring, the reaction mixture was warmed to at room temperature, and stirring was continued for 16 hours. Ethyl acetate (20 mL) was then added, followed the dropwise addition of water (15 mL). The solution was evaporated to dryness, and the oily residue partitioned between ether (20 mL) and water (50 mL). The ether layer was washed with saturated sodium hydrogen carbonate solution (20 mL), and the combined aqueous extracts acidified to pH 3 with citric acid. The product was extracted with ethyl acetate (3 x 25 mL), the organic extracts washed with water (2 x 25 mL), 5% aqueous sodium thiosulfate solution (2 x 25 mL), water (25 mL), dried over MgSC»4, and evaporated to yield 621 mg (94 %) of 202 as a pale yellow oil. 'H NMR (200 MHz, CDC13, 5): 1.23 (s, 6H), 1.47 (s, 3H), 2.76 (s, IH), 2.86 (s, 2H), 3.19 (m, IH), 3.49 (m, IH), 3.72 (s, 3H), 4.87 (dd, J= 5, 10 Hz: 0.6H), 5.11 (dd,/= 3, 5 Hz: 0.4H), 6.89 (s, 0.6H), 6.96 (s, 0.4H), 7.15 (m, 3H) 7.29 (s, IH), 7.64 (d, J = 6 Hz: IH). 1 3 C NMR (50 MHz, CDCI3, 8): 24.9,27.8,28.2,31.6,31.9,32.4,59.0,59.9,80.5, 102.6, 118.4, 118.9, 121.6, 126.9, 127.5, 136.9, 155.4, 155.3, 176.4. Exact mass calculated for Ci 8 H 2 5N 2 04 (M+H)+: 333.18143; found: 333.18135 [a]D 2 5 = -9.1°(c4.6, CHCI3). 234 5.4.2.8 iV'-Boc-/V'-desmethyl-hemiasterlin-ethyl ester (203). / Boc 203 Following the general procedure 3, ester 203 was prepared with the following quantities of reagents and solvents: A^-Boc-N,/^/J-trimethyl-S-tryptophan (186), 30 mg (0.08mmol); Af-Boc-S-terMeucine-N-methyl-4-amino-2,5-dimethylhex-2-enoic ethyl ester (183), 32 mg (0.08 mmol); PyBrop, 40 mg (0.08 mmol); DIEA, 45 pX (0.26 mmol); DMAP, 6.2 mg (0.05 mmol); CH2C12, 2 mL. Purification of the crude product by flash chromatography (silica gel, 2:3 ether-pet. ether) afforded 21 mg of 203 (42 %). 'H NMR (200 MHz, CDC13, 8): 0.41 (s, 9H), 0.85 (d, 3H, J = 6.6 Hz), 0.86 (d, 3H, J = 6.6 Hz), 1.14 (s, 3H), 1.26 (t, 3H, 7=7.1 Hz), 1.46 (s, 9H), 1.79-1.90 (m, IH), 1.84 (d, 3H, J= 1.5 Hz), 2.95 (s, 3H), 3.76 (s,3H), 4.16 (q, 1H, .7=7.1 Hz), 4.98 (t, 1H,7= 10.2 Hz), 5.12 (d, 1H,J = 8.3 Hz), 5.39 (d, 1H,J=8.3 Hz), 5.69 (d, IH, 7=8.1 Hz), 6.61 (dd, IH, 7=9.5, 1.5 Hz), 6.99 (s, IH), 7.11-7.31 (m, 3H), 8.13 (d, 1H,J=8.1 Hz). Exact mass calculated for C 3 6 H 5 7 N 4 O 6 (M+H)+: 641.42780; found: 641.42801. 5.4.2.9 A^'-Boc-lljll'-bis-desmethyl-hemiasterlin-ethyl ester (204). Following the general procedure 3, ester 204 was prepared with the following quantities of reagents and solvents: N'-Boc-N, N'-dimethyl-S-tryptophan (202), 44 mg (0.13 mmol); A'-Boc-iS'-^rr-leucine-7V-methyl-4-amino-2,5-dimethylhex-2-enoic ethyl ester (183), 50 mg (0.1.2 mmol); PyBrop, 62 mg (0.13 mmol); DIEA, 70 pL (0.40 mmol); DMAP, 10 mg (0.08 mmol); CH 2C1 2, 2 mL. Purification of the crude product by flash chromatography (silica gel, 1:1 ether-pet. ether) afforded 45 mg (54 %)of 204. 204 235 ' H N M R (200 MHz, CDCI3, 5): 0.67-0.89 (m, 6H), 0.91 (s, 9H), 1.13 (s, 1H), 1.15 (s, 1H), 1.21 (s, 4H), 1.22 (s, 1H), 1.28 (t, 3H, J=7.1 Hz), 1.40 (s, 2H), 1.82-1.92 (m, 1H), 1.86 (d, 3H, J = 1.2 Hz), 2.72 (s, 1H), 2.80 (s, 2H), 2.88 (s, 1H), 2.96 (s, 2H), 3.01-3.2 (m, 1H), 3.26-3.48 (m, 1H), 3.67 (s, 3H) ,4.17 (q, 2H, J= 7.3 Hz), 4.79 (t, 0.6H, J = 9 Hz), 4.86-5.09 (m, 1.5H), 6.60 (d, 1H, J = 9 Hz), 6.81 (bs, 1H), 7.02-7.25 (m, 3H), 7.55 (m, 1H); exact mass calculated for C 3 6 H 5 7 N 4 0 6 (M+H)+: 627.41216; found: 627.41384. 5.4.2.10 Ar-Boc-5-tiyptophan-5'-/er/-leucine-Ar-methyl-4-amino-2,5-dimethyIhex-2-enoic ethyl ester (205). Following the general procedure 3, ester 205 was prepared with the following quantities of reagents and solvents: iV'-Boc-tryptophan, 40 mg (0.13 mmol); TV-Boc-S-ferMeucine-A^methyl-4-amino-2,5-dimethylhex-2-enoic ethyl ester (183), 50 mg (0.12 mmol); PyBrop, 62 mg (0.13 mmol); DIEA, 70 pL (0.40 mmol); DMAP, 10 mg (0.08 mmol); CH 2C1 2, 2 mL. Purification of the crude product by flash chromatography (silica gel, 1:1 ether-pet. ether) afforded 43 mg (59%) of 205. *H NMR (200 MHz, CDCI3, 5): 0.76 (d, 3H, J = 6.6 Hz), 0.83 (d, 3H, J = 6.6 Hz), 0.84 (s, 9H), 1.27 (t, 3H, J= 7.1 Hz), 1.37 (s, 9H), 1.82-1.90 (m, 1H), 1.86 (d, 3H, J= 1.5 Hz), 2.93 (s, 3H), 3.17 (d, lH,J=6.8Hz), 4.17 (q, 2H,J=7.3Hz), 4.42 (q, 1H, .7=7.1 Hz), 7.75 (d, l H , y = 9.5 Hz), 5.00 (t, 1H, J = 9.5 Hz), 5.10 (d, 0.5H, J = 7.8 Hz), 6.62 (dd, 1H, J = 9.5, 1.5 Hz), 7.04-7.20 (m, 4H), 7.31 (dd, 1H,J= 6.7,1.2 Hz), 7.67 (d, 1H, J= 7.1 Hz), 8.3 (s, 1H); exact mass calculated for C 3 3H5iN 406 (M+H)+: 599.38086; found: 599.38061. H 205 236 5.4.2.11 N -Boc-21-desmethyl-hemiasterlin-ethyl ester (206). OEt 206 Following the general procedure 3, ester 206 was prepared with the following quantities of reagents and solvents: A^-Boc-A7,A '^/?,/?-tetramethyl-5'-tryptophan (170), 5 mg (0.014 mmol); 7Y-Boc-5'-valine-Ar-methyl-4-amino-2,5-dimethylhex-2-enoic ethyl ester (199), 6.8 mg (0.016 mmol); PyBrop, 7.1 mg (0.014 mmol); DIEA, 7.2 pL (0.041 mmol); DMAP, 1 mg (0.008 mmol); C H 2 C I 2 , 1 mL. Purification of the crude product by flash chromatography (silica gel, 3:7 ether-pet. ether) afforded 4.4 mg (51 %) of 206. *H NMR (400 MHz, CDCI3,5): 0.22 (d, 1H, J = 6.7 Hz), 0.30 (d, 0.5H, J = 6.7 Hz), 0.54 (d, 1H, J = 6.7 Hz), 0.59 (d, 0.5H, J = 6.7 Hz), 0.69- 0.94 (m, 9H), 1.23 (s, 9H), 1.27 (t, 3H, J = 7.3 Hz), 1.48 (s, 2H), 1.50 (s, 2H), 1.57 (s, 1H), 1.60 (s, 1H), 1.82 (s, 2H), 1.87 (d, IH,J= 1.5 Hz), 1.95-2.08 (m, 1H), 2.81 (s, 2H), 2.83 (s, 1H), 2.96 (s, 2H), 2.97 (s, 1H), 3.45 (q, 0.8H, J =1 Hz), 3.74 (s, 3H), 4.13-4.21 (m, 1.5H), 4.32 (t, 0.6H, J = 7 Hz), 4.41 (t, 0.4H, J = 1 Hz), 4.76 (dd, 0.6H, J= 10, 6 Hz), 4.93-5.00 (m, 1.5H), 5.53 (bs, 0.4H), 5.83 (s, 0.6H), 6.12 (d, 0.5H, J= 7.9 Hz), 6.57-6.63 (m, 1H), 7.02 (s, 0.4H), 7.05 (s, 0.6H), 7.14-7.29 (m, 3H), 7.91 (d, 0.4H, J= 8.2 Hz), 8.19 (d, 0.6H, J= 8.2 Hz); exact mass calculated for C 36H57N 40 6 (M+H)+: 641.42781; found: 641.42843. 5.4.2.12 /Y'-Boc^l^l'-bis-desmethyl-hemiasterlin-ethyl ester (207). Following the general procedure 3, ester 207 was prepared with the following quantities of reagents and solvents: 7Y'-Boc-A/,A '^/i)/i-tetramethyl-5'-tryptophan (170), 5 mg (0.014 mmol); 7Y-Boc-S'-buterine-7vr-methyl-4-amino-2,5-dimethylhex-2-enoic ethyl ester (200), 6.2 mg (0.016 mmol); PyBrop, 7.1 mg (0.014 mmol); DIEA, 7.1 pL (0.41 mmol); DMAP, 1 mg (0.008 mmol); 207 237 CH2CI2, 1 mL. Purification of the crude product by flash chromatography (silica gel, 3:7 ether-pet, ether) afforded 4.0 mg (46 %) of 207. 'H NMR (400 MHz, CDC13, 5): 0.41 (t, 2H, J = 7.3 Hz), 0.47 (t, 1H, J = 7.3 Hz), 0.77 (d, 3H, J = 6.5 Hz), 0.86 (d, 3H, J = 6.7 Hz), 1.18 (t, 1H, J = 7 Hz), 1.27 (t, 2H, J = 7 Hz), 1.47 (s, 9H), 1.51 (s, 1H), 1.59 (s, 4H), 1.61 (s, 1H), 1.79-1.85 (m, 1H), 1.80 (s, 3H), 2.74 (s, 2H), 2.77 (s, 1H), 2.97 (s, 2H), 2.99 (s, 1H), 3.45 (q, 0.7H, J= 7 Hz), 3.74 (s, 3H), 4.16 (q, 1.3H, J= 7 Hz), 4.45 (q, 0.6H, J = 6.5 Hz), 4.52 (q, 0.4H, J = 6.5 Hz), 4.91 (t, 1H, J = 9.3 Hz), 5.43 (bs, 0.6H), 5.73 (s, 0.6H), 6.10 (d, 0.4H, J = 7.6 Hz), 6.16 (d, 0.6H, J = 8.8 Hz), 6.58 (d, 1H, J = 7.6 Hz), 6.99 (s, 0.4H), 7.01 (s, 0.6H), 7.12-7.28 (m, 3H), 7.89 (d, 0.4H, J = 8.2 Hz), 8.14 (d, 0.6H, J = 7.6 Hz); exact mass calculated for C s s H ^ O e (M+H)+: 627.41216; found: 627.41384 5.4.2.13 Ar-Boc-5-7V,/Y',/J,/3-tertamethyl-tiyptophan-5-ter^leucine-N-methyl-4-amino-methyl-butyrate (208). 208 Following the general procedure 3, ester 208 was prepared with the following quantities of reagents and solvents: A '^-Boc-7ViA '^y5)/i-tetramethyl-5'-tryptophan (170), 10 mg (0.028 mmol); A/-Boc-5'-fer?-leucine-7Y-methyl-4-amino-methylbutyrate (201), 12.5 mg (0.036 mmol); PyBrop, 13 mg (0.028 mmol); DIEA, 14.5 pL (0.083 mmol); DMAP, 2 mg (0.016 mmol); CH 2C1 2, 1 mL. Purification of the crude product by flash chromatography (silica gel, 2:3 ether-pet. ether) afforded 6.1 mg (37 %) of 208. 'H NMR (400 MHz, CDCI3, 5): 0.49 (d, 3H, J = 6.5 Hz), 0.54 (d, 3H, J = 6.5 Hz), 0.82-0.85 (m, 1H), 1.43 (s, 1H), 1.44 (s, 1H), 1.49 (s, 9H), 1.51 (s, 1H), 1.54 (s, 1H), 1.57 (s, 1H), 1.63 (s, 1H), 1.76-1.78 (m, 1H), 2.24-2.30 (m, 1H), 3.01 (s, 2H), 3.07 (s, 1H), 3.22-3.32 (m, 1H), 3.33-3.42 (m, 1H), 3.65 (s, 3H), 3.72 (s, 1H), 3.73 (s, 2H), 4.39 (d, 0.5H, J = 8.8 Hz), 4.48 (d, 0.2H, J = 9.1 Hz), 4.55 (d, 0.2H, J = 9.7 Hz), 4.60 (d, 0.1H, J = 9.5 Hz), 5.52 (s, 0.1H), 5.57 (bs, 0.2H), 5.82 (s, 0.2H), 5.88 (s, 0.5H), 5.99 (d, 0.3H, J = 8.8 Hz), 6.06 (d, 0.5H, J = 8.5 Hz), 6.22 (d, 0.2H, J= 8.2 Hz), 6.84-7.08 (m, 1.3H), 7.12-7.29 (m, 3H), 7.91-7.94 (m. 0.3H), 8.15-8.20 (m, 238 0.3H), 8.23 (d, 0.4H, J = 7.3 Hz); exact mass calculated for C 3 2H5iN 4 0 6 (M+H)+: 587.38086; found: 587.38031. 5.4.2.14 /V'-desmethyl-hemiasterlin (209). Following the general procedure 4, N'-desmethyl-hemiasterlin (209) was prepared with the following quantities of reagents and solvents: A '^-Boc-A '^-desmethyl-hemiasterlin ethyl ester (203), 10 mg (0.015 mmol); lithium hydroxide (1.0 M) 122 pL (0.12 mmol); 1:1 THF/CH2C12 (1 mL). Purification of the crude product by reversed phase HPLC (1:1 0.05% TFA/H 20:MeOH) afforded 5.3 mg (69 %) of 209. 'H NMR (400 MHz, MeOD, 5): 0.85 (s, 9H), 0.90 (d, 3H, J = 6.4 Hz), 0.91 (d, 3H, J = 6.4 Hz), 1.46 (s, 3H), 1.54 (s, 3H), 1.89 (d, 3H, J= 1.3 Hz), 1.99-2.06 (m, IH), 3.01 (s, 3H), 3.80 (s, 3H), 4.64 (s, IH), 4.72 (s, IH), 5.03 (t, IH, J = 9.7 Hz), 6.76 (dd, IH, J = 9.7, 1.3 Hz), 7.13 (t, IH, J = 8.1 Hz), 7.14 (s, IH), 7.24 (t, 1H, .7=8.1 Hz), 7.42 (d, 1H, .7=8.1 Hz), 7.99 (d, 1H,J=8.1 Hz). 1 3 C NMR (100 MHz, MeOD, 5): 14.1,19.8,19.9, 24.8,24.9,26.6, 30.8, 32.0, 32.9, 35.6, 38.7, 57.5, 58.7, 60.4, 111.3, 118.0, 120.4, 121.0, 123.2, 125.9, 129.5, 133.4, 139.8, 168.4, 170.9, 172.4. exact mass calculated for C29H45N4O4 (M+H)+: 513.34408; found:- 513.34431. 209 5.4.2.15 ll,ll'-bis-desmethyl-hemiasterlin (210). 239 210 Following the general procedure 4, 11,1 l'-bis-desmethyl-hemiasterlin (210) was prepared with the following quantities of reagents and solvents: W-Boc-11,1 r-bis-desmethyl-hemiasterlin ethyl ester (204), 10 mg (0.015 mmol); lithium hydroxide (1.0 M) 122 uL (0.12 mmol); 1:1 THF/CH2C12 (1 mL). Purification of the crude product by reversed phase HPLC (1:1 0.05% TFA/H 20:MeOH) afforded 6 mg (80 %) of 210. *H NMR (400 MHz, MeOD, 5): 0.77 (d, IH, J = 6.9 Hz), 0.86 (d, IH, J = 6.9 Hz), 1.00 (s, 9H), 1.87-1.96 (m, IH), 0.88 (s, 3H), 2.59 (s, 3H), 3.05 (s, 3H), 3.15 (dd, IH, J = 15.3, 8 Hz), 3.34 (dd, IH, J = 15.3,6.1 Hz), 3.78 (s, 3H), 4.26 (dd, IH, J = 6.5, 6.1 Hz), 4.85 (s, IH), 5.03 (t, IH, J = 10.3 Hz), 6.73 (dd, IH, J = 9.5, 1.2 Hz), 7.10 (t, IH, J = 8 Hz), 7.13 (s, IH), 7.20 (t, IH, J = 8 Hz), 7.36 (d, IH, J = 8 Hz), 7.65 (d, IH, 7=8 Hz); exact mass calculated for C28H43N4O4 (M+H)+: 499.32843; found: 499.32802. 5.4.2.16 S-tryptophan-5,-tert-leucine-^-methyI-4-amino-2,5-dimethylhex-2-enoic acid (211). Following the general procedure 4, S-tryptophan-5'-terr-leucine- A^-methyl-4-amino-2,5-dimethylhex-2-enoic acid (211) was prepared with the following quantities of reagents and solvents: JV'-Boc- S-tryptophan-1S'-?err-leucine-A^-methyl-4-amino-2,5-dimethylhex-2-enoic ethyl ester (205), 10 mg (0.017 mmol); lithium hydroxide (1.0 M)136 pL (0.14 mmol); 1:1 THF/CH2CI2 (1 mL). Purification of the crude product by reversed phase HPLC (1:1 0.05% TFA/H20:MeOH) afforded 5.3 mg (66 %) of 211. 240 ! H NMR (400 MHz, CDC13, 8): 0.84 (d, 3H, J = 6.5 Hz), 0.88 (d, 3H, J = 6.5 Hz), 1.01 (s, 9H), 1.89 (d, 3H,J= 1.5 Hz), 1.93-2.00 (m, 1H), 3.02-3.11 (m, 1H), 3.06 (s, 3H), 3.36 (dd, I R , J = 14.9, 5 Hz), 4.32 (dd, 1H, J = 9.5, 5.3 Hz), 4.85 (s, 1H), 5.05 (t, 3H, J = 9.9 Hz), 6.75 (dd, 3H, J = 9.9, 1.5 Hz), 7.07 (t, 1H, J= 8 Hz), 7.14 (t, 1H, J= 8 Hz), 7.22 (s, 1H), 7.38 (d, 1H, J = 8 Hz), 7.73 (t, 1H, J = 8 Hz); exact mass calculated for C26H39N4O4 (M+H)+: 471.29713; found: 471.29823. 5.4.2.17 S-A^A^/Jj/i-tertamethyltryptoph^ acid (212). 212 Following the general procedure 4, S-A/i7Y')/i,/3-tertamethyltryptophan-5'-/er^-leucine-7Y-methyl-4-amino-butyric acid (212) was prepared with the following quantities of reagents and solvents: N-Boc- S-A^7vr')/i)/i-tertamethyl-tryptophan-iS'-rerr-leucine-Air-methyl-4-amino-butyric methyl ester (208), 6 mg (0.010 mmol); lithium hydroxide (1.0 M) 82 pL (0.08 mmol); 1:1 THF/CH2C12 (1 mL). Purification of the crude product by reversed phase HPLC (2:3 0.05% TFA/H20:MeOH) afforded 2.2 mg (45 %) of 183. 'H NMR (500 MHz, MeOD, 8): 1.04 (s, 9H), 1.46-1.52 (m, 2H), 1.49 (s, 1H), 1.57 (s, 3H), 1.80-1.83 (m, 2H), 2.27-0.35 (m, 2H), 2.45 (s, 3H), 3.19 (s, 3H), 3.79 s, 3H), 4.45 s, 1H), 4.89 (s, 1H), 7.12-7.16 (m, 2H), 7.23 (t, J = 8 Hz, 1H), 7.41 (d,J=8Hz, 1H), 8.01 (d,J=8Hz, 1H); exact mass calculated for C2 6H4iN404 (M+H)+: 473.31278; found: 473.31347. 5.4.2.18 21-desmethyl-hemiasterlin (213). 213 241 Following the general procedure 4, 21-desmethyl-hemiasterlin (213) was prepared with the following quantities of reagents and solvents: A '^-Boc-21-desmethyl-hemiasterlin ethyl ester (206), 4.4 mg (0.007 mmol);); lithium hydroxide (1.0 M) 56 uL (0.056 mmol); 1:1 THF/CH2C12 (1 mL). Purification of the crude product by reversed-phase HPLC (1:1 0.05% TFA/H 20:MeOH) afforded 2.5 mg (69 %) of 213. *H NMR (500 MHz, MeOD, 8): 0.85-0.92 (m, 6H), 0.96 (d, J = 6.4 Hz, 3H), 1.04 (d, J = 6.4 Hz, 3H), 1.44 (s, 3H), 1.45 (s, 3H), 1.58 (s, 3H), 1.91-1.96 (m,lH), 2.01-2.08 (m, IH), 2.40 (s, 3H), 3.13 (s, 3H), 3.08 (s, 3H), 4.36 (s, IH), 4.77 (m, IH), 4.99-5.02 (m, IH), 6.71-6.78 (m, IH), 7.08-7.14 (m, IH), 7.24 (t, J = 8 Hz, IH), 7.41 (d, J = 8 Hz, IH), 8.03 (d, J = 8 Hz, IH); exact mass calculated for C 2 9H4 5N 404 (M+H)+: 513.34408; found: 513.34491. 5.4.2.19 21,21'-bis-desmethyl-hemiasterlin (214) Following the general procedure 4, 21,21'-bis-desmethyl-hemiasterlin (214) was prepared with the following quantities of reagents and solvents: A/-Boc-21,21'-bis-desmethyl-hemiasterlin ethyl ester (207), 4 mg (0.006 mmol); lithium hydroxide (1.0 M) 48 uL (0.05 mmol); 1:1 THF/CH2C12 (1 mL). Purification of the crude product by reversed phase HPLC (1:1 0.05% TFA/H 20:MeOH) afforded 2.3 mg (76 %) of 214. : H NMR (500 MHz, MeOD, 8): 0.91 (d, 7 = 6.4 Hz, 6 Hz), 1.02 (t, J = 7.5 Hz, 3), 1.45 (s, 3H), 1.64 (s, 3H), 1.72-1.78 (m, IH), 1.79-1.91 (br s, 4H), 1.93-2.05 (m, IH), 2.40 (s, 3H), 3.08 (s, 3H), 3.80 (s, 3H), 4.35 (s, IH), 4.90-5.02 (m, 2H), 6.72-6.78 (m, IH), 7.11-7.15 (m, 2H), 7.24 (d, J = 8 Hz, IH), 7.42 (t, J = 8 Hz, IH), 8.00 (d, J = 8 Hz, IH). exact mass calculated for C28H43N404 (M+H)+: 499.32843; found: 499.32983 / 214 5.5 ENDNOTES: C H A P T E R S : SYNTHESIS 242 1) Coleman, J. E.; de Silva, E. D.; Kong, F.; Andersen, R. J.; Allen, T. M . Tetrahedron 1995, 51, 10653. 2) Anderson, H., J.; Coleman, J. E.; Andersen, R., J.; Roberge, M . Cancer Chem. Pharm. 1996, 39, 223. 3) Le Blanc, M. , Cymbastela sp. very rare. 4) Andersen, R. J. personal communication 1996. 5) Andersen, R. J.; Coleman, J. C ; Piers, E.; Wallace, D. J. Tetrahedron Lett. 1997, 38, 317. 6) Faulkner, D. J. Nat. Prod. Rep. 1997,14, 259. 7) Chan, W. R.; Tinto, W. F.; Manchand, P. S.; Todaro, L. J. J. Org. Chem. 1987, 52. 8) de Silva, E. D.; Andersen, R. J.; Allen, T. M . Tetrahedron Letters 1990, 31, 489. 9) Pettit, R.; Herald, D. L.; Singh, S. B.; Thornton, T. J.; Mullaney, J. T. J. Am. Chem. Soc. 1991,113, 6692. 10 11 12 13 14 15 16 17 18 19 Wenger, R. M . Angew. Chem., Int. Ed. Engl. 1985, 24, 77. Coste, J.; Frerot, E.; Jouin, P. J. Org. Chem. 1994, 59, 2437. Ryakhovskii, V. V.; Agafonov, S. V.; Kosyrev, Y. M. Russ. Chem. Rev. 1991, 60, 924. Jackson, A. G.; Kenner, G. W.; Moore, G. A.; Ramage, R.; Thorpe, W. D. Tetrahedron Lett. 1976,77,3627. Diago-Meseguer, J.; Palomo-Coll, A. L.; Fernandez-Lizarbe, J. R.; Zugaza-Bilbao, A. Synthesis 1980, 547. Castro, B.; Dormoy, J. Tetrahedron Lett. 1973,14, 3243. Coste, J.; Defour, M. ; Pantaloni, A.; Castro, B. Tetrahedron Lett. 1990, 31, 669. Castro, B.; Dormoy, J.; Evin, G.; Selve, C. Tetrahedron Lett. 1975,16, 1219. Coste, J.; Frerot, E.; Jouin, P. Tetrahedron Lett. 1991, 32, 1967. Frerot, E.; Coste, J.; Poncet, J.; Jouin, P. Tetrahedron Lett. 1992, 33, 2815. 20) Hagihara, M. ; Anthony, A. J.; Stout, T. J.; Clardy, J.; Schreiber, S. L. J. Am. Chem. Soc. 1992,114, 6568. 21) Nahm, S.; Weinreb, S. M . Tetrahedron Lett. 1981, 22, 3815. 22) Fehrentz, J.; Castro, B. Synthesis 1983, 676. 23) Coste, J.; Le-Nguyen, D.; Castro, B. Tetrahedron Lett. 1990, 31, 205. 24) Frerot, E.; Coste, J.; Pantaloni, A.; Dufour, M. ; Jouin, P. Tetrahedron 1991, 47, 259. 25) Bodanszky, M. ; Bodanszky, A. The Practice of Peptide Synthesis.; Springer-Verlag: Berlin, 1984, pp 151-187. 6 A P P E N D I X 1: S E L E C T E D N M R S P E C T R A 244 DITERPENOIDS Botanone B (56) PP 245 -248 Botanone C (57) PP 249 -252 Botanicol (58) PP 253 -256 Cadlinoglycine (59) PP 257 -260 Cadlinolide C methyl acetal (60) PP 261 -264 GEODIAMOLIDES Geodiamolide J (91) pp 265-268 Geodiamolide M (94) pp 269-271 PEPTDDES Hemiasterlin A (98) pp 272 -275 Hemiasterlin B (99) pp 276 -279 Criamide A (100) pp 280 -283 Criamide B (101) p284 246 Figure A2. COSY NMR spectrum of botanone B (56). Recorded in CDC13 at 500 MHz. 247 Figure A3. HMQC NMR spectrum of botanone B (56). Recorded in CDC13 at 500 MHz. 248 Figure A4. HMBC NMR spectrum of botanone B (56). Recorded in CDC13 at 500 MHz. 249 Figure A 6 . COSY NMR spectrum of botanone C ( 5 7 ) . Recorded in C6D6at 500 MHz. 251 100 h 120 140 (ppm) Figure Al. HMQC NMR spectrum of botanone C (57). Recorded in C 6D 6 at 500 MHz. 252 I I l l l j l l l l l l l l l | l l l l l l l I l | l l l l l l l l l [ I l l l l l l l l [ (ppm) 6 5 4 3 T I | 1 I I I I I I I I | I I I I I I 2 1 Figure A 8 . HMBC NMR spectrum of botanone C (57). Recorded in C 6D 6 at 500 MHz. 253 254 16 17 58 nzu 1 m OH J 1 I UhhJj 'IV1 / \ 7 (ppm) 6 Figure A10. COSY NMR spectrum of botanicol (58). Recorded in C 6D 6 at 500 MHz. 255 16 17 OAc Mel 7 \ s 20 40 60 80 100 120 140 I I I | I I I I I I I I I | I I I I I I I I I | I I I I I I I I I | 1 I I ' T T I 'I "T 'I "I T I" • r r - r - r - r - r - 1 ] | | I | | | | | l | | | | | l [ ( p p m ) 6 Figure A l l . HMQC NMR spectrum of botanicol (58). Recorded in C 6D 6 at 500 MHz. 256 16 17 2 0 0' • * • 0 o • o 0 ••6 • °o •20 '40 0 h60 « 0 80 100 D 6 a o 120 o 0 • O hl40 fc 160 1**1 1 1 f''' J ' 1 r* 1"I"'T " I ' I" 7'"fT—yT | ; | | | T I T | I I I 1 I I I I I | I I I I 1 I I' I I | I I I " f ***IT I I' I '[ I 1 I I I I 1 I t j I I I I I I (ppm) Figure A12. HMBC NMR spectrum of botanicol (58). Recorded in C 6D 6 at 500 MHz. 257 258 259 16 17 0 s o o O 0 Q - o c ? 0 O 0 o o 120 30 140 50 t60 70 F80 90 (ppm) 5 Figure A15. HMQC NMR spectrum of cadlinoglycine (59). Recorded in DMSO-4, at 500 MHz. 260 2 2 OH 16 17 HI4 0 o o ° <b- . • i °<*00 0 0 O o 0 o o « O 0 9 o G a 9 O -i—i—i—i—i—r -20 40 60 80 100 120 140 160 (ppm) 5 1 , 1 1 180 Figure A16. HMBC NMR spectrum of cadlinoglycine (59). Recorded in DMSO-d6 at 500 MHz. 261 262 Figure A18. COSY NMR spectrum of cadlinolide C methyl acetal (60). Recorded in CDC13 at 500 MHz. 263 H7 u M e l 6 M e l 7 6 20 40 60 80 100 (ppm) 4.5 4.0 3.5 3.0 2.5 2.0 1.5 1.0 Figure A19. HMQC NMR spectrum of cadlinolide C methyl acetal (60). Recorded in CDC1, at 500 MHz. / 264 OMs Mel6 Mel7 0 0 0 1 0 ° o a o D • O ' O 0 0 Q O 0 0 i i i | i i i i | I i I I | I I I I j i i i i | i i i i | i (ppm) 4.5 4.0 3.5 3.0 2.5 i i i | i i i i 1 i i i i 1 i i i i 2.0 1.5 1.0 0.5 6 a> 20 40 60 80 100 120 140 160 180 Figure A20. H M B C N M R spectrum of cadlinolide C methyl acetal (60). Recorded in CDC1, at 500 M H z . 265 266 OH 91 N-H10 H21 N-H14 JL_ HIS' H6 H 3 m [H7 Me25 y M c 2 2 / Me28 p p m 1 PPm Figure A22. COSY NMR spectrum of geodiamolide J (91). Recorded in CDC13 at 500 MHz. 267 I I I I 1 I | I I I | I I I I I I I I I [ i ' T T T n I I I | I I I I I I T - r - l — p i T I I I I I I I | I Tl [ I I I I 7 6 5 4 3 2 1 P P m Figure A23. HMQC NMR spectrum of geodiamolide J (91). Recorded in CDC13 at 500 MHz. 268 Figure A24. HMBC NMR spectrum of geodiamolide J (91). Recorded in CDC13 at 500 MHz. 269 H17 H20 | J J I h N-H14 N-H10 1 3 _ H5 H14 H 1 2 \ / H10 H15 JUJ H,0 H15' H 7 H 6 Me23 Me26 | Me25 Me22 U O p o O o 0 • £ 3 0 0 0 °& o o o e e 0 • • o o 0 o o o o a OO o • OD o • o ppm : 1 =" 2 1 PPm Figure A25. COSY N M R spectrum of geodiamolide M (94). Recorded in CDC1 3 at 400 MHz. 270 19 21' HO 20 27 13T O o 25 24 23" 14 28 OH H E 2 2 94 I H 2 0 I i IN-H14 N-H H n \ I H10 J U J J H,0 H3' HIS' H2 H 6 U u ' m m Mc23 Me26 Me25 Me22 I I Mc24 HI ppm 100 120 140 1 PPm Figure A26. H M Q C N M R spectrum of geodiamolide M (94). Recorded in CDCL. at 500 MHz. 271 HO 20 27 14 2 8 ^ N " OH H 1 2 2 94 H20 | J I N - H 1 4 N - H 1 0 H28 H I 5 H I 5' H2 Mc23 Me26 Mc25 M e 2 2 , t 0 0 0 1 g • » 0 t o 0 o ' 0 A O o 0 0 0 c 1 0 0 0 • * 0 0 0 0 • 0 0 0 0 0 0 * 0 I 1 I I 1 * I I I I I I I I'-l" I 111 7 ppm "20 "40 60 "80 '100 "120 "140 "160 "180 1 PPm Figure A27. HMBC NMR spectrum of geodiamolide M (94). Recorded in CDC13 at 500 MHz. 272 273 Figure A29. COSY NMR spectrum of hemiasterlin A (98). Recorded in DMSO-d6 at 500 MHz. 274 2 9 O H H2 / 0 H19 (ppm) 8.0 6.0 4.0 M e l 4 M e l 5 Me22 Me23 Me24 Me32 Me33 j 2.0 0 6 a, a 20 40 60 80 100 120 140 Figure A30. HMQC NMR spectrum of hemiasterlin A (98). Recorded in DMSO-d, at 500 MHz. 275 Figure A31. HMBC NMR spectrum of hemiasterlin A (98). Recorded in DMSO-d, at 500 MHz. I i 276 277 (ppm) 8.0 6.0 4.0 2.0 Figure A 3 3 . COSY NMR spectrum of hemiasterlin B ( 9 9 ) . Recorded in DMSO-d, at 500 MHz. ( 2 7 8 Figure A34. HMQC NMR spectrum of hemiasterlin B (99). Recorded in DMSO-4 at 500 MHz. 279 (ppm) 8.0 6.0 4.0 2.0 Figure A 3 5 . HMQC NMR spectrum of hemiasterlin B (99). Recorded in DMSO-tf, at 500 MHz. 281 i 1 1 1 1 r 1 1 1 1 1 1 1 1 1 1 1 8.0 (ppm) 6.0 4.0 2.0 Figure A37. COSY NMR spectrum of criamide A (100). Recorded in DMSO-tf, at 500 MHz. 282 H 43 4 i N . 4 2 / N H , M e 2 2 M e 2 3 M e 2 4 ( p p m ) 8.0 Figure A38. HMQC NMR spectrum of criamide A (100). Recorded in DMSO-4 at 500 MHz. 283 (ppm) 10.0 8.0 6.0 4.0 2.0 Figure A39. HMBC NMR spectrum of criamide A (100). Recorded in DMSO-4 at 500 MHz. 284 ure A40. COSY NMR spectrum of criamide B (101). Recorded in DMSO-tf, at 500 MHz. 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.831.1-0061561/manifest

Comment

Related Items