Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Novel G2 cell cycle checkpoint inhibitors and antimitotic agents isolated through two new HTS bioassays Xu, Lin 2000

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

Item Metadata

Download

Media
831-ubc_2001-612074.pdf [ 11.05MB ]
Metadata
JSON: 831-1.0061302.json
JSON-LD: 831-1.0061302-ld.json
RDF/XML (Pretty): 831-1.0061302-rdf.xml
RDF/JSON: 831-1.0061302-rdf.json
Turtle: 831-1.0061302-turtle.txt
N-Triples: 831-1.0061302-rdf-ntriples.txt
Original Record: 831-1.0061302-source.json
Full Text
831-1.0061302-fulltext.txt
Citation
831-1.0061302.ris

Full Text

N O V E L G2 C E L L C Y C L E C H E C K P O I N T INHIBITORS AND A N T I M I T O T I C A G E N T S I S O L A T E D T H R O U G H T W O N E W HTS BIOASSAYS by Lin Xu B.Sc, Nankai University, 1992 R A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in THE 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 T H E UNIVERSITY OF BRITISH COLUMBIA December 2000 © Lin Xu, 2000 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 The University of British Columbia Vancouver, Canada Date Otf, 7^0-0 DE-6 (2/88) II A B S T R A C T The search for new pharmacologically active agents by screening natural sources such as microbial fermentations and plant extracts has led to the discovery of many clinically useful drugs that play a major role in the treatment of human diseases. Analogous to the investigation of terrestrial products, some recent studies of marine natural products have focused on their potential applications, particularly the treatment of human diseases, such as cancer. The goal of my thesis research was to discover structurally novel G2 cell cycle checkpoint inhibitors and antimitotic agents in natural product extracts. Active crude extracts were identified via high throughput bioassays for G2 checkpoint inhibitors and antimitotic agents. The assays were developed by our collaborator Michel Roberge and the details of the assays are introduced in the thesis. Several G2 checkpoint inhibitors were isolated and identified from extracts of marine organisms and plants. A new inhibitor pachyclavularolide F (18) was isolated from the soft coral Pachyclavularia violacea. Several inactive analogs including pachyclavulariolides A (43), B (44), C (45), D (46), E (47), and G (48) were also isolated from P. violacea. The previous known inhibitor staurosporine (10) and its new semisynthetic oxazolidone derivative (17) were discovered from a marine microorganism. The known metabolites aaptamine (12), isolated from the sponge Aaptos auberitoides and the diterpene lactones (19, 54, 55) from the tropical bush Parinari curatellifolia were found to show the G2 checkpoint inhibition. Using Dr. Roberge's high throughput antimitotic assay, the paclitaxel derivatives 10-deacetyltaxuyunnanine A (71) and 7-(p-xylosyl)-10-deacetyltaxol C (72) were isolated from the Southern American plant Ilex macrophylla. This work resulted in the discover of a new natural source for paclitaxel derivatives, and it also validated the accuracy and efficiency of the new antimitotic assay. My research on new metabolites from a marine sponge Xestospongia ingens guided by a traditional cytotoxicity bioassay led to several new ingenamine alkaloid derivatives, including ingenamines H (56), I (57), J (58), K (59), and L (60). iv T A B L E O F C O N T E N T S Abstract ii Table of contents iv List of tables ix List of figures.... x List of pictures -. xvi List of abbreviations xvii Acknowledgements xxi Chapter I: Introduction 1 1.1 Marine natural products as drug lead candidates 3 1.2 Cell cycle checkpoint inhibitor bioassay 10 1.2.1 Cell cycle 10 1.2.2 Cell cycle checkpoint 11 1.2.3 GI checkpoint 12 1.2.4 G2 checkpoint 13 1.2.5 G2 checkpoint inhibitors in cancer therapy 17 1.2.6 Novel G2 checkpoint inhibitors discovered by the assay 22 1.2.7 G2 checkpoint inhibitors as tools in cell cycle research 23 1.3 Introduction to the antimitotic assay 25 1.3.1 Microtubules and tublin 25 1.3.2 Cell replication and the cell cycle 26 1.3.3 Antimitotic agents 28 1.3.4 A cell based antimitotic bioassay 30 1.4 Introduction to chemical genetics 32 1.5 Preview of research results in thesis 36 References 37 Chapter II: Cembranolide diterpenoids 43 2.1 Cembranolide diterpenoids from the soft coral Pachyclavularia violacea 43 2.2 A brief review of cembranolide diterpenoids from soft corals 43 2.3 Results and discussion 51 2.3.1 Isolation of cembranolide diterpenes from Pachylclavularia violacea 51 2.3.2 Structure elucidation of cembranolides from Pachylclavularia violacea 52 2.3.2.1 Pachyclavulariolide A 55 2.3.2.2 Pachyclavulariolide B 67 2.3.2.3 Pachyclavulariolide C 77 2.3.2.4 Pachyclavulariolide D 87 2.3.2.5 Pachyclavulariolide E 95 2.3.2.6 Pachyclavulariolide F 104 2.3.2.7 Pachyclavulariolide G 116 2.3.3 Discussion and conclusions 125 2.3.4 Experimental 128 References 131 VI Chapter III: G2 cell cycle checkpoint inhibitors 133 3.1 Introduction 133 3.2 Marine Actinomycete Clin 116 133 3.2.1 Isolation of active metabolites from bacterial extract clin 1116 134 3.2.2 The origin of two semisynthetic analogues 137 3.2.3 Structure elucidation of two semisynthetic analogues of staurosporine 139 3.2.3.1 Oxazolidone derivative 17 139 3.2.3.2 Carbamate derivative 49 149 3.2.4 Biological activity 160 3.2.5 Experimental 161 3.3 The Indonesian sponge Aaptos suberitoides 164 3.3.1 Isolation of active metabolites from sponge extract RJA-96-57 164 3.3.2 Structure elucidation of aaptamine and its derivatives 168 3.3.2.1 Aaptamine 168 3.3.2.2 Isoaaptamine 170 3.3.2.3 Demethyl(oxy)aaptamine .172 3.3.2.4 Demethyl(oxy)aaptamine analog 52 174 3.3.3 Biological activity 181 3.3.4 Experimental 182 3.4 Diterpene lactones from the South African plant Parinari curate 11 ifo I ia 184 3.4.1 Isolation of active metabolites from plant Parinari curatellifolia 184 vii 3.4.2 Structure elucidation of the diterpene lactone derivatives 186 3.4.3 Biological activity 193 3.4.4 Experimental 198 References 200 Chapter IV: Ingenamine alkaloids 201 4.1 Introduction 201 4.2 Isolation of cytotoxic metabolites from Xestospongia ingens 204 4.3 Structure elucidation of ingenamine alkaloids from Xestospongia ingens 206 4.3.1 Ingenamine H 208 4.3.2 Ingenamine 1 218 4.3.3 Ingenamine J 226 4.3.4 Ingenamine K .232 4.3.5 Ingenamine L 240 4.4 Biogenic considerations 248 4.5 Biological activity 250 4.6 Experimental..' 252 References 254 Chapter V: Antimitotic agents from Southern American plant Ilex macrophylla 255 5.1 Isolation of antimitotic metabolites from plant Ilex macrophylla 255 5.2 Results 258 5.3 Antimitotic activity of paclitaxel analogs 263 5.4 Cytotoxicity of paclitaxel analogs 263 5.5 Structure activity relationship of paclitaxel analogs 266 5.6 Effect of combining antimitotic compounds and saponin metabolites 267 5.7 Significance of the results and future research 269 References 270 Chapter VI: Conclusion 271 IX LIST O F T A B L E S Table 2.1 NMR data for pachyclavulariolide A (43) recorded in benzene-t/^  at 500MHz ('H). Temp:328K 63 Table 2.2 NMR data for pachyclavulariolide B (44) recorded in benzene-^ at 500 MHz('H). Temp: 328K 72 Table 2.3 NMR data for pachyclavulariolide C (45) recorded in benzene-J6 at 500 MHz ('H). Temp: 328K 83 Table 2.4 NMR data for pachyclavulariolide D (46) recorded in benzene-^ at 500 MHz ( !H). Temp: 298K 92 Table 2.5 NMR data for pachyclavulariolide E (47) recorded in benzene-d6 at 500 MHz ('H). Temp: 293K 100 Table 2.6 NMR data for pachyclavulariolide F (18) recorded in benzene-t/g at 500 MHz('H) 113 Table 2.7 NMR data for pachyclavulariolide G (48) recorded in benzene-^ at 500 MHz('H) 124 Table 3.1 NMR data for oxazolidine derivative 17 recorded in DMSO-dg 148 Table 3.2 NMR data for carbamate derivative 49 recorded in DMSO-d 6 159 Table 3.3 NMR data for 52 recorded in DMSO-d 6 180 r Table 3.4 NMR data for 53 recorded in CDC1 3 at 500 MHz (*H) 192 Table 3.5 Interaction between the checkpoint inhibiting effects of lactone 55 and IGR (16), DBH (15), UCN-01 (11), and caffeine (8) 194 Table 4.1 NMR data for ingenamine H (56) recorded in benzene-d6 at 500 MHz 217 Table 4.2 NMR data for ingenamine I (57) recorded in benzene-d6 at 500 MHz 225 Table 4.3 NMR data for ingenamine J (58) recorded in methanol-d4 at 500 MHz 231 Table 4.4 NMR data for ingenamine K (59) recorded in benzene-d6 at 500 MHz 239 Table 4.5 NMR data for ingenamine L (60) recorded in methanol-d4 at 500 MHz 247 Table 4.6 Cytotoxic activity against murine leukemia P388 cells in vitro 250 X LIST O F FIGUTRES Figure 1.1: Phyletic distribution of marine natural products 1997 7 Figure 1.2: Cell cycle and DNA damage checkpoints 11 Figure 1.3: The GI checkpoint pathway 13 Figure 1.4: Current model of the mammalian G2 checkpoint pathway 15 Figure 1.5: Overview of G2 checkpoint pathway in fission yeast S.pombe 16 Figure 1.6: Rationale for use of G2 checkpoint inhibitors in cancer therapy 19 Figure 1.7: G2 checkpoint inhibitor assay 21 Figure 1.8: Microtubule structure 26 Figure 1.9: Stages of mitosis 27 Figure 1.10: Both small molecules and mutations have been used to explore protein function. This figure emphasizes the relationship of genetics and chemical genetics 34 Figure 1.11: Small molecule-based studies of the cell cycle and cell cycle checkpoints and the chemical structure of the natural product that have been used to gain new sight into cell cycle progression 35 Figure 2.1: Biosynthetic pathway of diterpenoids 43 Figure 2.2: Cyclization pathway of the geranylgeraniol 44 Figure 2.3: Isolation of pachyclavulariolide diterpenoids from P. violacea 54 Figure 2.4: Low resolution FAB mass spectrum of pachyclavulariolide A (43) 56 Figure 2.5: 'H NMR data (500 MHz) for 43 in CD 2 C1 2 at the different temperatures. 1 3 C NMR data (125 MHz) for 43 in CD 2 C1 2 at -20°C 57 Figure 2.6: *H NMR data (500 MHz) for 43 in C 6 D 6 at different temperatures 58 Figure 2.7: The fragment C2 to C7 with selected COSY correlations 59 Figure 2.8: The [2,2,0] cyclic core with selected HMBC and COSY correlations 60 xi Figure 2.9: The trisubstituted butenolide with selected HMBC correlations 61 Figure 2.10: Selected NOE correlations (left) and conformation models (right) 62 Figure 2.1 1: ! H / 1 3 C NMR spectrum of pachyclavulariolide A (43) in C 6 D 6 at 328 K. . . .64 Figure 2.12: 2D COSY spectrum of pachyclavulariolide A (43) in C 6 D 6 at 328 K 65 Figure 2.13: 2D H M Q C spectrum of pachyclavulariolide A (43) in C 6 D 6 at 328K 66 Figure 2.14: Low resolution FAB mass spectrum of pachyclavulariolide B (44) 68 Figure 2.15: 'H NMR data (500 MHz) for 44 in C 6 D 6 at the different temperature 69 Figure 2.16: The fragment of 44 with selected COSY and HMBC correlations 70 Figure 2.17: Selected HMBC correlation in pachyclavulariolide B (44) 71 Figure 2.18: l H / 1 3 C NMR spectrum of pachyclavulariolide B (44) in C 6 D 6 at 328 K. . . .73 Figure 2.19: X-ray structure of pachyclavulariolide B (44) 74 Figure 2.20: 2D COSY spectrum of pachyclavulariolide B (44) in C 6 D 6 at 328 K 75 Figure 2.21: 2D HMQC spectrum of pachyclavulariolide B (44) in C 6 D 6 at 328 K 76 Figure 2.22: Low resolution FAB mass spectrum of pachyclavulariolide C (45) 78 Figure 2.23: *H NMR data (500 MHz) for 45 in C 6 D 6 at the different temperatures 79 Figure 2.24: The fragment of 45 with selected COSY and HMBC correlations 80 Figure 2.25: Selected H M B C correlation in pachyclavulariolide C (45) 81 Figure 2.26: The fragmention was stabilized by resonance 82 Figure 2.27: Selected NOE correlations in pachyclavulariolide C (45) 82 Figure 2.28: ' H / I 3 C NMR spectrum of pachyclavulariolide C (45) in C 6 D 6 at 328 K 84 Figure 2.29: 2D COSY spectrum of pachyclavulariolide C (45) in C 6 D 6 at 328 K 85 Figure 2.30: 2D H M Q C spectrum of pachyclavulariolide C (45) in C 6 D 6 at 328 K 86 xii Figure 2.31: Low resolution FAB mass spectrum of pachyclavulariolide D (46) 88 Figure 2.32: The fragment of 46 with selected COSY and HMBC correlations 89 Figure 2.33: Selected HMBC correlation in pachyclavulariolide D (46) 90 Figure 2.34: 'H / 1 3 C NMR spectrum of pachyclavulariolide D (46) in C 6 D 6 at 298 K....93 Figure 2.35: 2D COSY spectrum of pachyclavulariolide D (46) in C 6 D 6 at 298 K 94 Figure 2.36: Low resolution FAB mass spectrum of pachyclavulariolide E (47) 96 Figure 2.37: The fragment of 47 with selected COSY and HMBC correlations 97 Figure 2.38: The fragment of 47 with selected COSY and HMBC correlations 98 Figure 2.39: The structure of 47 with selected NOE correlations 99 Figure 2.40: 'H/ l 3 C NMR spectrum of pachyclavulariolide E (47) in C 6 D 6 at 293 K... 101 Figure 2.41: 2D COSY spectrum of pachyclavulariolide E (47) in C 6 D 6 at 293 K 102 Figure 2.42: 2D H M Q C spectrum of pachyclavulariolide E (47) in C 5 D 6 at 293 K 103 Figure 2.43: Low resolution FAB mass spectrum of pachyclavulariolide F (18) 105 Figure 2.44: 'H/ l 3 C NMR spectrum of pachyclavulariolide F (18) in C 6 D 6 at 293 K... 106 Figure 2.45: l 3C/DEPT-135 NMR spectrum of pachyclavulariolide F (18) in C 6 D 6 at 293 K 107 Figure 2.46: The fragment of 18 with selected COSY and HMBC correlations 108 Figure 2.47: Selected NOE results of pachyclavulariolide F (18) in C 6 D 6 (400 MHz)..l 12 Figure 2.48: 2D COSY spectrum of pachyclavulariolide F (18) in C 6 D 6 at 293 K 114 Figure 2.49: 2D HMQC spectrum of pachyclavulariolide F (18) in C 6 D 6 at 293 K 115 Figure 2.50: Low resolution FAB and ESI mass spectrum of pachyclavulariolide G . . . . 117 Figure 2.51: 'H/ ' 3 C NMR spectrum of pachyclavulariolide G (48) in C 6 D 6 at 293 K....118 xiii Figure 2.52: The trisubstituted butenolide core with related HMBC correlations 120 Figure 2.53: The selected NOE correlations in pachyclavulariolide G (48) 121 Figure 2.54: 2D COSY spectrum of pachyclavulariolide G (48) in C 6 D 6 at 293 K 122 Figure 2.55: 2D HMQC spectrum of pachyclavulariolide G (48) in C 6 D 6 at 293 K 123 Figure 2.56: Biogenic proposal for pachyclavulariolide E (47), F (18), and G (48)...... 127 Figure 3.1: Isolation scheme of staurosporine from bacteria isolate clinl 116 136 Figure 3.2: Low resolution EI mass spectrum of oxazolidone derivative 17 140 Figure 3.3: Selected NOE (left) and HMBC (right) correlations 141 Figure 3.4: 'H NMR spectrum of oxazolidone derivative 17 (500MHz, DMSO-d 6) 143 Figure 3.5: Selected NOE results of oxazolidone derivative 17 in DMSO-d 6 144 Figure 3.6: l 3 C NMR spectrum of oxazolidone derivative 17 (100MHz, DMSO-d6)....145 Figure 3.7: 2D H M Q C spectrum of oxazolidone derivative 17 (DMSO-d6) 146 Figure 3.8: 2D COSY spectrum of oxazolidone derivative 17 (500MHz, DMSO-d 6 ). . . 147 Figure 3.9: Low resolution FAB mass spectrum of cabamate derivative 49 150 Figure 3.10: Selected NOE (left) and HMBC (right) correlations 151 Figure 3.11: 'H NMR spectrum of cambamate derivative 49 (500MHz, DMSO-d6)....154 Figure 3.12: Selected NOE results of cambamate derivative 49 in DMSO-d 6 155 Figure 3.13: 1 3 C NMR spectrum of cambamate derivative 49 (DMSO-d6) 156 Figure 3.14: 2D H M Q C spectrum of cambamate derivative 49 (DMSO-d6) 157 Figure 3.15: 2D COSY spectrum of cambamate derivative 49 (500MHz, DMSO-d6)..158 Figure 3.16: G2 checkpoint inhibition by staurosporine and derivatives 160 Figure 3.17: ' H / 1 3 C NMR spectrum of staurosporine (10) (500MHz, CDC13) 163 XIV Figure 3.18: Isolation scheme of aaptamine and its derivatives from RJA-96-57 167 Figure 3.19: ' H / 1 3 C NMR spectrum of aaptamine (12) (100MHz, DMSO-d 6) 169 Figure 3.20: ' H / I 3 C NMR spectrum of isoaaptamine (50) (100MHz, DMSO-d 6) 171 Figure 3.21: ' H / I 3 C NMR spectrum of demethyl(oxy)aaptamine (51) (DMSO-d6) 173 Figure 3.22: Low resolution EI mass spectrum of 52 175 Figure 3.23: ' H / 1 3 C NMR spectrum of 52 (DMSO-d6) 176 Figure 3.24: 2D H M B C spectrum of 52 (500MHz, DMSO-d 6) 177 Figure 3.25: 2D HMQC spectrum of 52 (500MHz, DMSO-d 6) 178 Figure 3.26: 2D COSY spectrum of 52 (500MHz, DMSO-d 6) 179 Figure 3.27: Isolation scheme of diterpene lactone derivatives from P. curatellifolia.A 85 Figure 3.28: Selected H M B C correlation in 53. 187 Figure 3.29: Low resolution E l mass spectrum of diterpene ester 53 188 Figure 3.30: ' H / I 3 C NMR spectrum of diterpene ester 53 (CDC13) 189 Figure 3.31: 2D COSY spectrum of diterpene ester 53 (500MHz, CDC13) 190 Figure 3.32: 2D HMQC spectrum of diterpene ester 53 (500MHz, CDC13) 191 Figure 3.33: Effects of lactone 55 on cells arrested in G2 193 Figure 4.1: Isolation of ingenamine and madagamine alkaloids from X. ingens 205 Figure 4.2: Low resolution EI mass spectrum of ingenamine H (56) 209 Figure 4.3: ' H / I 3 C NMR spectrum of ingenamine H (56) (C 6D 6) 210 Figure 4.4: 2D COSY spectrum of ingenamine H (56) (500MHz, C 6 D 6 ) 211 Figure 4.5: Substructure of fragment C 8 H | 0 O and selected H M B C correlations 212 Figure 4.6: The tricyclic core and four appendages with selected HMBC correlations..213 XV Figure 4.7: Substructures of the two side chains and selected H M B C correlations 215 Figure 4.8: 2D H M Q C spectrum of ingenamine H (56) (500MHz, C 6 D 6 ) 216 Figure 4.9: Low resolution FAB mass spectrum of ingenamine I (57) 221 Figure 4.10: 'H / l 3 C APT NMR spectrum of ingenamine I (57) (C 6D 6) 222 Figure 4.11: 2D COSY spectrum of ingenamine I (57) (500MHz, C 6 D 6 ) 223 Figure 4.12: 2D HMQC spectrum of ingenamine I (57) (500MHz, C 6 D 6 ) 224 Figure 4.13: Low resolution FAB mass spectrum of ingenamine J (58) 227 Figure 4.14: ' H / 1 3 C NMR spectrum of ingenamine J (58) (CD 3OD) 228 Figure 4.15: 2D COSY spectrum of ingenamine J (58) (500MHz, CD 3 OD) 229 Figure 4.16: 2D H M Q C spectrum of ingenamine J (58) (500MHz, CD 3 OD) 230 Figure 4.17: Low resolution FAB mass spectrum of ingenamine K (59) 233 Figure 4.18: 'H/ ' 3 C NMR spectrum of ingenamine K (59) (C 6D 6) 234 Figure 4.19: Tricyclic core structure with selected HMBC and COSY correlation 235 Figure 4.20: 2D COSY spectrum of ingenamine K (59) (500MHz, C 6 D 6 ) 237 Figure 4.21: 2D H M Q C spectrum of ingenamine K (59) (500MHz, C 6 D 6 ) 238 Figure 4.22: Low resolution FAB mass spectrum of ingenamine L (60) 241 Figure 4.23: ' H / I 3 C NMR spectrum of ingenamine L (60) (Methanol-d4) 242 Figure 4.24: 2D COSY spectrum of ingenamine L (60) (500MHz, Methanol-d4) 243 Figure 4.25: Selected COSY and HMBC correlations of one side chain (C21-C30) 244 Figure 4.26: 2D HMQC spectrum of ingenamine L (60) (500MHz, Methanol-d4) 246 Figure 4.27: Proposed biogenesis for the ingenamine skeleton 249 Figure 5.1: HPLC traces on U V spectrum (228 nm) with flow rate of 1.0ml/min and corresponding MS of peaks E and F 256 XVI Figure 5.2: Isolation scheme of" antimitotic agents from Ilex macrophylla 257 Figure 5.3: ' H / 1 3 C NMR spectrum of ilexsaponin B1 (73) (DMSO-d6) 259 Figure 5.4: ' H / I 3 C NMR spectrum of 10-deacetyltaxuyunnanine A (74) 261 Figure 5.5: ' H / I 3 C NMR spectrum of 7-(p-xylosyl)-10-deacetyltaxol C (75) 262 Figure 5.6: Mitotic arrest by paclitaxel analogs 264 Figure 5.7: Cytotoxicity of paclitaxel analogs 265 Figure 5.8: Structure of paclitaxel (76), 10-deacetyltaxuyunnanine A (74) and 7-((3-xylosyl)-10-deacetyltaxol C (75) 266 Figure 5.9: Effect of ilexsaponin B1 (73) on anti-mitotic activity of 10-deacetyl-taxuyunnanine A (74) and 7-(p-xylosyl)-l0-deacetyltaxol C (75) 268 LIST O F PICTURES Picture 1: Soft coral Pachyclavularia violacea 50 Picture 2: Sponge Aaptos auberitoides 166 Picture 3: Sponge Xestospongia ingens 203 LIST O F ABBREVIATIONS [a]d ' specific rotation at wavelength of sodium D line at 25 C Ac acetyl APT attached proton test B.C. British Columbia br broad bz benzyl c concentration °C degree Celsius CeD6 deuterated benzene CDCI3 deuterated chloroform CD2CI2 deuterated dichloromethane cm"1 wave number CoA coenzyme A COSY Correlation SpectroscopY 5 chemical shift in parts per million d doublet ID one-dimensional 2D two-dimensional Da Daltons DEPT Distortionless Enhancement by Population Transfer DMSO dimethyl sulfoxide DMSO -d6 deuterated dimethyl sulfoxide E D 5 0 effective dose resulting in 50% response EIMS electron impact mass spectrometry Enz enzyme ESI ElectroSpray Ionization Et ethyl EtOAc ethyl acetate FTIR Fourier transform infrared g gram GC gas chromatography h hour Hex hexane HMBC heteronuclear multiple bond multiple quantum coherence HMQC heteronuclear multiple quantum coherence HOAc acetic acid HPLC high performance liquid chromatography HRDCIMS high resolution desorption chemical ionization mass spectrometry HREIMS high resolution electron impact mass spectrometry HRFABMS high resolution fast atom bombardment mass spectrometry Hz hertz i signal due to impurity IC50 Inhibitory Concentration resulting in 50% response IPP Isopentenyl diphosphate (iPr)2NH diisopropyl amine IR infrared J scalar coupling constant LH-20 Sephadex LH-20 m multiplet M + molecular ion Me methyl MeOH methanol MeOH-d 4 deuterated methanol ml milliliter m.p. melting point m/z mass to charge ration |aM micro (10"6) Mole NMR Nuclear Magnetic Resonance NOE Nuclear Overhauser Effect P388 murine leukemia cell line P388 PNG Papua New Guinea PP Pyrophosphate ppm parts per million q quartet ROSEY rotating frame Overhauser effect spectroscopy s singlet or signal due to solvent SCUBA self-contained underwater breathing apparatus sp. species t triplet TFA trifluoroacetic acid THF tetrahydrofuran T L C thin layer chromatography UBC University of British Columbia U V ultraviolet w signal due to water XXI A C K N O W L E D G E M E N T S I profoundly appreciate my research supervisor, Professor Raymond J. Andersen. His enthusiasm and dedication is the source of my interest on natural products. I learn the meaning of science and research from his serious attitude to science and high standard on research. Under his guidance, I build up a solid professional foundations and develop a strong confidence. I also extend my appreciation to Professor Michel Roberge. He not only provides two novel bioassay development as our isolation guidance, but also helps me deeply understand the biological function of secondary metabolite at the molecular biology level. Thanks Dr. Xiuxian Jiang and Lynette Lim for the assistance by doing bioassay for G 2 and antimitotic projects. Regards and appreciation are also due to the members of my colleagues for assistance in the collection of marine specimens used in this work as well as their cooperation and friendship. I am specifically indebted to Dr. David Williams, Bruno Cinel, Dr. Jeffery Gerard, Dr. Fangming Kong for their invaluable helps and suggestion over past few years. The helpful proof-reading provided by my colleagues: Bruno Cinel, Todd Barsby, Kyle Craig in the preparation of this thesis is highly appreciated. The technical assistance provided by the staffs, Dr. Nick Berlinson and Marietta Austria in NMR services, and by the staffs Dr. Eigendorf and Linda in Mass Spectrometry services are gratefully acknowledged. Chapter I: Introduction Cancer remains the second leading cause of death in North America.1 It has been estimated that the probability of Americans developing cancer sometime during their lifetime is roughly 1 in 2 for males and 1 in 3 for females. The major causes of cancer deaths in Americans are lung, breast, prostate, and colon/rectum solid tumors. Chemotherapy remains an important frontline tool for the systemic treatment of solid tumor cancers.2 The ideal anticancer drug would be a 'magic bullet' that selectively targeted cancer cells without having any adverse effects on normal cells. The currently available anticancer drugs are far from this ideal. Therefore, the discovery and development of new anticancer drugs is still one of medicine's most challenging issues. Natural products research has led to the discovery of many clinically useful drugs that play a major role in the treatment of human diseases. Many natural products are currently in clinical use against cancer.2 A survey by NCI scientists showed that natural products or synthetic analogs of natural products make up a significant proportion of anticancer 'lead' compounds currently in preclinical evaluation or clinical trials/' Recent trends in anticancer drug discovery from natural sources emphasize investigation of the marine environment yielding numerous, often highly complex chemical structures4,3'6 It is significant that a number of marine invertebrate metabolites, including the ecteinascidins,7 byostatin I,8 and dolastatin 10,9 are already in clinical trials as anticancer agents.10 Recent research in molecular genetics has provided enormous advances in our understanding of cancer biology. In particular, there are now many well documented genetic differences between cancer cells and normal cells, that have provided exciting 2 new opportunities for developing selective mechanism-based inhibitors of specific biochemical processes that are essential to the malignant phenotype of cancer cells."'1 2'1 3'1 4 These new targets offer the opportunities for developing anticancer drugs with higher selectivity and much improved therapeutic margins. Subsequently, constructing novel target directed screening systems resulted in the current anticancer drug screening paradigm known as 'high throughput screening' (HTS). I S The primary role of HTS is to identify 'lead' compounds and supply directions for their optimization. The key elements in successful HTS programs are the judicious choice of validated molecular targets and access to the greatest number and structural diversity of compounds. Combinatorial chemistry libraries, traditional synthetic chemical libraries, and natural product extracts represent the most important pools of chemical entities available for HTS programs. Regulation of critical events of the cell-division cycle is the molecular basis for controlling proliferative disorders such as cancer. Many small molecule natural products inhibit progression of the cell cycle by binding to a protein required for cell division. Irrespective of their targets, natural products have been invaluable as agents to achieve arrest at specific points in the cell cycle. The cell cycle G2 checkpoint is one such target against cancer cells, although the details of G2 checkpoint pathway in cells are still not fully understood. The research described in this thesis has focused on using high throughput cell-based bioassays designed to screen natural product extracts for G2 checkpoint inhibitors and antimitotic agents. The research was carried out in collaboration with Dr. Michael Roberge in the UBC Biochemistry Department who developed the assays. As a result, several G2 checkpoint inhibitors were isolated and identified from marine organism and terrestrial plant extracts. Once an inhibitor is discovered, a chemical genetics approach is possible. This entails using the small molecule as a tool to identify its molecular target and functionally relevant residues of proteins on the G2 checkpoint pathway. Identification of the direct targets of natural product inhibitors of the cell cycle implicates specific proteins or activities as being essential to cell cycle related functions. Thus, the crucial protein and related factors could be elucidated, and a deeper understanding of the mechanisms of G2 arrest in mammalian cells could be gained. Subsequently, the protein can be used as the basis of a new target-directed HTS assay for anticancer drug discovery. A brief description of marine natural products as a chemical diversity pool, the high throughput cell cycle G2 checkpoint inhibitor and antimitotic bioassays, and chemical genetics methods will be presented below as an introduction to the thesis research. 1.1 Marine natural products as drug lead candidates. The following review of marine natural products will focus on their potential as pharmaceuticals, and will highlight some compounds that have made significant contributions to our understanding of cellular processes at the biochemical level. Didemnin B 1 6 (1), isolated from the ascidian Trididemnum solidum was one of the first marine secondary metabolites to reach human clinical trials. The didemnins exhibit anticancer as well as antiviral and immunosuppressant activities by inhibiting protein biosynthesis independant of the cell cycle without interacting with D N A . 1 6 ' 1 7 Bryostatin 1 ( 2 ) , isolated from the Eastern Pacific bryozoan Bugula neritina by George Pettit,18 is active against leukemia and cancers of the breast, skin, colon, lungs, cervix, ovaries and lymphatic system, and has been in clinical trials for the past few years. Discodermolide ( 3 ) , which was obtained from the Caribbean sponge Discodermia dissoluta by researchers at the Harbor Branch Oceanographic Institution in Florida, 1 9 is a new and structurally unique antimitotic agent that stabilizes microtubules in breast carcinoma cells more potently than taxol. It is now in preclinical evaluation for treatment of taxol-resistant cancers. NH 2 3 Eleutherobin (4), isolated from the Western Australian soft coral Eleutherobia sp. by a team at Scripps Institute of Oceanography in California, is a diterpene glycoside.20 It 5 is a potent antimitotic agent that also stabilizes microtubules and has extraordinary selectivity against breast, renal, ovarian and lung cancer cell lines in vitro. OH Curacin A (5), isolated from the Caribbean cyanobacterium Lyngbya majuscula by Bill Gerwick's Group, is one of a family of natural lipids found to inhibit microtubule formation.21'22 In preclinical trials this compound has shown a potent and significant effect in vivo against human lung cancers transplanted into mice. There are also many successes for other pharmaceutical purposes. Pseudopterosin C (6), isolated by Bill Fenical's group at Scripps Institute of Oceanography from the Caribbean sea whip Pseudopterogorgia elisabethae,27, is an anti-inflammatory compound and helps to prevent skin aging. It has already reached the market as an additive to Este Lauder skin care products. Another successful example is contignasterol (7), isolated from the sponge Petrosia contignata by the Andersen group.24 A synthetic analog of contignasterol is being developed by Inflazyme Pharmaceutical Ltd. in Vancouver as an antiasthma agent. The compound designated IPL576 has successfully completed phase I clinical trials. 6 The main types of marine organisms being investigated for their natural products are sponges (Porifera), soft corals (Octocorallia), sea squirts (Ascidiaceae), sea mats (Broyozoa or more correctly, Ectoprocta), sea slugs (Mollusca) and marine microorganisms. All of the invertebrates are soft-bodied and either sessile (fixed to a surface such as rock) or slow moving. In addition, most of them are brightly colored which makes them very obvious to potential predators. As a result, these invertebrates need effective methods of self-defense, and it has been suggested that this is one reason they produce unusual and frequently bioactive secondary metabolites. It is also thought that organisms such as these that have not evolved a cellular immune system, instead develop a chemical immune system, and this also provides a second rationale for them being a rich source of potentially useful compounds. An analysis of the phyletic distribution of new marine natural products reported in the literature for the period of 1997 shows that majority of these compounds come from sponges (Figure 1.1).^  Sponges have a wider range of biosynthetic capabilities than any other group of marine invertebrates. They are also known to have few predators, perhaps because many species produce toxic or distinctive odorous metabolites that may prevent other organisms from settling on their surfaces or eating them. 7 Molluscs 7.79% Miscellaneous 0.32% Algae 10.06% Microbes 9.74% E3 Algae Bryozoans 1.46% • Microbes • Bryozoans • Tunicates HEch inoderms Tunicates 9.58% OCoelenterates Sponges 42.86% • Sponges • Molluscs Echinoderms 5.36% • Miscellaneous Coelenterates 12.82% Figure 1.1 Phyletic distribution of marine natural products 1997 Because only a small percentage of the marine organisms living in the biosphere have been described to date, there exists an enormous and still undiscovered reservoir of marine natural products. The few examples shown above illustrate some of the tremendous structural diversity found in marine natural products. These compounds affect numerous targets involved in eukaryotic cell signaling processes. For many marine secondary metabolites it is a challenge to obtain enough compounds for exhaustive biological activity profiling and clinical testing. Therefore, marine natural products are most useful as lead compounds that identify new bioactive chemical skeletons for synthetic production or modification using the strategies of combinatorial chemistry. Marine biotechnology focuses on another solution to the supply problem. Marine ecology2 6 , 2 7 research has shown that marine microorganisms are substantially involved in the biosynthesis of marine natural products isolated from macroorganisms such as invertebrates. For example, the presence of filamentous bacterial symbionts in lithistid sponges has been identified by scanning electron microscopy.28 Many researchers have 8 attempted to culture symbiotic microorganisms in order to generate useful metabolites which were first isolated from their marine host. To date, this has not been successful. It is also possible that those secondary metabolites with therapeutic potential from slow growing microbes or invertebrates could be produced more rapidly by incorporating the relevant biosynthetic genes into expression systems such as Escherichia coli. Pharmacological evaluation of marine natural products has also undergone an evolution over the past two decades, beginning with the early investigations of toxins, followed by studies of cytotoxic and antitumor activity, to the present day, where a myriad of activities based on whole-animal models and receptor-binding assays are being pursued. In the beginning, toxic principles dominated the spectrum of biological activities isolated from marine sources. This fact may partly be due to the major application of cytotoxcity guided screening assays. Cytotoxcity bioassays are traditional cell-based assays. The cytotoxcity data reported most frequently are from mouse lymphocytic leukemia cell cultures (P 3 8 8 and L ) 2io) and human epidermoid carcinoma of the mouth cell cultures (KB). Results are expressed in terms of the dose which inhibits cell growth to 50% of the control growth (such as ED50, ID 5 0 , IC50). These are usually expressed in ug/ml. With the advent of genomics research and newer molecular biology tools for developing bioassays, more sophisticated biological assays in addition to cell-based assays are being employed routinely in the drug discovery paradigm. Consequently, in recent years a notable number of natural product-derived agents, such as many protein kinase inhibitors have been discovered by employing a mechanism-based screening approach involving cellular or biochemical targets in their assay design. Therefore, any novel target-directed screening assay may result in identification of a new lead structure even from crude extract collections exclusively comprising compounds that are already described. One of my research objectives was to cooperate with Dr.Roberge's biochemical group to utilize a new high-throughput G2 checkpoint inhibitor bioassay to discover new inhibitors from marine organism extracts. High-throughput screening (HTS) identifies crude extracts of interest from natural sources very quickly. This demands rapid characterization and structure elucidation by natural product chemists. Elaboration of LC-MS and LC-NMR techniques to accelerate structure elucidation of bioactive metabolites is currently underway in our lab. Combination of those techniques with a potentially rich marine source will probably contribute substantially to more interest in marine natural products for application in drug discovery and the regulation study in cell-division cycle. 10 1.2 Cell cycle checkpoint inhibitor bioassay Michel Roberge's G 2 specific cell cycle checkpoint inhibitor bioassay is a cell based high-throughput screening assay that is suitable for target-directed screening of natural product extracts.29 This bioassay will be introduced with a brief discussion of the cell cycle. 1.2.1 Cell cycle The eukaryotic cell cycle is a tightly controlled series of events that ensures the continuous growth of cells and the faithful propagation of genetic information to daughter cells. Breakdown of cell cycle regulation often leads to the accumulation of genetic lesions that can cause unchecked proliferation, altered phenotype or cell death. The eukaryotic cell cycle can be divided into two major phases, M phase and interphase. M phase includes the process of mitosis, during which duplicated chromosomes are separated into two nuclei, and cytokinesis, during which the cell divides into two daughter cells. During interphase, the cell undergoes the processes of growth and replication of DNA. Interphase can be further divided into three stages. The first stage, GI or Gap 1 phase of the cell cycle starts at the conclusion of mitosis and ends with the onset of DNA synthesis. The second stage is the S phase during which the cell replicates its DNA. The G 2 phase is the second gap phase that begins upon the completion of DNA replication and ends with the onset of mitosis (Figure 1.2). 11 G2 checkpoint Figure 1.2 Cell cycle and DNA damage checkpoints The cell cycle can be divided into four stages. Interphase is composed of the gap stages G l and G2 when cellular growth occurs and the S phase when the D N A is replicated. The M Phase involves the cellular processes involved in the division of cell into two daughter cells. Cells would enter the GO stale if they become nonproliferative.The cell cycle is regulated by two D N A damage checkpoints. If cells have been exposed to D N A damage, the G l checkpoint regulates entry into S phase and the G2 checkpoint regulates entry into mitosis. G l checkpoint 1.2.2 Cell cycle checkpoint The ability of eukaryotic cells to maintain genomic integrity is vital for cell survival and proliferation. The inability to maintain genomic stability would result in mutations that could lead to cell death or give rise to cancers. There are built-in mechanisms known as checkpoints employed by the cell to protect itself from the harmful effects of DNA damage. When checkpoints are activated by DNA damage, a signal cascade is elicited which causes a transitory halt in cell-cycle progression. It is believed that this allows time for DNA repair to occur in order to prevent the replication of damaged DNA templates and segregation of broken chromosomes.30 After cells have been exposed to DNA damage, they can arrest at either the G1 phase or the G2 phase, depending on which position of the cell cycle they are in. The Gl checkpoint functions to prevent entry into the S phase and the G2 checkpoint prevents entry into mitosis from G2 (Figure 1.2). 12 1.2.3 GI checkpoint DNA damage triggers several signal transduction pathways that lead to DNA repair coupled with a halt in cell-cycle progression or to apoptosis (programmed cell death), depending on the severity of DNA damage. The GI checkpoint signaling cascade is mediated by the transcription factor p53 whose activity rises after DNA damage. This is achieved in part by the stabilization of p53, which is normally rapidly degraded, and in part by an increase in p53 transactivation activity/ The mechanism by which DNA damage is detected and signaled to p53 remains unclear but evidence implicates the kinases A T M and A T R . 3 2 The stimulation of p53 by DNA damage is believed to be mediated by the gene product of the A T M gene," " a 370 kD protein kinase with a COOH terminal domain similar to the catalytic subunit of phosphoinositide 3-kinase.34 Mutations in the A T M gene cause a wide variety of cellular defects including acute sensitivity to radiation, genomic instability and the inability to activate cell cycle checkpoints.35'36 A structurally related kinase, ATR has also been shown to mediate the activity of p53.32 A T M and ATR are believed to act as sensors recognizing DNA damage and transducing the signal to p53. A T M and ATR can phosphorylate p53 in vitro at Serl5 which is phosphorylated in vivo in response to DNA damage." '" The phosphorylation of p53 at SeiT5 leads to the stabilization of p53 by inhibiting its interaction with Mdm2, a protein that targets p53 for degradation via the ubiquitin-proteasome pathway. The increase in intracellular pools of p53 causes an increase in its activity as a transcription factor stimulating the transcription of downstream effector genes. The most notable of these genes is p21C I P 1, which inhibits the kinase activity of various cyclin/CDK complexes that are involved in the transition to 13 S phase, including cyclinA/CDK2." ' As a result, the Rb protein remains hypophosphorylated and cells are unable to enter the S phase, leading to G l cell cycle arrest (Figure 1.3). Figure 1.3: The G l checkpoint pathway After exposure to DNA damage, intracellular p53 level increase resulting in the activation or" expression of p21 protein. p21 protein inhibits the function of cycl in/cdck2 which is to phosphorylate R B . R B remains hypophosphorylated resulting in inability to enter S phase and consequently G l arrest. 1.2.4 G2 checkpoint The ultimate downstream target for the mammalian G2 checkpoint pathway is the Cdc2 kinase. After DNA damage, Cdc2 is maintained in its inactive phosphorylated state to prevent the cell from entering mitosis. A key component of the G2 checkpoint pathway has been identified, the Chkl kinase. In response to DNA damage, Chkl is activated and phosphorylated in an ATM-dependent manner.41 Chkl is believed to inhibit the Cdc2 kinase via the regulation of weel 4 2 and cdc25C 4 3 Chkl is believed to inactivate Cdc2 by two different mechanisms. First it has been G l arrest Entry into S phase 14 demonstrated by O'Connell et at2 that in fission yeast, Chkl can directly phosphorylate weel in vitro. In addition, they have demonstrated that weel is required for cell cycle arrest induced by overexpression of Chkl. These observations suggest that in response to DNA damage weel is phosphorylated and activated by Chkl in vivo and this results in the inhibition of the Cdc2 kinase. Alternatively, Chkl has been found to phosphorylate cdc25C at Ser216, enhancing its binding to 14-3-3 proteins which is believed to target cdc25C for nuclear export 4 3 , 4 4 Therefore, after DNA damage, cdc25C becomes cytoplasmic and sequestered away from its nuclear substrate Cdc2. Cdc2 remains inactive and phosphorylated which leads to G2 arrest (Figure 1.4). Recent studies have shown that cells expressing a mutant forms of Cdc2, which can not be phosphorylated at Thrl4 and Tyrl5, have a reduced but still substantial G2 arrest45 This suggests mechanisms other than inhibitory phosphorylation of Cdc2 may exist to cause G2 arrest. One such mechanism could be the regulation of cyclin Bl's intracellular localization in response to DNA damage. In order for cells to enter mitosis, cyclin B l needs to be translocated into the nucleus in order to initiate downstream mitotic events. It has been found that cyclin B l remains cytoplasmic in cells that have been arrested in G2 by DNA damage. In addition, cells expressing a mutant form of cyclin B1 that targets it to the nucleus in all cell cycle phases can partially override G2 arrest in response to DNA damage 4 5 This suggests that DNA damage acts not only by stabilizing the inhibitory phosphorylations of Cdc2 but also by maintaining the cytoplasmic localization of cyclin B1. 15 D N A d a m a g e • inactive active act ive act ive inact ive MITOSIS Thr161 Thr 161 inact ive act ive Figure 1.4: Current model of the mammalian G2 checkpoint pathway A T M acts as a sensor recognizing D N A damage and initiating a signal cascade that results in the phosphorylation and activation of the checkpoint kinase, C h k l . Activated Chkl phosphorylates cdc25C at residue S216 that is believed to mediate binding to 14-3-3 resulting in functional inactivation of the complex. Inactive cdc25C fails to activate cdc2 causing arrest in G2. The broken arrows between A T M and Chkl indicate that A T M activates Chkl by means that have yet to be elucidated. In the mammalian G2 checkpoint pathway, the signaling pathway upstream of Chkl remains little understood. What is known is that Chkl is phosphorylated and activated in an ATM-dependent manner.41 How the signal is propagated from A T M to Chkl is unclear but clues can be provided by genetic studies in the yeast S. pombe which also has a G2 checkpoint pathway. The organization of the checkpoint pathway downstream of Chkl in S. pombe is similar to that in mammalian cells. The components upstream of chkl are believed to involve a group of proteins radl, rad3, rad9, radl7, rad26 and husl that are required for cell cycle arrest in response to DNA damage.45'47 These proteins are involved in the recognition of DNA damage and the transduction of the checkpoint signal to chk 1. The nature of the interaction between chkl and the rad proteins remains unclear but it is believed that a protein Crb2, phosphorylated in response to DNA damage, may act as a 48 mediator between chkl and the rad proteins (Figure 1.5). Putative human homologues of various upstream components of yeast checkpoint proteins have been identified including hRadl, hRad9, hRadl7 and hHusl. 4 9 However, the roles of these proteins in mammalian checkpoint response have yet to be determined. DNA Damage wee 1 rad1 rad17 rad3 rad9 rad26 Crb2 chkl rad24 ® cdc25 G2 arrest Mitosis Figure 1.5 Overview of G2 checkpoint pathway in fission yeast S. pombe D N A damage is sensed by a number of proteins including rad 1, rad 17, rad3, rad9 and rad26. These proteins stimulate the activation of c h k l possibly by stimulation of Crb2. C h k l along with rad24 inhibits the cdc2 kinase via the inhibition of the phosphatase cdc25 and the activation of the kinase weel. 17 1.2.5 G2 checkpoint inhibitors in cancer therapy Cancer can be defined as a disease characterized by an abnormal cellular growth and the ability to invade normal tissues and metastasize to distinct sites. Treatment of cancers often involves the use of antitumor drugs that have a wide variety of cellular effects. They may be drugs that have anti-proliferative effects by interfering with mitotic microtubule function or they may be cytotoxic agents by binding to or damaging DNA or inhibiting nucleic acid synthesis. One drawback of these classes of compounds is that not only do they kill cancerous cells but they are also harmful to normal cells limiting their effectiveness in eradicating tumors. However, over the past 20 years numerous advances have been made in understanding the genetic changes underlying the transformation of normal cells into cancerous cells. A search has been undertaken for newer and more effective chemotherapeutic agents that would selectively target these genetic differences. One of the most common genetic mutations, found in over 60% of cancers, is the loss of p53 function.50 The loss of p53 leads to the cell's inability to activate the G l checkpoint in response to DNA damage. This mutation is believed to lead to increase genomic instability and proliferation which are hallmarks of cancers. One approach to treating cancers that have mutated p53 involves the introduction of WT p53 into cancer cells by gene therapy. It is hoped that this would restore the cell's ability to undergo apoptosis which is dependent on p53. However there are currently serious limitations to this approach as the genes have to be injected directly into the tumor limiting their usefulness to solid tumors and not to cancers that may have metastasized.51 An alternative approach is to take advantage of the loss of p53 function to 18 kill cancer cells with mutated p53 by introducing a genetically modified adenovirus. This human respiratory virus requires the disabling of p53 function in order to infect and kill the host cell because the p53 gene activity prevents viral replication. Removal of the viral gene responsible for disabling p53 from the adenovirus would allow it to infect and kill only cells with mutated p53. This treatment has showed some success but requires that the adenovirus be directly injected into a solid tumor in order to be effective.52 Again, this approach limits its use for metastatic cancer. Cancer cells with mutations in p53 have lost the ability to activate the GI checkpoint. It may be possible to exploit this by treating tumor cells with mutated p53 with agents that prevent the activation of the G2 checkpoint and exposing them to DNA damage. The tumor cells would be unable to arrest at either checkpoint and as a result, would enter mitosis with extensive DNA damage leading to cell death. On the other hand, normal cells exposed to DNA damage and G2 checkpoint inhibitors would still be able to arrest at the GI checkpoint to repair damaged DNA, resulting in increased cell survival over cells with mutated p53 (Figure 1.6). This principle has been demonstrated in vitro: using a variety of paired cell lines differing in their p53 status, several groups have shown selective sensitization of cells with mutated p53 to treatment with DNA damage and G2 checkpoint inhibitors.53'54 ,55 A number of G2 checkpoint inhibitors have been discovered but they remain unsuitable for use in chemotherapy because they exhibit a wide variety of side effects. The most notable is 7-hydroxystaurosporine (UCN-01,11) which is currently undergoing phase I clinical trials. However, UCN-01 has a drawback'in that it binds non-specifically to human a-acidic glycoprotein in the plasma, reducing its effectiveness in chemotherapy.56 19 a) Wildtype p53 cell Gl checkpoint 1 G1 b) mp53 cell Gl checkpoint G1 G2 checkpoint i G2 M G2 checkpoint G2 M Figure 1.6 Rationale for use of G2 checkpoint inhibitors in cancer therapy When cells with wildtype p53 are exposed to D N A damage and G2 checkpoint inhibitors, a large proportion wi l l arrest at G l checkpoint to repair their D N A to enhance cell survival. Cells with mutated p53 exposed to the same conditions wi l l be unable to arrest at either the G l or the G2 checkpoint and wi l l enter mitosis with D N A damage that wi l l result in cell death. Staurosporine (10) is also a G2 checkpoint inhibitor, but it is a broad-based protein kinase inhibitor that has been demonstrated to have no antitumor activity in vivo.51 Other G2 checkpoint inhibitors that have been identified include the purine analogues, caffeine (8) and pentoxifylline (9), and the aromatic alkaloid aaptamine (12). However, their numerous pharmacological effects preclude their effectiveness in chemotherapy.58 The protein phosphatase inhibitor okadaic acid (13) can also act as a G2 checkpoint inhibitor.59 Its drawback is that it can induce premature entry into mitosis in normal cells. Therefore, there is a need to find G2 checkpoint inhibitors that can act more 20 H H N H N H 10 11 specifically and can exert their effects at low concentrations to limit their cytotoxicity to normal cells. Michel Roberge's G2 checkpoint inhibition bioassay29 involves the use of an MCF-7 epithelial breast cancer cell line that expresses a dominant negative mutant p53. The cells are subjected to ionizing radiation to induce DNA damage and become arrested in G2. Marine extracts are added and cells entering mitosis are quantified by ELISA using the TG -3 antibody which recognizes a phosphorylated form of nucleolin present only in 21 13 mitotic cells. G2 checkpoint inhibitors are purified from positive extracts using assay-guided fractionation and their structures are elucidated via spectroscopic analysis (Figure 1.7). MCF-7 mp53 in 96 well plate I Irradiate with 6.5Gy I After 16 hours add extracts and nocodazole Cells exposed to G2 Cells not exposed to checkpoint inhibitors G2 checkpoint inhibitors enter mitosis remain arrested in G2 Quantitate mitosis by ELISA using TG3 i I Positive signal Negative signal Figure 1.7 G2 checkpoint inhibitor assay The assay used to efficiently screen for G2 checkpoint inhibitors is outlined. MCF-7 mp53 cells are platec in 96 well plates and exposed to 6.5 grays of Gama irradiation. After 16 hours, when cells are arrested in G2, extracts are added along with nocodazole for 6-8 hours. Cells exposed to a G2 checkpoint inhibitor will enter mitosis which is detected as a positive signal by ELISA using TG-3 antibody. 22 1.2.6 Novel G2 checkpoint inhibitors discovered by the assay. Thirteen hundred marine extracts were initially screened in the G2 checkpoint inhibitor assay. Four marine extracts were found to contain activity. These extracts were subjected to assay-guided fractionation to purify the active components and their chemical structures were elucidated. The first G2 checkpoint inhibitors identified by my colleague'Bruno Cinel were hymenialdisine (14) and its derivative debromohymenialdisine (15), which were isolated from the sponge Styllessa flabelliformis collected in the waters of Papua New Guinea. The second positive extract obtained from cultures of a marine bacterium, was found to contain staurosporine (10) and its semisynthetic analog, oxazolidine derivative (17). The bacterial isolate was obtained from the surface of a Northeastern Pacific Ocean sponge. Finding staurosporine was encouraging as it validated the ability of the assay to identify G2 checkpoint inhibitors. The third extract from the ascidian Didemnum granulatum which was collected off the coast of Brazil, was shown by Dr. Roberto Berlinck to contain a structurally novel G2 checkpoint inhibitor, isogranulatimide (16).29 The fourth active extract was from the soft coral Pachyclavularia violacea and it was found to contain pachyclavularolide F (18). O O 14 15 23 H N H N. N H 16 Subsequently, more than 80,000 extracts from the National Cancer Institute (USA) open repository were screened in the G2 checkpoint inhibitor assay, and one of the active extracts was from Parinari curatellifolia. The roots of this tropical bush are used by traditional healers in Zimbabwe to treat delirium.61 The active components from the crude extracts turned out to be diterpene lactones related to 19. The details of the last two projects are provided in the next chapter. 1.2.7 G2 checkpoint inhibitors as tools in cell cycle research As outlined before, much of our knowledge about the human G2 checkpoint pathway has been obtained by studying the equivalent pathway in yeast and then searching for the human homologues of yeast cell cycle proteins. However, it is clear that mammalian cells also have components not present in yeast. For example, the tumor o 24 suppressor protein p53 has no structurally.related homologue in yeast. Therefore, direct studies of the G2 checkpoint pathway in mammalian cells are also needed. The novel G2 checkpoint inhibitors obtained from the screen described above can be used to gain a deeper understanding of the mechanisms of G2 arrest in mammalian cells. To address this, the Roberge group has attempted to characterize the mechanism of action of the new checkpoint inhibitors by studying their effects on known components of G2 checkpoint pathway in mammalian cells. The goal is to identify the direct targets of the G2 checkpoint inhibitors or to reveal new components of the G2 checkpoint pathway, and to identify the potential in vivo targets of novel G2 checkpoint inhibitors. 25 1.3 Introduction to antimitotic assay: A second bioassay developed by Michel Roberge was the high throughput ELISA and ELICA antimitotic assay. He has used this assay to screen over 5000 crude extracts of marine organisms and some terrestrial plants. Several antimitotic agents were isolated and identified using this bioassay. These preliminary results confirmed the bioassay's reliability. In this section, I briefly introduce antimitotic agents and how Michel Roberge's assay can be used to screen for antimitotic agents. In chapter V, I will present the antimitotic agents that I identified and the relevant results. 1.3.1 Microtubules and tublin In the cell division sequence, the involvement of a dynamic pipe-like protein fiber, known as a microtuble, is known to be essential for the processes to occur. Within every nucleated cell in the human body exist two similar spherical proteins a and (3 tubulin, and these two proteins come together to form an oc-(3 heterodimer. These heterodimers, in the presence of additional GTP at 37°C, can combine in a head-to-tail arrangement to give a long protein fiber composed of alternating a and (3 tubulin, known as a protofilament. After an induction period, typically several minutes, the protofilaments group together to form a C-shaped protein sheet, which then curls around to give a pipe-like structure known as a microtubule. These microtubules typically consist of 12 or 13 protofilaments, with an external diameter of around 24 nm and an internal bore of around 15nm, as shown in Figure 1.8.62 Once formed, these complex protein tubes are not static. They exist in an equilibrium, with dimers constantly adding to one end of the microtubule ( known as the 26 "plus" (+) end), and leaving at the other (the "minus"(-) end). This finely balanced equilibrium, and the resulting control of the length of the microtubles, is vital for a number of their functions within the cell. 2Snm 15 nm \ Figure 1.8 Microtubule structure 1.3.2 Cell replication and the cell cycle6 3 During cell division, the cell must completely duplicate its internal components, including the whole of its DNA, such that it can form two identical daughter cells. Once duplication of the internal components has been completed, the cell must then order its DNA into two identical sets of chromosomes and separate them into two distinct parcels at opposite ends of the cell, ready to form the two nuclei in the daughter cells. Once these new nuclei have fully separated, the cell is then ready to split into the two new daughter cells. This ordering and relocation of the genetic material, which takes about an hour, is known as mitosis, and falls into five distinct phases, as depicted in Figure 1.9. 27 During the first phase, the prophase, the DNA in the nucleus is replicated and the two sets of genetic material organized into two identical daughter sets of chromosomes. Towards the end of prophase, the microtubles required for cell division begin to form and grow toward the newly formed chromosomes. This bundle of microtubles is the structure known as the mitotic spindle. This spindle grows concurrently from two microtuble organizing centers, which begin to separate and migrate toward opposite ends of the cell. Microtubules. Chromosomes. Prophase Prometaphase Metaphase Anaphase. Telophase. Figure 1.9 Stages of mitosis In the next stage, prometaphase, the nuclear envelope rapidly disintegrates and the microtubules attach themselves to the center of the chromosomes at a point known as the kinetochore. The cell then enters metaphase, where the chromosomes gradually become arranged in the plane between the two centrozomes. Once these chromosomes are accurately arranged, the cell abruptly enters anaphase, triggered by specific cellular signals. The daughter chromosomes then begin to separate slowly as the microtubules decay, slowly drawing and guiding the daughter chromosomes apart to opposite ends of the cell. During the final phase, telophase, the chromosomes reach the opposite ends of the cell and new nuclear envelopes form around them. This marks the end of mitosis, and it only remains for the cytoplasm surrounding the nuclei to begin to divide, in a process known as cytokinesis. The nuclei thus become partitioned, eventually dividing to give two new daughter cells. Microtubules, therefore, are intimately involved with the replication of cells. If the microtubules in a tumor cell can be prevented from forming or decaying, the chromosomes can not separate, the cell can not reproduce and hence the tumor can not grow. Thus, agents which interfere with the dynamics of tubulin may also act as inhibitors of cell division. Indeed, a number of these agents have been shown to act as clinically useful anticancer agents. 1.3.3 Antimitotic agents Antimitotics, also called mitotic inhibitors or antimicrotubule agents, are chemical agents that arrest cells in mitosis. Several are clinically important anticancer drugs, including the Vinca alkaloids vinblastine, vincristine, and vinorelbine64 and the taxanes paclitaxel and docetaxel.65 They cause mitotic arrest by interfering with the assembly or disassembly of a and [3 tubulin into microtubules. At high concentrations, the Vinca 29 alkaloids and most other antimitotics, such as curacin, cause complete microtuble depolymerization, whereas the taxanes and eleutherobin cause bundling of microtubules by stabilizing them against depolymerization. At low concentrations, neither depolymerization nor bundling is observed, but there is sufficient alteration in the dynamics of tubulin loss or addition at the ends of mitotic spindle microtubules to prevent the spindle from carrying out its function of attaching to and segregating the chromosomes, and cells arrest in mitosis.66 ,67 Prolonged arrest eventually leads to cell death, either in mitosis or after an eventual escape from mitotic arrest.68'69 These drugs, while extremely valuable, are not ideal. They have numerous toxicities, principally myelosuppression and neurotoxicity. More importantly, many cancers are inherently resistant to these drugs or become so during prolonged treatment.64,65 This is often the result of multi-drug resistance caused by overexpression of P-glycoprotein which functions as a drug efflux pump. Other sources of resistance include increased expression of tubulin isotypes to which a particular drug binds less effectively, and alterations in a and [3 tubulin structure, by mutation or posttranslational modification, that reduce binding. Antimitotics with different chemical structures might show increased specificity to mitotic mocrotubules rather than neuronal microtubules and reduce unwanted side effects, and might be effective against resistant cancers. Many other antimitotics have been discovered70 usually by cytotoxicity screening, or because they showed patterns of cytotoxic activity against panels of cancer cell lines similar to patterns shown by other previous known antimitotic agents.71 However, no rational assay exists to identify antimitotics agents. 30 Therefore, Michel Roberge has developed a high throughput and reliable cell-based screen for antimitotic agents. 1.3.4 A cell-based antimitotic bioassay The TG-3 monoclonal antibody, originally described as a marker of Alzheimer's disease,72 is highly specific for mitotic cells. Flow cytometry shows that TG-3 immunofluorescence is >50 fold more intense in mitotic cells than in interphase cells.73 TG-3 also recognizes mitotic cells in ELISA using microtiter plates.74 In this antimitotic assay,75 human breast carcinoma MCF-7 cells were seeded in 96-well plates and allowed to grow overnight. Crude extracts from natural sources were then added and cells were incubated for 16-20 h. Then cells grown in 96 well plates are lysed and the lysates are transferred to protein-binding ELISA plates for adsorbtion to the plastic surface. The antigen is detected by incubating with TG-3 antibody followed by an HRP-conjugated secondary antibody and colorometric determination of HRP activity. Dr. Roberge's group first tested the suitability of the ELISA for quantifying the activity of antimitotic agents. MCF-7 cells were incubated for 20h with different concentrations of the antimitotic drug paclitaxel, and the proportion of cells arrested in mitosis was measured by counting mitotic figures in the microscope, and by ELISA. Paclitaxel induced mitotic arrest in a concentration-dependent manner with half-maximal activity at lOnM measured by microscopy and at 4 nM measured by ELISA. Roberge found the ELISA to be accurate and reliable, and he used it for most of the screening on the extracts. However, it requires transferring cell lysates to ELISA plates and many solution changes, and is consequently rather slow and labor-intensive. He has 31 recently simplified it for faster and easier drug screening and for automation. The modified assay reduces the time of the procedure and the number of steps by half and does not require transfer of samples to ELISA plates. In this procedure, the cells are fixed with formaldehyde in their culture plate and permeabilized with methanol and detergents, and the TG-3 primary antibody and HRP-conjugated secondary antibody are added simultaneously. Colorimetric detection of HRP activity remains unchanged. Since cell fixation and permeabilization in situ are steps commonly used in immunocytochemistry, he termed the assay Enzyme-linked Immuno-Cytochemical Assay or ELICA. Measurements obtained by ELICA consistently showed smaller standard deviations than those obtained by ELISA, probably because the reduced number of manipulations reduced experimental variation. Overall Roberge finds the ELICA suited to rapid screening of large numbers of extracts and the ELISA more appropriate for precise quantitation of antimitotic activity. 32 1.4 Introduction to chemical genetics: Understanding how nature's network of cellular components carries out cellular functions is a challenging goal for chemists and biologists. The genome sequencing projects currently under way are already providing valuable information relevant to this challenge. However, the compilation of genetic sequence data alone will not elucidate the functional relevance of the newly identified gene products. The next challenge will be to understand how the proteins encoded by the genes function within the milieu of a living system. Achieving this goal requires the ability to alter protein function. Classical genetics has provided the predominant means of altering protein function, resulting in the activation or inhibition of proteins through genetic mutations. For example, inactivation can occur if the mutation abrogates a protein-protein interaction critical for cell signaling , whereas activation can occur if the mutation facilitates a new interaction. It has been widely used in biology. However, methods for comprehensive genetic analysis of mammalian systems are currently limited.76 With whole organisms such as mice, the space requirement and expense of large numbers of animals, their long generation time, small size and the difficulty inherent in identifying and mapping recessive mutations are problematic. In addition, many gene products are essential, redundant or expressed in a temporal or tissue-specific manner. A chemical genetics approach using small molecules that alter protein function directly has the potential to overcome many of the current limitations in genetic analyses of mammalian systems.77 This approach has been given a number of names, including the "pharmacological approach". The traditional genetic approach to understanding protein function can be time-consuming, but it is general. The chemical genetic approach relies 33 on the existence of highly specific ligands, which at this point come primarily from nature and exist for the different proteins. This complementary and direct approach involves the use of small molecules that alter the function of proteins to which they bind (Figure 1.10).77 This process is akin to the generation of mutations in genes, but relies on small molecules, often in the form of a chemical library, as the source of perturbation. As a small molecule in a cell-based assay can specifically alter the function of a gene product from all copies of a gene, a small molecule can be used analogously to an inducible dominant or homozygous recessive mutation. These characteristics circumvent the difficulty of generating these types of mutations in mammalian systems. Also, just as mutation sites can identify functionally relevant coding sequences of genes, small molecules can identify functionally relevant residues of proteins, based on their T O mechanism of interaction. By using small molecule libraries in an appropriate cell-based assay, it should be possible to identify novel gene products on pathways of interest, as well as novel biologically active small molecules from either natural sources or laboratory syntheses.79 This idea is supported by the existence of a wide variety of small molecules that cause a loss of function of their cognate targets, including kinases,80 phosphatases,81 membrane receptors,82 proteases,78 isoprenyl transferases83 and polymerases.84 For example, the antimitotic agent curacin A (5 ) binds to tublin and inactivates tublin function. However, the steroid hormones, such as dexamethasone, activate the transcription of the gene for tublin as the nuclear hormone receptors and consequently active tublin function. Therefore, small molecules could be regarded as ligands, which be capable of either inactivating or activating protein functions (Figure 34 i n a c t i v a t i n g m u t a t i o n i n g e n e C a c t i v a t i n g m u t a t i o n i n gene C 3 i * i zzzr 3 i * i — \genetic J c y t o p l a s m approach ft n u c l e u s /chemical genetic a p p r o a c h cell division* differentiation, cell death, etc. - M e d e x a m c t h a s o n e Figure 1.10: Both small molecules and mutations have been used to explore protein function. This figure emphasizes the relationship of genetics and chemical genetics. In the area of the cell cycle and cell cycle checkpoints, rapamycin has been used to discover the new protein FRAP, which mediates the G l cell cycle checkpoint.88 While searching for G2 cell cycle checkpoint components, Karlene Cimprich, Tae Bum Shin, and Curtis Keith discovered the protein chkl , 8 9 ' 9 0 which mediates the G2 DNA damage checkpoint (Figure 1.11).91 Since a small molecule ligand to chkl or the related protein is not known, we developed a G2 checkpoint inhibitor bioassay to search for this specific ligand from marine natural sources. Several novel G2 checkpoint inhibitors were discovered by this G2 bioassay, such as isogranulatimide (17) from the ascidian Didemnum granularum and the diterpene lactone (19) from Parinari curatellifolia. Now, these two compounds are being used to search for the components of G2 checkpoint pathway to which these small molecules bind. When the target of the inhibitor is known, we could use this inhibitor as a tool to alter the function of this protein target and gain a deeper understanding of this protein function on G2 arrest in mammalian cells. It is these features that provide the means to apply chemical genetics in situations where small molecules that bind to a protein of interest have not yet been identified . Small Molecule-Based Studies of the Cell Cycle. Figure 1.11: Small molecule-based studies of the cell cycle and cell cycle checkpoints and the chemical structures of the natural products that have been used to gain new sight into cell cycle progression. . 36 1.5 Preview of research results in thesis Currently, the most valuable collection of ligands for use in the study of protein function are natural products.92 Many small molecule natural products inhibit progression of the cell cycle by binding to a protein required for cell division, thus helping to determine the function of the protein. However, few natural products were known as cell cycle G2 checkpoint inhibitors, and the details of the G2 checkpoint pathway are still not fully understood. Therefore, the goal of my thesis research was to discover structurally novel G2 cell cycle checkpoint inhibitors in natural product extracts. Active crude extracts were identified via a high throughput bioassay for new G2 checkpoint inhibitors. The assay was developed and conducted by our collaborator Michel Roberge. We proposed that this novel target-directed screening assay might result in new lead structures from crude natural product extracts. As a result, several G2 checkpoint inhibitors were isolated and identified from marine organism and plant extracts by bioassay-guided fractionation. One of the inhibitors pachyclavularolide F (18) was isolated from the soft coral P. violacea. The details of isolation and characterization of pachyclavularolides is described in chapter II. Other checkpoint inhibitors, such as aaptamine isolated from sponge Aaptos auberitoides, staurosporine semisynthetic derivatives from a marine microorganism, and diterpenoids from the tropical bush P. curatellifolia are summarized in chapter III. The diterpenoids from P. curatellifolia are now being used as tools for chemical genetics studies on the G2 checkpoint pathway. Chapter IV contains my research on new metabolites discovered from a marine sponge by the traditional cytotoxicity bioassay guidance. The purpose of this project was 3 7 to discover new cytotoxic natural products from marine sources. Several new ingenamine alkaloid derivatives were isolated. Using Dr. Roberge's high throughput antimitotic assay, we attempted to discover the novel specific arrest agents for the mitotic phase of the cell cycle from crude extracts. As a result, taxol derivatives were found from the Southern American plant Ilex macrophylla, which has never been reported to contain antimitotic agents. This work resulted in the discovery of a new natural source for taxol derivatives, and it also validated the accuracy and efficiency of the new antimitotic assay. The details of this research is described in chapter V. The main parts of this thesis focus on the structure elucidation and characterization of the metabolites isolated during the above studies. References: 1. Parker, S.L.; Tong, T.; Bolden, S.; Wingo, P.A. "Cancer Statistics, 1996" CA Cancer J. Clin. 1996, 65, 5-27. 2. 'A Clinician's Guide to Chemotherapy Pharmacokinetics and Pharmacodynamics' Grochow, L.B.; Ames M . M , Editors, Williams & Williams, Baltimore, Maryland, 1998. 3. Cragg, G.M.; Newman, D.J.; Snader, K . M . / . Nat. Prod. 1997, 60, 52-60. 4. Attaway, DH; Zaborsky, OR. (1993) Marine biotechnology, Vol 1, pharmaceutical and bioactive natural products. Plenum Press, New York. 5. (a) de Vries, DJ; Beart, PM (1995) TiBS, 16: 275; (b) Konig, G M ; Wright AD. Planta Medica 1996, 62, 193. 6. Jensen, PR.; Fenical, W. J IndMicrobiol. 1996, 77, 346. 38 7. a) Rinehart, K.L.Jr.; Holt, T.G.; Fregeau, N.L.; Stroh, J.G.; Keifer, P. A.; sun, f.; Li , L. H.; Martin, D.G. / . Org. Chem. 1990, 55, 4512; b) Wright, A.E.; Forleo, D.A.; Gundawardana, G.P.; Gunasekera, S.P.; Koehn, F.E.; McConnell, O J . J. Org. Chem. 1990, 55, 4508. 8. Pettit, G.R. J. Nat. Prod. 1996, 59, 812-821. 9. 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, 6883-85. 10. Shu Y-Z. J. Nat. Prod. 1998, 61, 1053-71, and references therein. 11. Gibbs, J.B.; Oliff, A. Cell. 1994, 79, 193-198. 12. Omer, C.A.; Kohl, N.E. TIPS. 1997, 75, 437-444. 13. Levitzki, A.; Gazit, A. Science. 1995, 267, 1782-1787. 14. Nurse, P. Cell. 1997, 91, 865. 15. Borach, J.R.; Thorner, J. Nature. 1996, Vol 384, Supplement, page 14. 16. Crews C M , Collins JL, Lane WS, Snapper M L , Schreiber SL. J. Biol. Chem. 1994, 269, 411-20. 17. Sakai R, Rinehart K L , Kishore V, Kundu B, Faircloth G, Gloer JB, Carney JR, Namikoshi M , Sun F,Hughes RG, Gravalos DG, de Quesada TG, Wilson GR, Heid RM. JMed Chem. 1996, 39, 2819-22. 18. Pettit, G.R., etal, J. Am. Chem. Soc. 1982, 104, 6846-8. 19. Gunasekara, S.P., et al. J. Org. Chem. 1990, 55, 4912-5. 20. T.Lindel, P.R. Jensen, W. Frenical, B.H. Long, A. M . Casazza, J. Carboni, and R. R. Fairchild, J. Am. Chem. Soc. 1997, 119, 8744. 21. W. H. Gerwick, P. J. Proteau, D.G. Nagel, E. Hamel, A. Blokhin and D.L. Slate, J. Org. Chem. 1994, 59, 1243. 22. D.G. Nagel, R.S. Geralds, H.D.. Yoo, W.H. Gerwick, T-S. Kim, M . Namby, and J.D. White, Tetrahedron Lett. 1995, 36, 1189. 23. Look. S.A., et al, J. Org. Chem. 1986, 57, 5140-5. 24. Burgoyne, D.L., Andersen, R.J. & Allen, T.M. , J. Org. Chem., 1992, 57, 525-8. 39 25. Faulkner, D.J., Nat. Products Rep. 1999, 76, 155-198. 26. Wahl, M . , Mar. Ecol. Prog. Ser. , 1989, 55, 175-79. 27. Johnson, CR.; Sutton, DC. Mar Biol, 1994, 120, 305. 28. C.A. Bewley., D.J. Faulkner., Angew. Chem. Int. Ed., 1998, 37, 2162-78. 29. M.Roberge, R. G.S. Berlinck, Lin Xu, Hilary J. Anderson, et al., Cancer Research, 1998, 55, 5701-6; R.G.S. Berlinck, et al., J. Org. Chem., 1998, 63, 9850-6. 30. Kaufmann, WK. Paules RS. FASEB Journal. 1996, 10(2), 238-47. 31. Levine, AJ. Cell. 1997, 88(3), 323-331. 32. Sherr, CJ. Cell. 1994, 79(4), 551-555. 33. Sherr, CJ. Cell. 1993, 73(6), 1059-1065. 34. Savitsky, K.; Bar-Shira, A.; et al. Science. 1995, 268(5218), 1749-1753. 35. Lavin, MF.; Shiloh, Y. Annual Review of Immunology. 1997, 75, 177-202. 36. Shiloh, Y.; Rotman, G. Journal of Clinical Immunology. 1996, 16(5), 254-260. 37. Banin, S.; Moyal, L.; Shieh, S.; Taya, Y.; Anderson, CW., et al. Science. 1998, 281(5383), 1674-1677. 38. Waterman, MJ.; Stavridi, ES.; Waterman, JL.; Halazonetis, TD. Nature Genetics. 1998,19(2), 175-178. 39. Dulic, V.; Kaufmann, WK.; Wilson, SJ.; Tlsty, TD.; Lees, E.; Harper, JW.; Elledge, SJ.; Reed, SI. Cell. 1994, 7676'), 1013-1023. 40. Xiong, Y.; Hannon, GJ.; Zhang, H.; Casso, D.; Kobayaxhi, R..; Beach, D. Nature. 1993, 366(6456), 701-704. 41. Chen, P.; Gatei, M. ; O'Connell, M L ; Khanna, KK.; Bugg, SJ.; Hogg, A.; Scott, SP.; Hobson, K.; Lavin, MF. Oncogene. 1999, 18(1), 249-256. 42. O'Connell, MJ.; Raleigh, JM.; Verkade, HM.; Nurse, P. EMBO Journal. 1997,16(3), 545-554. 40 43. Peng, CY.; Graves, PR.; Thoma, RS.; Wu, Z.; Shaw, AS.; Piwnica-Worms, H. Science. 1997,277(5331), 1501-1505. 44. Dalai, SN.; Schweitzer, CM.; Gan, J.; DeCaprio, JA. Molecular & Cellular Biology. 1999, 79(6"), 4465-4479. 45. Jin, P.; Hardy, S.; Morgan ,DO. Journal of Cell Biology. 1998, 141(4), 875-885. 46. Al-Khodairy, F.; Carr, A M . EMBO Journal. 1992, 77(4), 1343-1350. 47. Rowley, R.; Subramani, S.; Young, PG. EMBO Journal. 1992, 11(4), 1335-1342. 48. Saka, Y.; Esashi, F.; Matsusaka, T.; Mochida, S.; Yanagida, M . Genes & Development. 1997, 11(24), 3387-3400. 49. Parker, L L . ; Piwnica-Worms, H. Science. 1992, 257(5078), 1955-1957. 50. Greenblatt, MS.; Bennett, WP.; Hollstein, M. ; Harris, CC. Cancer Research. 1994, 54(18), 4855-4878. 51. Barinaga, M . Science. 1997, 278(5340), 1036-1039. 52. Pennisi, E. Science. 1996, 274(5286), 342-343. 53. Powell, SN.; De-Frank, JS.; Connell, P.; Eogan, M.; Preffer, F.; Dombkowski, D.; Tang, W.; Friend, S. Cancer Research. 1995, 55(8), 1643-1648. 54. Fan, S.; Smith, ML. ; Rivet, DJ.; Duba, D.; Zhan, Q.; Kohn, KW.; Fornace, AJ.; O'Connor, PM. Cancer Research. 1995, 55(8), 1649-1654. 55. Wang, Q.; Fan, S.; Eastman, A.; Worland, PJ.; Sausville, EA.; O'Connor, PM. Journal of the National Cancer Institute. 1996, 88(14), 956-965. 56. Fuse, E.; Tanii, H.; Takai, K.; Asanome, K.; Kurata, N.; Kobayashi, H.; Kuwabara, T.; Kobayashi, S.; Sugiyama, Y. Cancer Research. 1999, 59(5), 1054-1060. 57. Akinaga, S.; Gomi, K.; Morimoto, M.; Tamaoki, T.; Okabe, M. Cancer Research. 1991,51(18), 4888-4892. 58. Arnaud, MJ. Progress in Drug Research. 1987, 31, 273-313. 59. Roberge, M. ; Tudan, C ; Hung, SM.; Harder, KW.; Jirik, FR.; Anderson, H. Cancer Research. 1994, 54(23), 6115-6121. 41 60. Anderson, H. J.; De Jong, G.; Vincent, I.; Roberge, M . Experimental Cell Research. 1 9 9 8 , 238(2), 498-502. 61. Gelfand, M. ; Mavi; M. ; Drummond, R.B.; Ndemera, B. 'The Traditional Medical Practitioner in Zimbabwe', Mambo Press, Gweru, 1 9 8 5 . 62. Wilson, L.; Jordan, M.A., Chem. Biol. 1 9 9 5 , 2, 569-573. 63. Mcintosh, J.R..; Mcdonald, K.L. Scientific American. 1 9 8 9 , 261, 48. 64. Rowinsky, E.K.; donehower, R.C.; Chabner, B.A..; Longo, D.L. Cancer Chemotherapy and biotherapy, pp. 263-296. Philadelphia: Lippincott Raven Publishers, 1 9 9 6 . 65. Rowinsky, E.K. Annu. Rev. Med. 1 9 9 7 , 48, 353-374. 66. Jordan, M.A.; Thrower, D.; Wilson, L. J. Cell. Sci. 1 9 9 2 , 102, 401-416. 67. Jordan, M.A.; Toso, R.J.; Thrower, D.; Wilson, L. Proc. Natl. Acad. Sci. USA. 1 9 9 3 , 90, 9552-9556. 68. Jordan, M.A., Wilson, L. Cell Biol. 1 9 9 8 , 10, 123-130. 69. Torres, K ; Horwitz, S.B. Cancer Res. 1 9 9 8 , 58, 3620-3626. 70. Jordan, A.; Hadfield, J.A.; Lawrence, N.J.; McGown, A.T. Med. Res. Rev. 1 9 9 8 , 18, 259-296. 71. Paull, K.d.; Lin, c.M.; Malspeis, L.; Hamel, E. Cancer Res. 1 9 9 2 , 52, 3892-3900. 72. Vincent, I.; Rosado, M. ; Davies, P. J. Cell Biol. 1996,132, 413-425. 73. Anderson, H.J.; deJong, G.; Vincent, I.; Roberge, M . Expt. Cell. Res. 1 9 9 8 , 238, 498-502. 74. Roberge, M. ; et. al; Andersen, R.J., Cancer Res. 1 9 9 8 , 58, 5701-5706. 75. Perlmann, H.; Perlmann, P.; Celis, J. E. Cell Biology - a practical approach, Vol.2, pp. 322-328. San Diego: Academic Press, 1 9 9 4 . 76. Schimenti, J.; Bucan, M . Genome Res. 1 9 9 8 , 8, 698-710. 77. Schreiber, S.L. Chem. Biol. 1 9 9 8 , 6, 1127-1152. 78. Fenteany, G.; Standaert, R.F.; Lane, W.S.; Choi, S.; Corey, E.J.; Schreiber, S.L. Science. 1995, 268, 726-731. 79. Mitchison, T.J. Chem. Biol. 1994, J, 3-6. 80. Pang, L.; Sawada, T.; Decker, S.J.; Saltiel, A.R. J. Biol. Chem. 1995, 270, 13585-88. 81. Liu, J.; Farmer, J.d.; Lane, W.S.; Friedman, J.; Weissman, I.; Schreiber, S.L. Cell, 1991,66, 807-815. 82. Panek, R.L.; etal.; Brown, K.J. J. Pharmacol. Exp. Titer. 1996, 286, 569-577. 83. Kohl, N.E.; et al.; Lee, T.J. Proc. Natl. Acad. Sci. USA. 1994, 91, 9141-9145. 84. Tantillo, C ; etal; Arnold, E. J. Mol. Biol. 1994, 243, 369-387. 85. Crabtree, G.R.; Schreiber, S.L. Trends Biochem. Sci. 1996, 21, 418-422. 86. Tian, S.S.; etal; Rosen, J. Science. 1998, 281, 257-259. 87. Rohrer, S.P.; etal; Schaeffer, J.M. Science. 1998, 282, 737-7'40. 88. Brown, E.J.; Schreiber, S.L. Cell. 1996, 86, 517-519. 89. Cimprich, K. A.; Shin, T.B.; Keith, C. T.; Schreiber, S.L. Proc. Natl. Acad. Sci. U.S.A. 1996, 93, 2850-2851. 90. Keith, C.T.; Schreiber, S.L. Science. 1995, 270, 50-55. 91. Cliby, W.A.; etal; Friend, S.H. EMBO. 1998,17, 159. 92. Hung, D.T.; Jamison, T.F.; Scheriber, S.L. Chem. Biol. 1996, 3, 623-639. 43 Chapter II Cembranolide diterpenoids 2.1 Cembranolide diterpenoids from the soft coral Pachyclavularia violacea. Specimens oi Pachyclavularia violacea (Quoy & Gaimard, 1833) were collected as part of a general collecting expedition to Papua New Guinea. The crude extracts of this sample showed G2 checkpoint inhibition activity and mild cytotoxicity against murine leukemia P388 in preliminary bioassay screening. Assay-guided fractionation using the new G2 checkpoint inhibition assay led to the isolation of the active components which turned out to be cembranolide diterpenoids. The isolation and structure elucidation of these secondary metabolites obtained from the crude extract of this species are described in the following chapter. 2.2 A brief review of diterpenoids related to cembranolides from soft corals. The diterpenoids are C20 compounds which, if carefully analyzed, can be recognized to consist of four C5 isoprene units linked in a repeating head-to-tail fashion. Biosynthetically, they arise from geranylgeranyl disphosphate (Figure 2.1). A wide variety of different cyclization, reduction, and oxidation transformations of this common precursor, leads to the great diversity of diterpenoids found in terrestrial and marine organisms. Diterpenoids exhibit varying degrees of biological activity.1'2 OPP OPP farnesyl PP (FPP) allylic cation IPP OPP geranylgeranyl PP Figure 2.1 Biosynthetic pathway to diterpenoids 43 44 Among abundant marine organisms, soft corals are one of the richer sources of diterpenoids. Soft coral diterpenoids represent many carbon skeletons, from the acyclic skeleton through monocyclic and bicyclic skeletons, to more complex cyclic systems. Cembranolide diterpenoids belong to the family of diterpenoids with monocyclic carbon skeletons. The monocyclic diterpenoids isolated from soft corals are an interesting group of compounds that possess 9-, 10-, 14- and 15-membered rings depending on the cyclization pathway involved in their biosynthesis (Figure 2.2)/ 10-2 Xenicin D ictyodial (9-membered ring) 10-1 D i I o p h o 1 (1 0-membered ring) 14-1 Cembranolides ^ (14-membered ring) 15-1 Flexibilene (1 5-membered ring) Figure 2.2 Cyclization pathway of geranylgeraniol Xenicin (20), the first octocoral diterpenoid with a 9-membered ring, was isolated from the Australian soft coral Xenia elongata.4 The xenicin skeleton can formally be produced via cyclization between C-2 and C-10 of the acyclic geranylgeranyl precursor. These types of 9-membered ring terpenoids are usually called xenia diterpenoids. Recently, the new xenia diterpenoids florlides A (21) and B (22),5 were isolated from the soft coral Xenia florida. They contain the typical xenia cyclononane skeleton but also have another unusual C19 and C7 linkage which forms a [4,3,1] bicyclic ring system. Florlide B showed antibacterial activity at 100 ug/disk against Staphylococcus aureus. 44 45 Dilophol (23) is a typical 10-membered octocoral diterpenoid.6 Soft corals in the genus Lobophytum usually contain many lobane moncyclic diterpenoids (i.e. 24-26) that may be generated by a Cope rearrangement of a diliphol-like precursor. The first lobane diterpenoids were isolated by the Australian group at the Roche Institute from an indigenous soft coral Lobophytum species." Since this initial discovery, the Lobophytum genus continues to provide new lobane diterpenoids. For example, the Indian Ocean soft coral L. hirsutum yielded 13-hydroxyloba-8,10,15,17-teraen-19-al (27) and two diastereomeric epoxides of 13,15-epoxyloba-8,10,16-trien-18-ol (28), while L. pauciflorum contained (13E,16E)-18-methoxyloba-8,10,13(15), 16-tetraene (29).78 45 Flexibilene (30), is an example of a diterpenoid with a 15-membered ring, isolated from the soft coral Sinulariaflexibilis9 It represents the first example of this unique diterpenoid skeleton. The most common monocyclic diterpenoids isolated from soft corals have the cembrane skeleton, which contains one 14-membered ring formed via cyclization of a geranylgeraniol-derived precursor between carbons 1 and 14. Soft corals abound in tropical waters such as the Caribbean and Indo-Pacific regions and are easily collected in shallow waters. Their extensive investigation has also been prompted by the high levels of biological activity found in the cembrane lactone derivatives (cembranolides), which are common metabolites of this group.' One of simplest cembrane derivatives is nephthenol (31) originally isolated from the soft coral Nephthea,i0 and subsequently found in Sarcophytum, Clavularia and Lobophytum species. Nephthenol (31) may be the biosynthetic precursor of many other cembranoids.2 Hawaiian, Indonesian, and Australian soft corals in the genus Sinularia have been the subject of numerous chemical studies. About 35 of the 90 known species in this genus have been chemically examined. From Indonesian collections of S. flexibilis, the Tursh group first described the structure of an interesting new lactone, sinulariolide"(32). It is 46 47 unique among the cembranolides in possessing a seven-membered ring lactone system, probably formed via a carboxylate attack on a CI 1-C12 epoxide precursor. Figure 2.3 The biscembranoid diterpenoid sinuflexlin (33) was isolated from the species S. flexibilis.12 It exhibited cytotoxicity against P388 murine leukemia cells in vitro with an E D 5 0 of 1.32 ug/ml. Hetero Diels-Alder coupling of two units and subsequent hydrolysis (Figure 2.3) may be the route for dimerization. C 0 9 C H , 33 The Hawaiian soft coral S. abrupta yielded an interesting furanocembranolide, pukalide (34). ' It contains a furan ring at C3-C6, and also an a,(3-unsaturated lactone bridging C10-C12, instead of the typical C1-C2 a,(3-unsaturated lactone. Two bis-pukalide diterpenoids, mayotolides A (35) and B (36), were isolated14 from the Indo-Pacific soft coral S. erecta. Esterification between the C18 carboxylic acid and a C7 hydroxyl group leads to the dimerization of two pukalide units. This dimerization is different from the formation of sinuflexlin (33). 47 48 36 R,=CH3, R2=OH Species of the soft coral genera Sarcophytum, and Cobophytum have also yielded cembranolide chemistry. The first structure reported from these genera was the epoxy cembranolide sarcophine (37) obtained from extracts of S. glaucuum collected in the Red Sea.15 Sarcophine (37) contains the typical oc,P-unsaturated lactone at CI and C2 as well as an epoxy ring at C7 and C8. Screening of cembranoids in therapeutically relevant assay systems has shown some to exhibit cytotoxic,16'17 cancer chemopreventative,18 and antiinflammatory19 potential. Recently, the discovery of the potent inhibition of Ras farnesyl transferase by a cembranolide (38) isolated from L. christagalli (IC 5 0 0.15|iM) has further enhanced the interest in this group of secondary metabolites.20 The terpenoid nature of this compound suggests that it might compete with farnesyl pyrophosphate (FPP) for binding to farnesyl protein transferase. Since FPP is a critical intermediate in the biosynthesis of a number of isoprene-derived cellular metabolites, an FPP competitive compound could impact on a number of metabolic pathways perhaps leading to cytotoxicity. The soft coral genera Clavularia and Pachyclavularia are also rich sources of cembranolide diterpenes. Both genera belong to the family Clavulariidae. Pachyclavulariadiol (39) and its naturally occurring mono-(40) and di-acetylated (41) derivatives have been isolated from the soft coral P. violacea." They contain a tri-48 49 substituted furan ring at C1-C2, and another tetrahydrofuran ring at the C9-C12 position. Pachyclavularolide (42) from the soft coral Clavularia violacea is related to pachyclavulariadiol (39) except that the furan ring has been oxidized to a butenolide.22 As part of our effort to find new G2 checkpoint inhibitors, we found that the extracts of P achy clavularia violacea gave positive response in the new G2 checkpoint inhibitor bioassay. Assay-guided isolation led to the discovery of six new cembranolide diterpenes. Their structures will be described in the following sections. O 41 R=Ac 49 51 2.3 Results and discussion: 2.3.1 Isolation of cembranolide diterpenes from Pachylclavularia violacea. Specimens of P. violacea were collected by hand using SCUBA on reefs at depths of 15 to 20m near Sek Point off Madang, Papua New Guinea in 1994. Freshly collected soft coral was frozen on site and transported to Vancouver over dry ice. The soft coral was identified by Dr. L .V. Ofwegen at the National Museum in the Netherlands. Frozen specimens of P. violacea (lOOg, wet) were thawed and extracted exhaustively with MeOH (500 ml x 3, 24hr. between extractions). The MeOH extract was filtered and concentrated in vacuo to give a deep green crude gum (5 g) which showed G2 checkpoint inhibition activity at 10 and 1 |lg/ml, and cytotoxic activity against murine leukemia P388 at 10 |i,g/ml. The crude extract was diluted with water up to 500 ml and partitioned sequentially against hexane (200 ml x 3), chloroform (200 ml x 3) and ethyl acetate (200 ml x 3). Only the hexane extracts (800 mg) showed both G2 and cytotoxic activities. Purification of the hexane extract was accomplished by repeated fractionation on Sephadex LH-20, silica gel flash chromatography, and reversed phase HPLC. Sephadex LH-20 size-exclusion chromatography (in MeOH) afforded one active fraction C (lOOmg) that was analyzed by 'H NMR and T L C . The 'H NMR data indicated that fraction C contained mainly terpenoids. A small portion of fraction C was fractionated by gradient silica gel flash chromatography and gradient reversed phase HPLC to get a preliminary idea of the chromatographic properties of the active components. This analysis showed that the active components eluted from the normal phase silica gel column with hexane/ ethyl acetate (4:1), and eluted from reversed phase HPLC with MeOH/H 2 0 (85:15). 51 52 Guided by the above analytical results, a purification.method for the active fraction C was established. In the first step, fraction C was chromatographed on silica gel using a gradient elution (hexane/ethyl acetate 1:9 to 1:1) to give fraction L (20 mg) which eluted with hexane/ethyl acetate (2:8). Secondly, fraction L was chromatographed on reversed phase HPLC with MeOH/H 2 0 (85:15) to give one bioactive diterpene F (5.2mg) and five inactive diterpenes A (6.2 mg), B (3.1 mg), and C (0.8 mg), D (0.9 mg), E (1.0 mg), G (0.8 mg). 2.3.2 Structure elucidation of cembranolides from Pachylclavularia violacea. The structures of pachyclavulariolides A (43), B (44), C (45), D (46), E (47), F (18) and G (48) were solved by extensive analysis of ID and 2D NMR, and mass spectrometric data. Proton-carbon attachments were determined by HMQC experiments and proton spin systems were identified from COSY data. H M B C data also proved useful for the identification of all connectivity within the molecules. The assignment of the quaternary carbons was mainly dependent on HMBC and DEPT results. Relative stereochemistry was determined using NOE difference and X-ray diffraction data. 52 Pachyclavulariolide A (43), R = H Pachyclavulariolide B (44), R = OH Pachyclavulariolide C (45), R = OMe 21 Pachyclavulariolide F (18) 0 Pachyclavulariolide G (48) 53 54 Figure 2.3 Isolation of pachyclavulariolide diterpenoids from Pachyclavularia violacea. MeOH extract 1. Concentrated in vacuo 2. Diluted with water 3. Partitioned between aq. and org. solvents Hexane fraction CHC1 3 fraction EtOAc fraction Water fraction Sephadex LH-20 Eluent: MeOH A B C D Silica gel Eluent: Hex/EtOAc (1:9 to 1:1) < 1 K L M N 0 R.P. HPLC Eluent: MeOH/H 2 0 1 1 r | Pachyclavulariolide F Pachyclavulariolide G Pachyclavulariolide E Pachyclavulariolides A-D 54 55 Pachyclavulariolide A Pachyclavulariolide A (43) was isolated as a white amorphous solid. It gave a very intense [M+H]+ peak at m/z 317.21150 in the positive ion HRFABMS corresponding to a molecular formula of C 2 oH 280 3 (AM, -0.53ppm), requiring seven sites of unsaturation (Figure 2.4). Table 2.1 provides a summary of the NMR data acquired for 43 at the elevated temperature 328 ± 2 K (55 ± 2 °C). The first 'H NMR spectrum of pachyclavulariolide A acquired in CD 2 C1 2 at room temperature showed very broad signals for each peak. The peak broadening indicated the molecule probably existed in several different interconverting conformations at room temperature. In an attempt to slow the conformational interconversions the 'H and 1 3 C NMR spectra were reacquired over the temperature range from 20°C to -20°C. The 'H NMR signals appeared broad until the temperature reached -10°C, which was the molecular coalescence temperature. When temperature went down to -20°C, the broad peaks gradually separated into two distinct sets of sharper resonances. I 3 C NMR data was also recorded at -20°C (Figure 2.5). The number of carbon resonances observed was almost twice the number of carbons indicated by the HRFABMS analysis. Each carbon resonance was paired with a nearby resonance or else it appeared that two resonances 55 S3 o > O CO o II a> IS "OJ o o _ x < S ID ™ Cp S T l - « CM t » -—r~-cvi _^ o> _j CO CO 5 CO rr> 8 + at CM -o • LO CO o • o CO o • o co o o - i n eo-*»8-CM O • O CM O LO 8-8 1 -5?-o - o co o • LO LO o • o LO o • LO o o o o •a 3 > >-. JS u «J o OH oo < PL, c "o 00 o J rr r4 a> t-S M l s o 57 1 7 0 160 1 5 0 1 4 0 1 3 0 1 2 0 110 1 OO 9 0 8 0 7 0 60 5 0 4 0 3 0 2D (ppm) Figure 2.5: 'H NMR data (500 MHz) for 43 in CD 2 C1 2 at the different temperatures. 1 3 C NMR data (125 MHz) for 43 in CD 2 C1 2 at -20°C. 57 58 60 C J V 4.8 4 . 4 4 . 0 3 .6 3 .2 2 .8 2 .4 (ppm) 2 . 0 1.6 1.2 . 0 .8 55 C L A J 5.0 4 . H 4 .6 4 .4 4 . 2 4 .0 3.8 3 .6 3.4 3.2 3 .0 2 .8 2.f. 2 . 4 2 .2 2 .0 l . H 1.6 1.4 1.2 | l . O 0.8 0.( (ppm) 45 C 4 . 8 4 . 4 4 .0 3.6 3 .2 2 .8 2 .4 (ppm) 2 . 0 1.6 1.2 0.8 I 35 C 4.4 4 .0 3.6 3 .2 2 .8 2 . 4 2 . 0 (ppm) 1.6 1.2 O.fci Figure 2.6 J H NMR data (500 MHz) for 43 in C 6 D 6 at different temperatures. 58 59 overlapped each other. The low temperature NMR data clearly indicated that pachyclavulariolide A occupied two conformations at -20°C. To solve the structure of 43, NMR experiments were performed at an elevated temperature so that the average resonance of the two conformations would give one single peak. When the temperature incr eased from 20°C to 55°C, the 'H NMR resonances gradually turned sharper. At the even higher temperature 60°C, the 'H NMR spectrum did not change further (Figure 2.6). Therefore one dimensional 'H, 1 3 C and two dimensional COSY, HMQC, H M B C NMR data were collected at 55°C. Several unsaturated functional groups were evident from the NMR data. These included an a, (3 - unsaturated ester (5 173.0 ppm /CI6, 163.6/C1, 123.2/C15) and one zra/w-trisubstituted double bond (8 129.1/C4, 129.6/C5, 18.5/C18). Consequently, it was apparent that 43 was a tetracyclic structure. Moreover, the single remaining oxygen atom and two oxygen-bearing carbons (87.0/C9, 86.2/C12) had to be tied together in a heterocyclic ring. HMQC correlations identified the chemical shifts of the carbon resonances and their attached protons, and also indicated two aliphatic quaternary carbons (C8 (5 49.4) and C12 (5 86.2)). Furthermore, the COSY NMR spectra of this compound established the connectivity beginning at C2 and extending to C7, including CI8 (Figure. 2.7). Figure 2.7: The fragment of C2 to C7 with selected COSY correlations. 59 60 H M B C NMR spectra of pachyclavulariolide A (43) showed correlations between C7 and H19, C8 and H19, C9 and H19, C13 and H19 (Figure 2.8). Therefore, the aliphatic quaternary carbon C8 was found to connect C7, C9, C13 and Me-19. Further COSY correlations between HI3 and HI4, and HMBC correlations between HI3 /H14 and CI, and between H14 and CI / C2 (Figure 2.9) confirmed one 10-membered ring fromCl to C8, C13 and C14. The Me20 resonance (6 1.05) showed H M B C correlations to a resonance at 5 86.2 assigned to an oxygenated quaternary carbon (C12), to a methine resonance at 5 54.6 (C13) and to a methylene resonance at 5 38.8 (CI 1). These correlations indicated that the oxygenated quaternary carbon C12 was connected to CI 1, C13, C20. COSY correlations from H9 through H10 to HI 1 established the C9-C10-C11 connectivity. HMBC correlations were also observed between H9 and C12. Together this data led to the elucidation of bicyclo [2,2,0] fragment with an oxygen bridge (Figure 2.8). 1.92 Figure 2.8:The [2,2,0] cyclic core structure with selected HMBC and COSY correlations. H M B C correlations were observed between C H 3 (17) and CI, C15, and C6. This set of H M B C correlations confirmed the 5-membered a,[3-unsaturated lactone (C2-C1-C15-C16) with a single methylation at the (3 (C15) position (Figure 2.9). 1.68 60 61 Figure 2.9: The trisubstituted butenolide with selected H M B C correlations. Finally, the whole skeleton was assembled. It contained one trisubstituted butenolide and one typical 14-membered ring. One unsaturated double bond was positioned at C4 and C5. A further carbon connectivity between C8 and C13 was formed and one oxygen linked C9 to C12. The final structure was a diterpenoid with a briarane IP carbon skeleton. ~ The configuration of the A 4 ' 5 olefin and the relative stereochemistry of the chiral centers in 43 were determined via a series of difference NOE experiments. Irradiation of the Mel8 resonance at 8 1.35 induced a NOE in the H6' resonance at 8 1.89 demonstrating that the A 4 ' 5 olefin had the E configuration. When the Mel9 resonance at 8 0.76 was irradiated, strong NOEs were observed in the H14' (8 1.92) and H9 (8 3.52) resonances, but no NOEs were observed in the HI 3 (8 1.18), HI0 (8 1.50), or 10' (8 1.68) resonance. These observations were consistent with Me 19 and C14 being exo 61 62 substituents on the oxanorbornane fragment. Finally, irradiation of the H2 resonance at 8 4.65 induced a strong NOE in the H13 resonance at 8 1.18 indicating that H2 and H13 were on the same face of the 10-member ring, completing the relative stereochemical assignment of pachyclavulariolide A (43) as shown in F igure 2.10. Molecular modelling23 showed that the C3-C7 region of the 10-membered core ring would adopt two relatively strain free conformations (Figure 2.10). The interconversion of these two conformations occurred slower than the NMR time scale at room temperature. Therefore, the 'H NMR spectra showed many broad peaks. When the temperature was lowered to -20°C, the interconversion of the two conformations was slower and each gave a distinct 'H NMR spectrum. When the temperature was raised to 55°C, the interchange of the two conformations occurred fast relative to the NMR time scale, resulting in a single 'H NMR spectrum, reflecting the average environment. Figure 2.10: Selected NOE correlations (left) and conformation models (right). 62 63 Table 2.1. NMR data for pachyclavulariolide A recorded in benzene-^ at 500MHz ('H). Temp:328 ±2 K. c# 8 l 3 C(ppm) 5 'H(ppm) a C O S Y Correlation b H M B C Correlation "NOE-D1FF 1 163.6 — . . . H13,H14,H14 ' ,H17 2 78.6 4.65(dd,l 1.0, 2.06Hz) H 3 , H 3 ' H14 H3,H5,H13 3 42.5 2.75(dd,l 1.0, 1.78Hz) H 2 , H 3 ' H5,H18 H 2 , H 3 ' 3' 1.52 H 2 . H 3 4 129.1 — — H18 5 129.6 5.04(m) H 6 , H 6 ' , H 1 8 H18 H2,H13 6 24.3 1.92 H 5 , H 6 ' , H 7 , H 7 ' H5 ,H7 6' 1.89 H 5 , H 6 , H 7 , H 7 ' 7 37.7 1.22 H 6 , H 6 ' , H 7 ' H13 ,H19 7' 1.05 H 6 , H 6 ' , H 7 8 49.4 — — H6,H10 ,H7 ,H19 9 87.0 3.52(d,5.10Hz) H10 H I 0 , H 1 9 H19,H10 10 26.1 1.50 H 9 , H 1 0 ' , H 1 1 , H 1 1 ' H9 10' 1.68 H10.H11,H1 r 11 38.8 1.32 H 1 0 , H 1 0 ' H H ' H9 ,H10 ,H13 ,H20 i r 1.12 H10,H10 ' ,H11 12 86.2 — H9,H13 ,H14 ,H14 ' ,H20 13 54.6 1.18 H14.H14 ' H9 ,H14 ,H19 ,H20 14 26.5 2.22(dd,9.98, 15.76Hz) H13 ,H14 ' H13 14' 1.92 H13,H14 15 123.2 — H14,H14 ' ,H17 16 173.0 — . . . H17 17 9.1 1.65 — . . . 18 18.5 1.35 H5 H5 H2 ,H3 ,H6 19 20.5 0.76 — H9 H5 ,H9 ,H14 ' 20 18.5 1.05 — — H I 1, H14,H17 a. Correlated to proton resonance in S ' H column. b. Correlated to carbon resonance in 8 I 3 C column. 63 (ppm) 4.8 4.0 3.2 2.4 1.6 0.8 Figure 2.12: 2D COSY spectrum of pachyclavulariolide A (43) in C 6 D 6 at 328 ± 2 K. 66 . A r \ (ppm) ( p p m ) 4.8 4.0 3.2 2.4 20 40 60 80 100 120 1.6 0.8 Figure 2.13: 2D H M Q C spectrum of pachyclavulariolide A (43) in C 6 D 6 at 328 ± 2 K. 66 67 Pachyclavulariolide B Pachyclavulariolide B (44) was obtained as a white solid that gave a [M+H]+ peak (Figure 2.14) in the HRFABMS at m/z 333.20663 appropriate for a molecular formula of C 2oH 280 4 (AM 0.15 ppm). This molecular formula differed from that of pachyclavulariolide A (43) by the addition of one oxygen atom. *H NMR spectra obtained for pachyclavulariolide B (44) at room temperature also showed very broad peaks and it indicated that two conformations existed like pachyclavulariolide A (43). To solve the structure, high temperature NMR experiments were performed. At or above 55°C, the *H NMR spectrum for pachyclavulariolide B (44) turned sharper and the average resonances of two conformations gave single peaks (Figure 2.15). The H and ~ C NMR spectra obtained for pachyclavulariolide B (44) were similar to those of pachyclavulariolide A (43) collected at 55°C. However, the H2 resonance in the ' H NMR spectrum of pachyclavulariolide A was conspicuously absent in the spectrum of pachyclavulariolide B. Additionally, the l 3 C NMR spectrum of pachyclavulariolide B had one carbon resonance at 106.5 ppm corresponding to a hemiacetal carbon. One exchangeable proton resonance existed in the 'H NMR spectrum. It was apparent that the additional oxygen was attached to C2 as part of a hydroxyl group. A detailed analysis of the COSY, HMQC, and HMBC data obtained for pachyclavulariolide B confirmed the structure. HMQC correlations identified the 67 00 69 25'C 70 chemical shifts of the carbon resonances assigned to C3 (5 46.2), C5 (S 130.1), C6 (5 24.3), C7 (8 37.2), C18 (§ 19.5) and their attached protons. An examination of the COSY and HMBC data established the connectivity from CI (8 160.6) to C7. COSY correlations were observed between H3 (8 2.79) and H3'(5 1.95), H5 (5 5.47) and H6/H6'(8 2.03/1.96), H6 and H7/H7'(8 1.10/1.12). HMBC correlations were observed between 2-OH and C3, H3 and CI, HI8 and C3/C4/C5 (Figure 2.16). Therefore, connectivity from CI to C7 was constructed. The l 3 C chemical shift of C18 (8 19.5) indicated a trans-trisubstituted double bond between the C4 (8 128.3) and C5 (8 130.1) olefinic carbons. An a,(3-unsaturated lactone was confirmed by the chemical shifts of C16 (8 170.5), C15 (8 124.6), CI (8 160.6) and C2 (8 106.5). H M B C correlations were observed between CH 3 ( 17) and CI/C15/C16. They verified a 5-membered a,(3-unsaturated lactone (C1,C2,C15,C16) methylated at the (3(15) position (Figure 2.17). Therefore, pachyclavulariolide B also contained a trisubstituted butenolide. Figure 2.16: The fragment structure of 44 with selected COSY and H M B C correlations. The remaining fragment of 44 was elucidated to have the same [2,2,0] cyclic core structure as that of pachyclavulariolide A (43) by following the same protocol used for 43. C7 was attached to the core structure of 44 at C8, and the C14 methylene appendage of core structure was connected to the trisubstituted butenolide at CI. Therefore, the 70 71 overall framework was the same as that of pachyclavulariolide A (43). The only difference in structure is that pachyclavulariolide B (44) has a hydroxyl group attached to C2 instead of a hydrogen. The presence of this hydroxyl group was also confirmed by mass spectrometric analysis. FABMS peaks (Figure 2.14) at m/z 315.19610 (C20H97O3) could be assigned to a fragment which has lost a molecule of water from the protonated molecular ion to form a stable cation at C2. This cation was easily stabilized by electron resonance effects with the a,(3-unsaturated lactone. So this fragment indicated the molecule contained a hydroxyl group. Figure 2.17: Selected HMBC correlation in pachyclavulariolide B (44). Initial attempts to determine the relative stereochemistry of pachyclavulariolide B (44) involved NOE difference experiments. When the NOE experiments were performed in benzene-d6 at 55°C overnight, the solvent gradually evaporated and crystals formed in the NMR test tube. As a result, the NOE experiments failed because the concentration of the sample in the NMR tube varied throughput the data collection. Fortunately, one single crystal was of sufficient quality for single crystal X-ray diffraction analysis. The X-ray diffraction analysis24 elucidated the relative stereochemistry for each chiral center in 10 71 72 pachyclavulariolide B and also confirmed the structure of the carbon skeleton of the molecule (Figure 2.19). ' Table 2.2 NMR data for pachyclavulariolide B recorded in benzene-<i6 at 500 MHz('H). Temp: 328 ± 2 K. c# 5 1 3 C(ppm) 8 'H(ppm) a C O S Y Correlation b H M B C Correlation ' ' N O E - D I F F 1 160.6 . . . — H3,H13 ,H14 ,H14 ' ,H17 2 106.5 . . . . . . H3,H14 3 46.2 2.79 H 3 ' H 1 8 , 2 - O H 3' 1.95 H3 4 128.3 . . . — H18 5 130.1 5.47(b) H 6 , H 6 ' H3,H18 6 24.3 2.03 H 5 , H 6 \ H 7 , H 7 ' H 7 6' 1.96 H 5 , H 6 , H 7 , H 7 ' 7 37.2 1.40 H 6 , H 6 ' , H 7 ' H19 T 1.12 H 6 , H 6 ' , H 7 8 48.9 . . . — H7,H10 ,H10 ' ,H14 ,H19 9 86.7 3.57(d,5.04Hz) H10 H10.H19 10 25.8 1.55 H 9 , H 1 0 ' , H 1 I H 9 , H ! 1 10'. 1.75 H10,H11 i i / i r 38.1 1.40 H10 ,H10 ' H 9 , H 1 0 , H 1 0 \ H 2 0 12 86.1 . . . H 9 , H 1 0 ' , H 1 1 , H 1 3 , H 1 4 , H 14',H20 13 49.9 2.24 H14,H14 ' H 9 . H 1 1 , H 1 4 , H 1 4 , , H 1 9 , H 20 14 26.8 2.28 H13,H14 ' H13 14' 2.06 H13.H14 15 124.6 . . . — H14.H17 16 170.5 . . . — H17 17 8.1 1.59 H14 ' — 18 19.5 1.51 — 19 20.0 0.82 . . . H 9 H6,H9,H14 20 18.1 1.18 . . . H I 1.H13.H17 2 - O H — 2.82 . . . — a. Correlated to proton resonance in 8 'H column. b. Correlated to carbon resonance in 5 I 3 C column. H O 72 o CM LO 2 a +1 OO CN co <S Q U PQ o T3 =3 > 'o o <+-. O E o N) o o oo <u i n a 0* 1 CN i t CM s I I I 4 LO CO o r-s o. o - o o o - T | w CN +1 OO CN m Q U pq 3 > o >^  _C CJ ca a, <+-o £ o N IT) CN U 00 CN -3 OX) 74 75 ~2> (ppm) J'*> A. ^ S " SB P (ppm) 5.6 4.8 4.0 3.2 2.4 1.6 0.8 1.6 2.4 3.2 4.0 4.8 0.8 Figure 2.20: 2D COSY spectrum of pachyclavulariolide B (44) in C 6 D 6 at 328 ± 2 K. 75 76 J L ( p p m ) ( p p m ) 5.6 4.8 4.0 3.2 *8> H5° 2.4 1.6 0.8 Figure 2.21: 2D H M Q C spectrum of pachyclavulariolide B (44) in C 6 D 6 at 328 ± 2 K. 76 77 Pachyclavulariolide C Pachyclavulariolide C (45) was obtained as a white solid that gave a [ M +H] + peak in the HRFABMS at m/z 347.22162 (Figure 2.22) appropriate for a molecular formula of C 2 1 H 3 0 O 4 (AM -1.77ppm). This molecular formula differed from that of pachyclavulariolide A (43) by the addition of C H 2 0 . 'H NMR spectra obtained for pachyclavulariolide C at room temperature showed very broad peaks indicating that pachyclavulariolide C was present as two conformations, like pachyclavulariolide A. High temperature of NMR experiments were performed to solve the structure. At 55°C, the ' H NMR spectrum turned sharp and the average resonance of two conformations gave one single peak (Figure 2.23). The ' H and 1 3 C NMR spectra obtained for pachyclavulariolide C (45) were similar to those of pachyclavulariolide A (43) collected at 55°C. However, the H2 resonance in the 'H NMR spectrum of pachyclavulariolide A (43) was conspicuously absent in the spectrum of pachyclavulariolide C (45). Additionally, the l 3 C NMR spectrum of pachyclavulariolide C had a carbon resonance at 106.5 ppm corresponding to an ketal carbon. Furthermore, the *H NMR spectrum of pachyclavulariolide C contained a methylene resonance at 8 2.91 ppm, which was assigned to a OMe group. Therefore, it was apparent that the additional elements CH9O were attached to the C2 position as a methoxyl group. 77 79 55 C X 5.6 5.2 4.8 4.4 4.0 3.6 3.2 (ppm) 50 C 2.8 2.-1 2.0 1.6 1.2 0.8 45 C J L 2.* 2.0 5.6 5.2 4.8 4.4 4.0 3.6 3.2 (ppm) 2.8 2.4 2.0 1.6 1.2 0.8 40 C 5.6 5.2 4.8 4.4 4.0 3.6 3.2 (ppm) 2.8 2.4 2.0 1.6 1.2. 0.8 35 C 5.6 5.2 4.8 4.4 4.0 3.6 3.2 (ppm) 2.8 2.4 2.0 1.6 l]2 0.8 25 C JL AAMJ 5.6 5.2 4.8 4.4 4.0 3.6 3.2 • (ppm) 2.8 2.4 2.0 1.6 1.2 Figure 2.23: ] H NMR data(500 MHz) for 45 in C 6 D 6 at the different temperatures. 79 80 A detailed analysis of the COSY, HMQC and HMBC data obtained for pachyclavulariolide C confirmed the structure. HMQC correlations identified the chemical shifts of the carbon resonances assigned to C3 (5 44.0), C5 (5 130.3), C6 (5 24.6), C7 (5 38.3), C18 (5 20.0) and their attached protons. An analysis of the COSY and H M B C spectroscopic data established the connectivity from C2 to C7. COSY correlations were observed between H3 (8 3.08) and H3'(S 1-75), H5 (5 5.63) and H6/H6'(5 1.98/2.01), H6 and H7/H7'(5 1.35/1.15). HMBC correlations were observed between 2-OCH-, (S 2.91) and C2 (8 108.8), H3 (8 3.08) and C2, HI8 (8 1.50) and C3/C4/C5. Therefore, the connectivity from C2 to C7 was assembled (Figure 2.24). The l 3 C chemical shift of C18 (8 20.0) indicated a rrans-trisubstituted double bond between the C4 (8 127.2) and C5 (8 130.3) olefinic carbons. An a,(3-unsaturated lactone was confirmed by the chemical shifts of C16 (8 169.7), C15 (8 127.0), CI (8 158.0) and C2. HMBC correlations were observed between the CH3(17) resonance and the C1/C15/C16 resonances (Figure 2.25). They verified a 5-membered a,(3-unsaturated lactone (C1,C2,C15,C16) methylated at the (3(C15) position. Therefore, pachyclavulariolide C (48) also contained a trisubstituted butenolide. Figure 2.24: The fragment structure of 45 with selected COSY and H M B C correlations. 80 81 The remaining fragment of 45 was elucidated to have the same [2,2,0] bicyclic core structure as that of pachyclavulariolide A (43) by following the same protocol used for 43. The C7 was attached to the core structure of 45 at C8, and the C14 methylene appendage of core structure was connected to the trisubstituted butenolide at CI. Figure 2.25: Selected HMBC correlation in pachyclavulariolide C (45). Therefore, the overall framework was the same as that of pachyclavulariolide A (43), and the only difference in the structures is that pachyclavulariolide C (45) has a methoxyl group attached to C2 instead of a hydrogen. The presence of the methoxyl group was also confirmed by mass spectrometric analysis. A LRFABMS peak at m/z 315 (C20H97O3) could be assigned to a fragment ion which has lost methanol from the protonated molecular ion to form a stable cation.(Figure 2.26) This cation was at the C2 position and easily stabilized by resonance effects. The observation of this fragment ion in the MS supported the presence of the methoxyl group at C2. 10 81 82 H o Figure 2.26: The fragment ion was stabilized by resonance. The relative stereochemistry in pachyclavulariolide C (45) was based on the results of NOE difference experiments and the known stereochemistry of pachyclavulariolide B (44). Pachyclavulariolide C (45) has the same bicyclic [2,2,0] substructure as pachyclavulariolide B (44), so it was assumed that they possess the same relative stereochemistry in that fragment. Irradiation at H19 (S 0.82) gave NOEs in H9 (5 3.54) and H14 (5 2.26) confirming that conclusion. The O C H 3 was assigned to the (3 orientation. This was based on a NOE observed between H21 and H13, confirming the methoxyl group was oriented on the same side of the molecule as HI 3 (Fig. 2.27). The remaining stereochemistry of pachyclavulariolide C (45) was the same as that of pachyclavulariolide B (44). 10 O 17 Figure 2.27: Selected NOE correlations in pachyclavulariolide C (45). 82 83 Table 2.3. NMR data for pachyclavulariolide C recorded in benzene-fife at 500 MHz (*H). Temp: 328 + 2 K. c# 5 1 3 C(ppm) 5 'H(ppm) " C O S Y Correlation b H M B C Correlation a N O E - D I F F 1 158.0 . . . H14,H14 ' ,H17 2 108.8 — . . . H3,H14,H21 3 44.0 3.08(d,15.3Hz) H 3 ' H18 H 3 \ H 2 1 3' 1.75 H3 4 127.2 — . . . H 3 ' , H 6 ' , H 1 8 5 130.3 5.63(b) H 6 , H 6 ' H18 6 24.6 1.98 H 5 , H 7 , H 7 ' H 7 6' 2.01 H 5 , H 7 , H 7 ' 7 38.3 1.35 H 6 , H 6 ' , H 7 ' H6 ,H9 ,H13 ,H19 7' 1.15 H 6 , H 6 ' , H 7 8 49.3 — . . . H6,H10 ,H13 ,H19 9 86.4 3.54(d,5.28Hz) H10 H10.H19 10 25.7 1.53 H l O ' . H l l . H l l ' H9.H1 l . H l 1' 10' 1.75 H l O . H l l . H l l ' 11 38.3 1.35 H10 .H10 ' H9 ,H13 ,H20 i r 1.32 H10 .H10 ' 12 86.1 — — H9.H1 r , H 1 4 , H 1 4 ' , H 2 0 13 49.9 1.98 H14.H14 ' H9 ,H13,H14,H20 14 27.1 2.26(dd,9.79, 16.1 Hz) H13 .H14 ' H13 14' 2.02 H13.H14 15 127.0 — — H14.H17 16 169.7 — . . . H17 17 8.4 1.63 H 1 4 ' . . . H14,H20,H21 18 20.0 1.50 — . . . H3 ' ,H6 ,H6 ' ,H21 19 19.9 0.82 . . . H9 H9,H14 ,H14 ' 20 18.5 1.18 . — . . . H I l ' , H 1 4 , H 1 7 , H 2 l 21 49.5 2.91 . . . . . . H13,H17,H20 a. Correlated to proton resonance in 8 ' H column. b. Correlated to carbon resonance in 8 I 3 C column. 83 oo 85 86 87 Pachyclavulariolide D Pachyclavulariolide D (46) was obtained as a white solid that gave a [M+H]+ peak in the HRFABMS (Figure 2.31) at m/z 347.22079 appropriate for a molecular formula of C 2 1 H 3 0 O 4 (AM -4.17ppm), identical to the molecular formula of pachyclavulariolide C (45). However, the 'H NMR spectra obtained for pachyclavulariolide D (46) at room temperature showed sharp highly resolved resonances. This suggested that pachyclavulariolide D was not undergoing slow conformational change at room temperature. The room temperature 'H and l 3 C NMR spectra obtained for pachyclavulariolide D (46) were very similar to those of pachyclavulario lide C (45) obtained at 55°C. The l 3 C NMR spectrum of pachyclavulariolide D (46) had a carbon resonance at 5 106.5 ppm (C2) corresponding to an acetal carbon, and 'H NMR spectrum contained a methyl resonance at 5 2.81 ppm assigned to a methoxyl group. The similarity of the acetal and methoxyl resonance in the spectra of pachyclavulariolide D (46) to their counterparts in the spectra of 45 indicated the presence of a -OCH3 group at C2 in 46. A detailed analysis of the COSY, HMQC and HMBC data obtained for pachyclavulariolide D (46) established the constitution of 46. H M Q C correlations identified the chemical shifts of the carbon resonances assigned to C3 (8 52.2), C5 87 oo 0 0 89 (5 131.0), C6 (5 24.9), C7 (5 37.8), C18 (5 17.7) and their attached protons. An examination of COSY and H M B C spectroscopic data established the connectivity from C2 to C7. COSY correlations were observed between H3 (5 2.90) and H3' (5 2.10), H5 (8 4.52) and H6/H6' (8 1.69/2.00), H6 and H7/H7' (8 1.11). HMBC correlations were observed between 2-OCH 3 (8 2.81) and C2 (8 112.2), H3 and C2, HI8 (8 17.7) and C3/C4/C5. Therefore, the connectivity from C2 to C7 was assembled (Figure 2.32). Figure 2.32: The fragment structure of 46 with selected COSY and HMBC correlations. The l 3 C chemical shift of CI 8 (8 20.0) indicated the presence of a trans-trisubstituted double bond between the C4 (8 127.2) and C5 (8 130.3) olefinic carbons. An a,(3-unsaturated lactone was suggested by the chemical shifts of C16 (8 169.7), CI 5 (8 127.0), CI (8 158.0) and C2 (8 108.8). HMBC correlations were observed between CH3 (17) and C1/C15/C16 (Figure 2.33). They verified a 5-membered a,(3-unsaturated lactone (CI, C2, CI 5, CI6) methylated at the (3(15) position. Therefore, pachyclavulariolide D also contained a trisubstituted butenolide. The remaining fragment of 46 was elucidated to have the same [2,2,0] cyclic core structure as that of pachyclavulariolide A (43) by following the same protocol used for 43. The C7 was attached to the core structure of 46 at C8, and the C14 methylene appendage of core structure was connected to the trisubstituted butenolide at CI. 8 9 Figure 2.33: Selected HMBC correlation in pachyclavulariolide D (46). Therefore, the overall framework was the same as that of pachyclavulariolide A (43). The only difference in structure is that pachyclavulariolide D has a methoxyl group attached to C2 instead of a hydrogen atom. A LRFABMS peak (Figure 2.31) at m/z 315 (C20H27O3), that could be assigned to a fragment which had lost one equivalent of methanol from the protonated molecular ion to form a stable cation, confirmed the presence of the methoxy group. The relative stereochemistry of pachyclavulariolide D (46) was assigned by NOE difference experiments and by reference to the known stereochemistry of pachyclavulariolide B (44). Pachyclavulariolide D (46) has the same bicyclic [2,2,0] ring system as pachyclavulariolide B, so it was assumed that they possess the same stereochemistry in the bicyclic core. NOEs observed between H19 (5 0.82) and H9 (5 3.46)/H14a (5 2.05) confirmed that conclusion. The OCH3 group was assigned to the a position. This assignment was based on the observation of a NOE between H21 (5 2.81) and HI9 (5 0.82), HI4a (8 2.05). The remaining stereochemistry of pachyclavulariolide 90 91 D ( 4 6 ) was the same as that of pachyclavulariolide B ( 4 4 ) . Therefore, pachyclavulariolide D was an epimer of pachyclavulariolide C ( 4 5 ) . The stereochemical assignment was consistent with the fact that pachyclavulariolide D ( 4 6 ) did not show slow intercoversion between two conformations. The a orientation of the OCH3 group relative to the 14-membered ring prevents the C3-C7 chain from easily rotating. Therefore, the C3-C7 chain remains locked in one conformation. The olefinic proton H5 (6 4.55) appears to be shielded by its proximity to the CI-CI 5 olefin. This was evidence that the C3-C7 side chain was oriented above the a,(3-unsaturated lactone ring. 91 92 Table 2 . 4 . NMR data for pachyclavulariolide D recorded in benzene-^ at 500 MHz (!H). Temp: 298 ± 2 K. c# 5 1 3 C(ppm) 5 'H(ppm) a C O S Y Correlation b H M B C Correlation " N O E - D I F F 1 161.7 — . . . H 3 , H 1 3 , H 1 4 , H 1 4 ' 2 112.2 . . . . . . H 3 , H 3 \ H 1 4 , H 2 1 3 52.2 2.90(d, 13.5Hz) H3' H18 H3' ,H18 3' 2.10(d,13.5Hz) H3 4 130.0 . . . . . . H 3 , H 3 \ H 1 8 5 131.0 4.52(b) H 6 , H 6 ' H3 ,H3' ,H18 H3' ,H13,H17 6 24.9 1.69 H 5 , H 6 ' , H 7 . . . 6' 2.00 H 5 , H 6 , H 7 111' 37.8 1.11 H 6 , H 6 ' H13.H19 8 50.7 — — H13,H14,H19 9 85.9 3.46(d,4.72Hz) H10 H19 H19 10 25.4 1.55 H 9 , H 1 0 \ H 1 1 HI 1 10' 1.65 H l O . H l l 11/1T 39.5 1.36 H10.H10' H9 ,H20 12 87.3 — . . . H9,H14 ,H14 ' ,H20 13 53.7 1.25(b) H14.H14' H9,H14,H14' ,H19 ,H20 14 23.7 2.05 H13.H14' H13 14' 1.87 H13.H14 15 127.2 — . . . H14.H17 16 170.8 — . . . H17 17 10.5 1.70 — . . . 18 17.7 1.70 — H 3 . H 3 ' 19 19.7 0.82 — H9,H13 H9.H14 20 20.6 1.35 — . . . . . . 21 51.1 2.81 . . . . . . H14 a. Correlated to proton resonance in 5'H column. b. Correlated to carbon resonance in 5 I 3 C column. o 92 r~- o o C M co L O 3 CN +1 -OO OS CN Q u c -a > u >^  x: o a, o OH 1/5 o o m P i C O LH .S DC 5 % % f CP) C M C M J CO : o - •& | CN +1 OO ON CN Q u Q CD > o >^  x o OH O o 1) o. CN u -t-m •0/ i -Sii E 94 95 Pachyclavulariolide E o Pachyclavulariolide E (47) was obtained as a white solid that gave a [M+H]+ peak in the HRFABMS at m/z 451.23322 appropriate for a molecular formula of C 2 4 H 3 4 O 8 (AM 0.06 ppm), requiring eight sites of unsaturation (Figure 2.36). Table 2.5 provides a summary of the NMR data acquired for 47 at 293 ± 2 K (20 ± 2 °C). The unsaturated functionalities evident from the NMR data were an a,(3-unsaturated ester (5 168.4 ppm/C16, 156.6/C1, 128.5/C15), two other esters (5 173.0/C21, 169.4/C23), and one ?ra/M-trisubstituted double bond (5 132.4 ppm/C4, 131.6/C5, 19.5/C18). Consequently, a tricyclic structure was apparent. The 'H NMR spectra obtained for pachyclavulariolide E (47) was different from those of previous compounds. Two additional methyl resonances and two downfield resonances at 8 6.5 and 8 5.0 ppm were present. An exchangeable proton resonance at 5.38 ppm indicated one hydroxyl group in the molecule (Figure 2.40). A detailed analysis of the COSY, HMQC, and HMBC data obtained for pachyclavulariolide E (47) confirmed that its structure was similar to pachyclavulariadiol (39) and pachyclavulariolide (42). Connectivity from C2 to C7 was established from the same NMR correlations as for the pachyclavulariolides. An a, p-unsaturated lactone was 95 o o o 97 also supported by the chemical shifts of CI6 (5 168.4), C15 (§ 128.5), CI (5 156.6) and C2 (8 106.6). H M B C correlations were observed between Me-17 (5 1.75) and CI (8 156.6) /CI5 (8 128.5) /CI6 (8 168.4). They verified the same 5-membered a, (3 -unsaturated lactone (CI, C2, CI5, CI6) methylated at the (3 (CI5) position (Figure 2.37). Therefore, pachyclavulariolide E (47) also contained a trisubstituted butenolide. HMBC correlations were observed between the OH resonance at 5.38 ppm and the C3/C2 resonances at 46.4 ppm and 106.6 ppm. These HMBC correlations confirmed that C2 was a hemiketal carbon. Figure 2.37: The fragment structure of 47 with selected COSY and H M B C correlations. One of the differences in structure between pachyclavulariolide E (47) and pachyclavulariolide A (43) is that 47 does not contain a C8-C13 bond. The doublet Me-19 signal was coupled by H8 and it suggested C8 was a methine carbon. The tetrahydro-furan ring still existed from C9 to C12 confirmed by the COSY correlations between H9 (8 3.50) andH10(8 1.27)/H10' (8 1.63), H10/H10' and HI 1 (8 1.50)/HI 1' (8 1.32). The carbon resonance of C9 at 85.3 ppm and CI2 at 82.9 ppm indicated those two carbons bore the oxygen (Figure 2.38). 97 98 Proton signals HI 3 (8 5.02) and H1.4 (8 6.52) were coupled to each other in the COSY spectrum. H M B C correlations between CH 3-20 (8 1.45) and C13 (8 75.4), H14 (8 6.52) and CI (8 156.6) /C2 (8 106.6) /C15 (8 128.5) established the connectivity between Figure 2.38: The fragment structure of 47 with selected COSY and H M B C correlations. C12 and C13; between C14 and CI. H M B C correlations showed that both C13 and C14 were attached to acetyl esters. H13 (8 5.02) was correlated to a carbonyl resonance at 8 173.0 (C21), and H14 (8 6.52) was also correlated to acarbonyl resonance at 8 168.4 (C23). Both carbonyl carbons C21 and C23 were directly attached to one methyl group Me-22 (8 20.0) and Me-24 (8 20.0), respectively, which was confirmed by H M B C correlations. Finally, the overall framework of 47 could be completely established by the combination of three fragments: the C2 to C7 chain, a trisubstituted butenolide, and the tetrahydrofuran ring from C9 to C14. The whole skeleton was also confirmed by the comparison with the known pachyclavulariadiol (42) isolated from the same species by 99 Dr. Philip Crews' group.""" Pachyclavulariadiol (42) contains a proton at C2 instead of a hydroxyl group, and has a vicinal diol at C13 and C14. The remaining structure of 42 was the same as that of pachyclavulariolide E (47). 98 99 The relative configuration of pachyclavulariolide E (47) was established by a single crystal x-ray diffraction analysis on the pachyclavulariadiol (39). Both 39 and 47 were isolated from the same specimens of P. violacea. Therefore, we have assumed that the relative configuration of the chiral centers at C8 (R*), C9 (S*), C12 (R*), CI 3 (S*), and C14 (R ) in pachyclavulariolide E (47) are the same as the corresponding centers in 39 as shown. Even though a difference in NOE was observed between the OH-2 and HI 3 resonances, the conformational mobility of the 14-membered carbocyclic ring precluded an unambiguous assignment of the relative configuration at C2 in 47. Figure 2.39: The structure of 47 with selected NOE correlations. 99 100 Table 2.5. NMR data for 47 recorded in benzene-d6 at 500 MHz ('H). Temp: 293 ± 2 K. c# 8 1 3 C(ppm) 5 'rl(ppm) a C O S Y Correlation b H M B C Correlation " N O E - D I F F 1 156.6 — H3,H14,H17 2 106.6 . . . — 2-OH,H3,H14 3/3' 46.4 3.09 2 - O H , H5 H 2 - O H , H 1 8 H5,H14,H18 4 132.4 — — H3,H18 5 131.6 6.30 H3,H6,H6',H18 H3,H18 H3,H14 6 25.3 2.15 H 5 , H 6 ' , H 7 , H 7 ' H7 6' 1.92 H 5 , H 6 , H 7 , H 7 ' 7 32.4 2.12 H 6 , H 6 ' , H 7 ' H6 .H19 T 1.05 H 6 , H 6 ' , H 7 , H 8 8 39.5 1.26 H 7 ' , H 1 9 H 6 ' , H 1 0 , H 1 0 ' , H 1 9 9 85.3 3.50 H8 ,H10 ,H10 ' H19 H10 ' ,H19 10 31.4 1.27 H 9 , H 1 0 ' , H l l , H i r H8 10' 1.63 H9 ,H10 ,H11 ,H11 ' 11 39.7 1.50 H10,H10',H11' H10,H10 ' ,H13 ,H20 i r 1.32 H10,H10 ' ,H11 12 82.9 . . . — H10 ' ,H13 ,H20 13 75.4 5.02(d, 1.53Hz) H14 H20 H14 .2 -OH 14 70.7 6.52(d,1.53Hz) H13 H13 H3,H5,H13,H24 15 128.5 . . . — H14.H17 16 168.4 — . . . H17 17 9.7 1.75 — . . . 18 19.5 1.90 H5 H3 H3 19 15.9 0.65(d,6.49Hz) H8 H7 20 23.5 1.45 — H13 21 173.0 . . . — H13.H22 22 20.0 1.40 — . . . 23 168.4 — — H14,H24 24 20.0 1.59 — . . . 2 - O H . . . 5.38 H3 — H13 a. Correlated to proton resonance in 5 ' H column. b. Correlated to carbon resonance in 8 L C column. O 100 102 (ppm) (PPm) P J J L •O-a J L * e 7.00 6.00 5.00 4.00 3.00 2.00 1,00 Figure 2.41: 2D COSY spectrum of pachyclavulariolide E (47) in C 6 D 6 at 293 ± 2 K. 102 103 Figure 2.42: 2D H M Q C spectrum of pachyclavulariolide E (47) in C 6 D 6 at 293 ± 2 K. 103 104 Pachyclavulariolide F Pachyclavulariolide F (18) was obtained as a white solid that gave a [M+H]+ peak in the HRFABMS at m/z 507.33198 (Figure 2.43) appropriate for a molecular formula of C29H46O7 (AM -0.40 ppm), requiring seven sites of unsaturation. Table 2.6 provides a summary of the NMR data acquired for 18 at 293 ± 2 K (20 ± 2 °C). The unsaturated functionalities evident from the NMR data were an a, (3-unsaturated ester (5 170.7 ppm/C16, 154.9/C1, 133.4/C15), one rra/M-trisubstituted double bond (8 130.1 ppm/C4, 131.2/C5, 17.8/C 18), and another ester (8 172.8/C21). Consequently, a tricyclic structure was apparent. Fragments ions in the FABMS corresponding to m/z 475, [M+H]+-32 [M+H-MeOH] +; 363, [M+H]+-144 [M+H-C 8 H, 6 0 2 ] + ; 331, [M+H]+-32-144 [M+H-MeOH-C 8 H 1 6 0 2 ] + ; and 313, [M+H]+-32-144-18 [M+H-MeOH- C 8 H 1 6 0 2 - H 2 0 ] + , suggested the presence in 18 of methyoxyl (-OCH3), octanoate (CsHi602), and hydroxyl (OH) groups. Subtraction of the nine carbons associated with the 8-carbon aliphatic ester and the methyoxyl group left 20 carbons, suggestive of a diterpene skeleton. 104 o o do G O b> O) co + f£ s CVJ CO II o cSo C J C O > '—' co C O _ h-cn cr _ i s w co co o • o 5 o j LO a C D SI 3 O 3 O = CD O • LO LO fe-o • o LO fe-es • LO -cr o o 00 1) T 3 x. 3 > x. *o >^  x: % ex <+» o £ o ID CX oo 03 E < G o - o oo <D l -& O J IT, rt u S M l i n o ™ co g a. o o o o VO o 0)Q O LO +1 CO ON CN ^ cd VD Q U c c 0 0 PH T 3 3 > o E. o ^ OH o o in z, T f T t cN cu !_ 3 Oi) vo o r-o 1 0 8 A detailed analysis of the COSY, HMQC and HMBC data obtained for 18 confirmed that the diterpene skeleton belonged to the cembranolide family. The 1 3 C NMR data had one carbon resonance at 110.1 ppm (C2) corresponding to a ketal carbon. HMBC correlations were observed between CH 3-17 (8 2.28) and CI (8 154.9) /CI5 (8 133.4) /CI6 (8 170.7). They identified a 5-membered a, (3-unsaturated lactone (CI, C2, C15, C16) that is methylated at the (3 (C15) position. A H M B C correlation between C H 3 -29 (8 3.18) and C2 (8 110.1) verified that the ketal carbon C2 was connected to a methoxyl group. The 'H NMR data also exhibited three oxymethine protons at 8 4.96 (H9), 8 3.35 (H13), and 8 4.14 ppm (H14). HMQC correlations identified the chemical shifts of the carbon resonances assigned to C3, C5-C11, CI 8, C19 and their attached protons. An examination of the COSY and HMBC spectroscopic data established the connectivity from C2 to CI 1 (Fig. 2.46). COSY correlations were observed between Figure 2.46: The fragment structure of 18 with selected COSY and H M B C correlations. H3 (8 2.81) and H3'(8 2.58), H5 (8 4.86) and H6 (8 2.00) /H6'(8 1.70) /H18 (8 1.65), H6 and H7 (8 1.20) /H7'(8 0.96), H77H8 (8 1.50), H8 and H9 (8 4.96), H9 and H10 (8 1.42) /H10'(8 1.28), H10/H10' and HI 1 (8 1.82) /Hll'(8 1.12). The CH 3-19 (8 0.72, d, J = 6.92 Hz) resonance showed a COSY correlation to H8. HMBC correlations observed between 1 0 8 109 CH 3-18 (8 1.65) and C3 (8 46.1) /C4 (8 130.1) /C5 (8 131.2) confirmed that CH 3-18 was attached to the C4 olefinic carbon. Therefore, the connectivity from C2 to CI 1 including CH 3-18 and CH 3 -19 was established. The 1 3 C chemical shift of C18 (8 19.5) indicated that the trisubstituted A 4' 5 olefin was trans. A DEPT spectrum of 18 exhibited only one aliphatic quaternary carbon at 8 63.2 (C12), and H M B C correlations were observed between Me-20 (8 1.10) and CI 1 (8 34.8), C12 (8 63.2) and C13 (8 64.3). Therefore, CH 3-20 had to be attached directly to the aliphatic quaternary carbon C12 (8 63.2), and C12 had to be also connected to C11, CI 3, and oxygen. The oxymethine H13 (8 3.35) was further coupled to another oxymethine H14 (8 4.14, dd, J=3.91, 6.17Hz). HMBC correlations between H14 and CI (8 154.9), C2 (8 110.1), and C15 (8 133.4) verified that C14 was connected to CI. This final connectivity generated the typical 14-membered ring found in cembranolides. An exchangeable proton resonance at 8 2.14 (d, J= 3.91Hz) showed coupling in the COSY spectrum to HI4 (8 4.14), showing that there was a hydroxyl group connected to C14. Consequently, the single remaining oxygen atom had to be part of an epoxide linking the oxymethine carbon C13 (8 64.3) and the oxygen bearing aliphatic carbon C12 (8 63.2). Thus, the constitution of 18 could be completely established as shown. The relative stereochemistry at the six chiral centers in 18 were assigned using the following general strategy. Stereochemical assignments were based on NOE experiments, 'H NMR coupling constants, and the known stereochemical centers of the pachyclavulariolide E (47). The relative configurations at C8 and C14 were assumed to be 8RP and 14R" which were identical to the corresponding centers in 47. 109 110 A NOE correlation observed between HI4 (5 4.14) and Me-20 (5 1.10) (Fig. 2.47) suggested that the C14 and the CH3-2O were on the same side of the epoxide ring. The vicinal coupling constant (J=6.17Hz) between H13 and H14 indicated that the dihedral angle H14-C14-C13-H13 was approximately 30 or 130 degrees according to the Karplus equation. The H14-C14-C13-H13 dihedral angle could not be 30 degrees because this would make the H14 and CH3-2O distance too large to have a NOE correlation. Also, since no NOE was observed between 14-QH (8 2.14) and HI 3 (8 3.35), the dihedral angle between HO(14)-C14-C13-H13 had to be greater than 90 degrees. Therefore, the dihedral angle H14-C14-C13-H13 could be =130 degrees and there is only one configuration (shown below) that corresponds to this data. By this analysis the C12 and C13 chiral centers would be 12S and 13S*, respectively. The configuration of the chiral center C9 was difficult to determine due to the flexibility within the cembrane skeleton. Therefore, assignment of the relative configuration at C9 was based on a putative biogenetic relationship between pachyclavulariolide E (47) and pachyclavulariolide F (18). One reasonable biogenetic scenario (Figure 2.56) in Section 2.3.3 would have the C12/C13 epoxide in a common precursor 48A undergoing nucleophilic opening by a C9 alcohol to generate the C9/C12 ether and C13 alcohol found in 47. This proposal would predict that the C9 (S*) and C13 (S*) configuration in pachyclavulariolide F (18) should be identical to the C9 and C13 configuration in pachyclavulariolide E (47), and assuming that there 110 111 would be backside attack of the C9 alcohol on C12 of the epoxide, it would predict that the CI 2 (S*) configuration in 18 should be the opposite of the CI 2 configuration in pachyclavulariolide E (47). An alternate scenario where a CI3 alcohol displaces the C12 ether bond to give the C12/C13 epoxide and a C9 alcohol in 18 results in the same prediction for the relative stereochemistries at C9, C12, and CI3 in 18. A difference NOE was observed between the CH 3-29 (5 3.18) and H14 (5 4.14) resonances, however, the conformational flexibility of the fourteen membered ring in 18 precludes a completely unambiguous C2 assignment on the basis of this NOE data alone. Pachyclavulariolide F (18) was the diterpenoid from P. violacea which showed G2 cell cycle checkpoint inhibition activity. It was active at approximately 1.0 to 2.0 pLg/ml." Unfortunately, it also showed in vitro cytotoxcity against murine leukemia P388 cells with an E D 5 0 of 1.0 pig/ml." Therefore, it was not a specific G2 cell cycle checkpoint inhibitor. The biological data was measured by Dr. Xiuxian Jiang in the Michel Roberge's Laboratory. 113 Table 2.6: NMR data for pachyclavulariolide F recorded in benzene-^ at 500 MHz (*H). c# 5 1 3 C(ppm) 5 'H(ppm) a C O S Y Correlation b H M B C Correlation : ' N O E - D I F F 1 154.9 — — H3,H3' ,H13 ,H14 H 1 7 , 1 4 - O H 2 110.1 — . . . H3,H3' ,H14 ,H29 3 46.1 2.81(d,14.4Hz) H 3 ' H5,H18 H3' ,H18 3' 2.58(d, 14.4Hz) H 3 H3,H5,H14 4 130.1 . . . . . . H3,H3' ,H18 5 131.2 4.86(t,6.34Hz) H 6 , H 6 ' H3 ,H3 ' ,H6 ,H7 ,H18 H3' ,H14 6 26.8 2.00 H 5 , H 6 ' , H 7 H 5 , H 7 , H 7 ' 6' 1.70 H 5 , H 6 , H 7 , H 7 ' 7 31.9 1.20 H 6 , H 6 ' , H 7 ' , H 8 H 5 , H 6 , H 6 ' , H 8 , H 1 9 7' 0.96 H 6 ' , H 7 , H 8 8 38.3 1.50 H7 ,H7' ,H19 H6 ,H7 ,H19 9 73.7 4.96(m) H10,H10' H7,H10,H11.H19 H13.H14 10 23.5 1.42 H 9 , H 1 0 ' , H l l , H i r H I 1,H1 r 10' 1.28 H 9 , H 1 0 , H l l , H i r . . . 1 1 34.8 1.82 H10,H10' ,H11' H10,H13,H20 i r 1.12 HI 1,H10,H10' 12 63.2 — — H10,H11,H13,H20 13 64.3 3.35(d,6.17Hz) H14 H11,H14,H20 H9,H14,H17 14 63.3 4.14(dd, 3.91, 6.17Hz) H 1 3 , 1 4 - O H H 1 3 , 1 4 - O H H3' ,H29 ,HI3 , 14-O H , H 2 0 15 133.4 . . . — H14.H17 16 170.7 . . . . . . H17 17 9.8 2.28 . . . _ _ _ H13 18 17.8 1.65 . . . H 3 , H 3 ' , H 5 19 16.3 0.72(d,6.92Hz) H8 . . . H8 20 16.1 1.10 — . . . 21 172.8 — . . . H22,H23 22/22' 34.5 2.09(t,7.27Hz) H23 H23,H24 23/23' 25.0 1.50 H22,H24 H22.H24, 24/24' 31.5 1.19 H23.H25 H22,H23,H25,H26 25/25' 31.5 1.19 H24,H26 H23.H26 26/26' 31.9 1.19 H25.H27 H25,H28 27/27' 22.6 1.18 H26.H28 H26.H28 28 14.0 0.90(t,7.1 1Hz) H27 H26 29 50.3 3.18 . . . . . . H14 14-OH — 2.14(d,3.91Hz) H14 . . . H14.H20 a. Correlated to proton resonance in 5 H column. b. Correlated to carbon resonance in 5 1 3C column. 113 114 115 ihCO ( p p m ) J j J A_ 0 <3t) 0t?> 0®> tf) 20 40 60 80 100 120 ( p p m ) 4.8 4.0 3.2 2.4 1.6 0.8 Figure 2.49: 2D H M Q C spectrum of pachyclavulariolide F (18) in C 6 D 6 at 293 ± 2 K. 115 116 Pachyclavulariolide G 48 Pachyclavulariolide G (48) was obtained as a white solid that gave a [M+H]+ peak in the HRFABMS at m/z 493.31781 (Figure 2.50) appropriate for a molecular formula of C28H44O7 (AM 2.59ppm), requiring seven sites of unsaturation. Table 7 provides a summary of the NMR data acquired for 48 at 293 ± 2 K (20 + 2 °C). The unsaturated functionalities evident from the NMR data were two esters (5 174.6ppm/C16, 173.0/C21) and one frans-trisubstituted double bond (8 130.3ppm/C4, 129.5/C5, 15.9/C18). Consequently, a tetracyclic structure was apparent. Fragment ions in the LRFABMS corresponding to m/z 475, [M+H]+-18 [M+H-H 2 0] + ; 349, [M+H]+-144 [M+H-C 8 H 1 6 0 2 ] + ; 331, [M+H]+-144-18 [M+H-C 8 H 1 6 0 2 -H 2 0] + , confirmed the presence of hydroxyl (OH) and octanoate (CsH| 60 2) groups in 48. Subtraction of the eight carbons associated with the aliphatic ester left 20 carbons, suggestive of a diterpene skeleton. Meanwhile, 48 in MeOH solution gave a strong [M+Na]+ peak at m/z 515.0 in the low resolution ESI-MS. When dissolved in deuterated methanol (CD 3 OD, 99.8%), 48 gave two strong peaks at m/z 515.1 and 516.0. This observation indicated 48 had one exchangeable proton. A detailed analysis of the COSY, HMQC, and HMBC data obtained for 48 confirmed that the molecule contained a cembranolide diterpene core. Its structure is similar to pachyclavulariolide F (18). 116 a s-i a i 1 S3 00 o > cd o >, x: o cd OH o O H cd s i 7 g. Si W X) c cd 03 < U H c . o o © i n C N u S OJj 119 COSY correlations were observed between H3 (5 2.68) and H3'(5 2.70), H5 (5 5.05) and H6 (5 2.10) /H6'(8 1.68) /H18 (8 1.65), H6 and H7 (8 1.30) /H7' (8 1.29), H7' /H8 (8 1.46), H8 and H9 (8 4.85), H9 and H10 (8 1.55) /H10' (8 1.65), H10/H10' and HI 1 (8 2.05) /HI 1 '(8 0.86). The CH 3-19 (8 0.82, d, 6.91 Hz) signal was a doublet in the 'H NMR due to coupling by H8 (8 1.46). HMBC correlations observed between CH3-18 (8 1.65) and C3 (8 38.5) /C4 (8 130.3) /C5 (8 129.5) confirmed CH3-18 was attached to the olefinic C4. Therefore, connectivity from C2 to CI 1 including CH 3-18 and CH 3-19 was assembled. The l 3 C chemical shift of CI8 (8 15.9) indicated that the trisubstituted A 4 ' 5 olefin was trans. H M B C correlations were observed between CH 3-20 (8 1.00) and CI 1 (8 36.7), C12 (8 61.5) and C13 (8 63.9), indicating CH 3-20 would connect to the aliphatic quaternary carbon C12 (8 61.5) directly, and C12 would connect CI 1, C13 and oxygen. The oxymethine H13 (8 2.58) was further coupled to another oxymethine H14 (8 3.55, J= 7.25 Hz). H M B C correlations between H14 and CI (8 70.0) /C2 (8 90.1) /C15 (8 40.2) verified C14 was attached to CI. One exchangeable proton (8 1.86) was coupled to H14 in the COSY spectrum (solvent: CeD6), indicating the one hydroxyl group was connected to C14. This correlation was clearly shown in the COSY spectrum when 48 was dissolved in acetone-d6. The doublet exchangeable proton (8 4.92) was correlated with the quartet proton H14 with a coupling constant of 4.88 Hz. The chemical shift of C12 at 61.5 ppm and CI3 at 63.9 ppm suggested that both carbons were attached to an oxygen and had an epoxide ring. 119 120 The difference in structure from pachyclavulariolide F (18) is in the trisubstituted butenolide component. HMBC correlations were observed between CH 3-17 (5 1.10) and CI (5 70.0), C15 (5 40.2) and C16 (5 174.6). However, CI and C15 were changed to aliphatic carbons with the chemical shifts of 70.0 ppm and 40.2 ppm, respectively. Thus the 5-membered lactone ring was present, except that the a, P unsaturated double bond was lost. The l 3 C chemical shift of C2 at 90.1 ppm still indicated that it was connected to two oxygen atoms. Therefore, the single remaining oxygen could be part of another Figure 2.52: The butenolide core with HMBC correlations and model A and B structure, epoxide linking the ketal carbon C2 (8 90.1) and the oxygen bearing aliphatic carbon CI (8 70.0). Terrestrial terpenoids eremofarfugin A 2 5 and senauricolide25 isolated from rhizomes of several different species have the similar epoxy-butenolide core. The 1 3 C chemical shift of CI in these molecule usually is 62 ppm. So the 1 3 C chemical shift of CI in 48 was downfield compared to these molecules suggesting that CI in 48 may have oxygen linkage with other carbons (C12 or C13). Therefore, three different oxygen linked structures (48, Model A, and Model B) were proposed (Figure 2.52). Because C12 and C13 have l 3 C NMR resonances around 62-63 ppm, which is the typical resonance for OH 48 M o d e l A M o d e l B 120 121 epoxide ring, and the analog 18 has the epoxide ring at C12 and C13 with the similar 1 3 C chemical shift, 48 was the most likely structure for pachyclavulariolide G. The relative configuration of pachyclavulariolide G (48) was based on NOE correlations, 'H NMR coupling constants, and known stereochemical centers of pachyclavulariolide F (18). The relative configuration of the chiral centers at C8 (R*), C9 (S*), C12 (S*), C13 (S*), C14 (R*) in 48 were assumed to be the same as the corresponding centers in 18 as shown. A NOE correlation observed between H14 (8 3.55) and CH 3-20 (8 1.00) suggested that C14 and CH3-I2 were on the same side of the epoxide ring (Fig. 2.53). The 'H vicinal coupling constant (J = 7.25 Hz) between HI 3 and H14 indicated that the dihedral angle H14-C14-C13-H13 was around 145 or 20 degrees according to the Karplus equation. The dihedral angle H14-C14-C13-H13 could not be 20 degrees because this would give a H14 to CH3-2O distance too large to give the observed NOE correlation. Furthermore, no NOE was observed between 14-OH and HI3, suggesting that the dihedral angle between HO (14)- C14- C13- HI3 was greater than 90 degrees. Therefore, the H14-C14-C13-H13 dihedral angle must be approximately 145 degrees as shown. o 77 1.10 Figure 2.53: The selected NOE correlations in pachyclavulariolide G (48). 121 122 Thus the C12 and C13 chiral centers would be 12S* and 13S*, respectively, if the configuration at C14 was assumed as 14R*. A strong NOE difference was observed between H14 (5 3.55) and H15 (5 2.75), and a weak NOE difference was observed between H17 (8 1.10) to H14 (8 3.55). However, the conformational flexibility of the fourteen membered ring in 48 precludes a completely unambiguous CI, C2, and C15 assignment on the basis of these NOE data. P (ppm) 4.8 P A J U U L i u . (ppm) 4.0 3.2 2.4 P o 1.6 0.8 2.4 3.2 4.0 4.8 0.8 Figure 2.54: 2D COSY spectrum of pachyclavulariolide G (48) in C 6 D 6 at 293 ± 2 K. 122 123 _ A _ _ A _ ( p p m ) ^ V. 20 40 60 80 100 120 ( p p m ) 4.8 4.0 3.2 2.4 1.6 0.8 Figure 2.55: 2D HMQC spectrum of pachyclavulariolide G (48) in C 6 D 6 at 293 ± 2 K. 123 124 Table 2.7: NMR data for 48 recorded in benzene-d6 at 500 MHz ('H). c# 5 i 3 C ( p p m ) 8 'H(ppm) a C O S Y Correlation b H M B C Correlation a N O E - D I F F 1 70.0 _ _ _ _ _ _ H14,H15,H17 2 90.1 . . . . . . H3,H3' ,H14 3 38.5 2.68 H3 ' ,H5 H5,H18 3' 2.70 H3 4 130.3 — . . . H3,H3' ,H18 5 129.5 5.05(m) H 3 , H 6 , H 6 ' H3 ,H3 ' ,H6 ' ,H18 H3,H13 6 26.0 2.10 H 5 , H 6 ' , H 7 , H 7 ' H5, 6' 1.68 H 5 , H 6 , H 7 , H 7 ' 7 33.0 1.30 H 6 , H 6 ' , H 7 ' , H 8 H6' ,H19 7' 1.29 H 6 , H 6 ' , H 7 , H 8 8 35.3 1.46 H 7 , H 7 ' , H 9 , H 1 9 H19 9 77.0 4.85(m) H8,H10,H10' H10.H19 H20 10 25.4 1.55 H9,H10' ,H11 ,H11' . . . 10' 1.65 H9,H10,H11,H11' _ _ _ 11 36.7 2.05 H10,H10' ,H1 r H9,H13 ,H20 i r 0.86 H10,H10' ,H1 1 12 61.5 — . . . H11,H13,H20 13 63.9 2.58(d,7.25Hz) . H14 H I 1,H14,H20 H5,H14 14 69.0 3.55(d,7.25Hz) H13 H I 3 H15.H20 15 40.2 2.75(q,7.20Hz) H17 H14.H17 H14.H17 16 174.6 . . . . . . H15,H17 17 10.5 1.10(d,7.20Hz) H15 H15 H14,H15 18 15.9 1.65 . . . H 3 , H 3 ' , H 5 H 6 19 17.4 0.82(d,6.91Hz) H8 H 9 H9.H11.H13 20 16.3 1.00 . . . . . . H14 21 173.0 . . . . . . H9,H22,H22' 22/22' 35.3 2.20(t,7.44Hz) H23,H23' H24.H24' 23/23' 25.7 1.60 H 2 2 , H 2 2 \ H 2 4 , H 2 4 ' H22,H22' ,H24,H24' 24/24' 29.4 1.20 H23,H23' ,H25,H25' H22,H22' ,H23,H25 25/25' 29.3 1.20 H24,H24' ,H26,H26' H23,H23' ,H26,H26' 26/26' 32.0 1.20 H25,H25' ,H27,H27' H25,H27,H28 27/27' 22.9 1.20 H26,H26' ,H28 H28 28 14.2 0.90(t,7.08Hz) H27,H27' . . . a. Correlated to proton resonance in S ' H column. b. Correlated to carbon resonance in 8 I 3 C column. 124 125 2.3.3 Discussion and conclusions: The soft coral P achy clavularia violacea is a rich source of cembranolide diterpenes. The crude extract of specimens collected in PNG showed moderate activity in assays for in vitro cytotoxicity and G2 cell cycle checkpoint inhibition. Fractionation of the P. violacea extract yielded six new diterpenoid secondary metabolites, pachyclavulariolides A (43) to G (48). Pachyclavulariolide F (18) was the only one that had significant biological activity. It showed in vitro cytotoxicity against murine leukemia P388 with an E D 3 0 of 1.0 Ug/ml, and had G2 checkpoint inhibition activity at 1.0 to 2.0 jlg/ml in vitro. Two previous studies of P. violacea resulted in the isolation of the cembranoids pachyclavulariadiol (39) and its naturally occurring mono- (40) and diacetylated (41) derivatives from specimens collected in Australia,21 and pachyclavulariolide (42) from specimens collected in Vanuatu.22 The P. violocea specimens examined in the current study were collected at a site on the southwestern coast of Papua New Guinea that borders on the Coral Sea, not far from either the Australian Barrier Reef or Vanuatu locations that border on the same water mass. The three new cembranolides, pachyclavulariolide E (47), F (18), and G (48), are closely related to the cembranoids 39 and 42 found in the Australian and Vanuatu specimens. Pachyclavulariolide E (47) had a hydroxyl group at C2. Pachyclavulariolide F (18) and G (48) are missing the tetrahydrofuran ring, and they have an octanoate ester at C9 and an epoxide ring between C12 and C13. Pachyclavulariolide F (18) has an oc,(3-unsaturated 5-membered lactone, while pachyclavulariolide G (48) has an epoxide ring at CI and C2. 125 126 The new diterpenoids pachyclavulariolide A (43) to D (46) have briarane carbon skeletons. The 9, 12 ether bridge that forms an oxanorbornane substructure in 43 to 46 is without precedent among known briarane diterpenoids. Pachyclavulariolides A (43), B (44), and C (45) were also found to be present as two major conformations in solution. Taken together, the two earlier studies and the current study of P. violacea indicate that there is a significant geographic variation in diterpenoid content in this octocoral. From a biosynthetic point of view, all compounds obey the mevalonate biosynthetic pathway. They are all derived from Cs isoprene units joined in a head to tail fashion. They are all oxygenated at several carbon sites to form a cyclic oxygen bridge, lactone, or epoxide, and two compounds are further esterified to produce the final products. Pachyclavulariolides E (47), F (18), and G (48) probably originated from the same precursor 48A (Fig. 2.56). The epoxide ring at C12 and C13 was most likely formed from the unsaturated double bond in the original isoprene by oxygenation. A hydroxyl group at C9 can attack C12 by a nucleophic substitution mechanism to force the epoxide ring to open. C2 in 48A could be a methine carbon and can be oxidized to the hemiketal. The newly formed hydroxyl group at CI3 and the original hydroxyl group at CI4 can be further acetylated to generate pachyclavulariolide E (47). The hydroxyl group at C9 in 48A could also be esterified to an octanoate ester (48B). The further oxidation and methylation of 48B at C2 generates pachyclavulariolide F (18). In addition, rearrangement of A 1 ' 1 5 in 48B can form 48C with a double bond at C1-C2 position. The oxidation of this double bond could give the pachyclavulariolide G (48). This raises the very challenging issue to understand the biosynthetic pathways in marine invertebrate natural products. 126 127 O Pachyclavulariolide E (18) Figure 2.56: Biogenic proposal for pachyclavulariolide E (47), F (18), and G (48). 127 128 2.3.4: Experimental NMR data were collected on either, a Bruker AMX500, a Bruker WH400 or a Bruker AM400 spectrometer each equipped with a 5mm probe. All spectra were obtained in benzene-d6. Proton spectra were referenced using internal residual C 6 D 6 (8 7.15) and carbon spectra were referenced to the C 6 D 6 carbon resonance (8 128.0). FABMS data were collected on a Kratos Concept IIHQ hybrid mass spectrometer with cesium ion secondary ionization and a magnetic sector mass analyzer. Samples were dissolved in a MeOH-thioglycerol matrix and spectra were obtained using a source voltage of 8 K V and a cesium ion gun voltage of 12 KV. ESI-MS data were collected on a EsquireLC_00085 mass spectrometer with an iontrap detector. Optical rotations were measured on a Jasco J-710 spectropolarimeter (1 cm quartz cell), and the [a]o values are given in 10"1degcm2g"1. Reversed phase and normal phase thin layer chromatography (TLC) was performed using Whatman MKC18F and Kieselgel 60 F 2 5 4 plates. Visualization was detected by U V (X=254 nm) and/or heating after spraying with vanillin reagent. Normal phase column chromatography was carried out either on Merck silica gel G60 (230-400 mesh) or Sigma silica gel (size: 10-40u). Reversed phase chromatography was performed using reversed-phase silica prepared according to literature.27 Sephadex LH-20 (bead size 25-lOOu.) was used for size exclusion chromatography. High performance liquid chromatography (HPLC) separations were performed on one of two possible systems using either a Whatman Partisil 10 ODS-3 Magnum column or Rainin Partisil 10-ODS column. The first system consisted of a Waters 600E HPLC pump/system controller with a Waters 486 tunable absorbance detector. The second system consisted of a Waters 600E HPLC pump/system controller equipped either with a 128 129 Waters 996 photodiode array detector or RI (Perkin Elmer LC-25) detector. Both systems were interfaced with a personal computer using Millenium™ 2010 chromatography software. The solvents used for extraction and for column chromatographies were Fisher reagent grade. HPLC solvents were Fisher HPLC grade which were filtered and degassed prior to use. All other solvents, reagents and standards were reagent or commercial grade and were used without further purification. Single crystal x-ray diffraction analysis of pachyclavulariolide B (44): The crystal of pachyclavulariolide B (44) was analyzed by Brian Patrick in the X-ray lab. A crystal of 44 with dimensions 0.45 x 0.20 x 0.12 mm was mounted on a glass fibre. Data were collected at-100 °C on a Rigaku/ADSC CCD area detector in two sets of scans (<|> = 0.0 to 190.0°, % = -90°; and co = -18.0 to 23.0°, x = -90°) using 0.50° oscillations with 58.0 second exposures. The crystal-to-detector distance was 40.41 mm with a detector swing angle of -5.53°. The material crystallized in space group P2{2{2\ with a = 7.390(4), b = 9.465(1), and c = 25.138(3) A. Of the 3509 unique reflections measured (Mo-Ka radiation, 20 m a x = 55.8°, R,„, = 0.080, Friedels not merged), 23 1 1 were considered observed (I> 3a(I)). The final refinement residues were R = 0.033 (on F, I >3a(I)) and R 2 = 0.094 (on F 2 , all data). The data was processed using the d'TREK program and corrected for Lorentz and polarization effects. The structure was solved by direct methods9 and all non-hydrogen atoms were refined anisotropically, while all methyl and hydroxyl were refined isotropically. All other hydrogens were included in calculated positions. This enantiomorph was chosen based on the known configuration of 129 130 the various stereocentres. All calculations were performed using the teXsan10 crystallographic software package of Molecular Structure Corporation. Pachyclavulariolide A (43): White amorphous solid; [a]o = +17.10 (10"1 deg cm 2 g"1, MeOH); HRFABMS: [M+H] m/z 317.21150 ( C 2 0 H 2 9 O 3 , AM -0.53ppm); 'H NMR (C 6 D 6 , 500MHz) and 1 3 C NMR (C 6 D 6 , 125MHz), and 2D NMR data listed in Table 2.1. Pachyclavulariolide B (44): Crystalline solid; m.p. 152.3-154.6 °C; [a]D = +23.1° ( c = 10_1 deg cm 2 g ', MeOH); HRFABMS: [M+H] m/z 333.20663(C2 0H2 9O4, AM 0.15ppm); LRFABMS, m/z [formula, relative intensity %]: 315 (C 2 0H 2 7O 3 ,50); 'H NMR (C 6 D 6 , 500MHz) and 1 3 C NMR (C 6 D 6 , 125MHz), and 2D NMR data listed in Table 2.2. Pachyclavulariolide C (45): White amorphous solid; [a]D = +11.1° (10"1 deg cm 2 g"1, MeOH); HRFABMS: [M+H] m/z 347.22162 ( C 2 l H 3 1 0 4 , AM -1.77ppm)); LRFABMS, m/z [formula, relative intensity %]: 315 (C 2 oH 2 7 0 3 , 19); 'H NMR (C 6 D 6 , 500MHz) and l 3 C NMR (C 6 D 6 , 125MHz), and 2D NMR data listed in Table 2.3. Pachyclavulariolide D (46): White amorphous solid; [a]o = +12.1° (10"' deg cm 2 g"1, MeOH); HRFABMS: [M+H] m/z 347.22079 ( C 2 1 H 3 1 0 4 , AM -4.17ppm); LRFABMS, m/z [formula, relative intensity %]: 315 ( C 2 0 H 2 7 O 3 , 19); *H NMR (C 6 D 6 , 500MHz) and 1 3 C NMR (C 6 D 6 , 125MHz), and 2D NMR data listed in Table 2.4. Pachyclavulariolide E (47): White amorphous solid; [OC]D = -15.1° (10"' deg cm 2 g"1, MeOH); HRFABMS: [M+H] m/z 451.23322 ( C 2 4 H 3 5 0 8 , A M 0.06ppm); LRFABMS, m/z [formula, relative intensity %]: 433 ( C 2 4 H 3 3 0 7 , 24); *H NMR (C 6 D 6 , 500MHz) and l 3 C NMR (C 6 D 6 , 125MHz), and 2D NMR data listed in Table 2.5. 130 131 Pachyclavulariolide F (18): White amorphous solid; [CX]D = +32.1° (10"' deg cm 2 g"1, MeOH); HRFABMS: [M+H] m/z 507.33198 (C 2 9 H 4 70 7 , AM -0.40ppm); LRFABMS, m/z [formula, relative intensity %]: 475 ( C 2 8 H 4 3 0 6 , 70), 363 ( C 2 | H 3 | 0 5 , 18), 331 (C 2 oH 2 7 0 4 , 45), 313 ( C 2 0 H 2 5 O 3 , 33); 'H NMR (C 6 D 6 , 500MHz) and l 3 C NMR (C 6 D 6 , 125MHz), and 2D NMR data listed in Table 2.6. Pachyclavulariolide G (48): White amorphous solid; [a]o = +81.1° (10"' deg cm 2 g"1, MeOH); HRFABMS: [M+H] m/z 493.31781 ( C 2 8 H 4 5 0 7 , AM 2.59ppm); LRFABMS, m/z [formula, relative intensity %]: 475 (C 2 8H 4 3 0 6 , 30), 349 (C 2 0 H 2 <A, 33), 331 (C2(,H2704, 17); 'H NMR (C 6 D 6 , 500MHz) and , 3 C NMR (C 6 D 6 , 125MHz), and 2D NMR data listed in Table 2.7. Reference: 1. Paul M . Dewick, Medicinal Natural Products: a biosynthetic approach, 1M edn, John Wiley, New York, pi 84-201. 2. D. J. Faulkner. Nat. Prod. Rep., 1997, 14, 259-297. 3. William Fenical. Marine Natural Products, Chapter 3, pi73, 1978. 4. Vanderah, D. J . ; et al. J. Am. Chem. Soc, 1998, 99, 5780-5782. 5. T. Iwagawa; J-I. Kawasaki; T. Hase, J. Nat. Prod., 1998, 61, 1513-1515. 6. Amico, V.; Oriente, G.; et al. J. Chem. Soc, Chem. Commun. 1976, 1024-25. 7. B.L. Raju; G.V. Subbaraju; C B . Rao. lnd. J. Chem., Sect. B., 1995, 34, 221-225. 8. A. S. R. Anjaneyulu; G. V. Rao; K.V.S. Raju; R.K. Murthy. lnd. J. Chem., Sect.B, 1995,34, 1074-1076. 9. Herin, M. ; Tursch, B. Bull. Soc Chim. Belg., 1976, 85, 707-711. 131 132 10. Schmitz, F.J.; Vanderah, D.J.; Ciereszko, L.S. J. Chem. Soc. Chem. Commun. 1974, 407-409. 11. Tursh, B. Pure Appl. Chem. 1976, 48, 1-6. 12. C-Y Duh; S-K wang; H-K Tseng; J-H sheu, Tetrahedron Lett. 1998, 39, 7121-7122. 13. Missakian, M.g.; Burreson, B.J.; Scheuer, P.J. Tetrahedron, 1975, 31, 2513-1215. 14. A. Rudi; T. L-A . Dayan; M. Aknin; E .M. Gaydou; Y..Kashman. J. Nat. Prod. 1998, 61, 872-875. 15. Kashman, Y.; Groweiss, A. Tetrahedron Lett. 1977, 1159-1 162. 16. Coll, J . C ; Price, I.R.; Konig, G. M.; Bowden. B. F. Mar. Biol. 1987, 96, 129-135. 17. Coll, J. C ; Bowden. B. F.; Alino, A.; Heaton, A.; et al. Chem. Scr. 1989, 29, 383-388. 18. Schmitz, F.J.; Bowden. B.F.; Toth, S.I. Marine Biotechnology, Attaway, D.H., Zaborsky, O.R., Eds. Plenum: New York, 1993; Vol. 1, 197-308. 19. Norton, R.S.; Kazlauskas, R. Experientia 1980, 36, 276-278. 20. Coval, S.J.; Patton, R.W.; Petrin, J.M.; et al. Bio.Med. Chem. Lett 1996, 6, 909-912. 21. Bowden, B.F.; Coll, J . C ; et al. Aust. J. Chem., 1979, 32, 2265-227'4. 22. Wayne Inman; Phillip Crews. J. Org. Chem. 1989, 54, 2526-2529. 23. Molecular models were calculated by the Chem 3D standard™ (version 4.0.1). Two conformations with the lowest free energy were obtained by MM2 method. 24. X-ray analysis was performed by Brain Patrick in the X-ray lab at the Department of Chemistry in UBC. The method was listed on the Experimental Section 2.3.4. 25. Motoo, T.; Shiotani, Y.; et. al. Tetrahedron Letters, 2000, 41, 1797-1799. 26. Torres, P.; Ayala, J.; et al. Phytochemistry, 1998, 47(1), 57-61. 27. Kiisher, T.C.; Lindsten, G.R. / . Org. Chem. 1983, 48, 3589-3591. 132 133 Chapter III: G2 cell cycle checkpoint inhibitors 3.1 Introduction: This chapter describes three G2 cell cycle checkpoint inhibitor projects. The first includes the chemistry found in culture extracts of the marine microorganism species known as clinl 116, the second involves the chemistry in an Indonesian sponge Aaptos suberitoides, and the third describes the chemistry of a South African plant Parinari curatellifolia obtained from the National Cancer Institute (USA) open repository of natural product extracts. The crude extracts of these three samples exhibited strong G2 checkpoint inhibition activity in preliminary bioassay screening. 3.2 Marine Actinomycete Cl in l l l 6 The G2 bioassay was used to search for G2 checkpoint inhibitors in extracts from marine invertebrates and their associated microorganisms, both of which are rich sources of chemically-diverse bioactive compounds. Screening of 1300 extracts by Dr. Roberge's group from marine organisms at two different dilutions yielded 11 samples with significant activity. Extracts of one of the bacterial isolates, clinl 116, exhibited strong inhibition of the G2 cell cycle checkpoint. Clinl 116 was therefore chosen as a candidate for further chemical studies in order to identify the bioactive metabolites produced by this microorganism. This bacterial isolate, clinl 116, was obtained from the surface of a Northeastern Pacific Ocean sponge. It is an actinomycete by appearance and gram stain, but has not been identified to specific genus and species. 133 134 3.2.1. Isolation of active metabolites from bacterial extract cl inll l6. Marine microorganisms were isolated from the invertebrates on site using various marine culture media. The pure marine isolate clinl 116 was grown as lawns on trays of tryptic soy agar (24cm long, 37cm wide, 0.5cm deep agar) supplemented with NaCl to a concentration of 1%. It was grown in moderate scale culture in order to generate sufficient quantities of the secondary metabolites required for chemical identification of the active components. Ten 400 ml trays (24cm x 37cm x 0.5cm deep agar) were cultured for five days at 16° C after which the cultures were harvested. The combined agar and bacterial cells from the 10 trays of microorganism cultures (cells and agar medium, approximately 200 g) were exhaustively extracted with three 600 ml portions of MeOH that were combined, filtered, and reduced in vacuo to give a brown/gray gummy residue. A small amount of the extract was dissolved in DMSO for the G2 checkpoint inhibition assay. The active crude extract was fractionated by a modified Kupchan partitioning scheme. This involved partitioning the crude extract between organic solvents of increasing polarity and an aqueous methanol solution of varying concentrations.1 The extract was dissolved in 500 ml of MeOH/H 2 0 (1:4) and sequentially extracted with hexane (3x200 ml), chloroform (3x200 ml), and ethyl acetate (3x200 ml). The fractions obtained after this solvent partitioning were again tested in the G2 assay. The chloroform fraction, which was found to be active, was further fractionated by the sequential application of size exclusion chromatography, silica gel flash chromatography (eluent: ethyl acetate/MeOH, 90:10), and reversed phase HPLC (eluent: H 20/MeOH/(iPr) 2NH 134 135 30:70:0.005). Throughout this fractionation procedure fractions were also screened by 'H NMR in order to identify NMR signals suggestive of novel chemistry. The bioassay guided isolation procedure yielded one pure active compound (5.2 mg) identified as staurosporine (10), which is one of the few known G2 checkpoint inhibitors. Staurosporine was confirmed by analysis of its MS and NMR data and comparison with published values.2 This provided a validation for the G2 bioassay. 135 Figure 3.1: Isolation scheme of staurosporine from bacteria isolate clinl 116. MeOH extract 1. Concentrated 2. Diluted with water 3. Partitioned between aq. and org. solvents Hexane fraction CHC1 3 fraction EtOAc fraction Water fraction Sephadex LH-20 (MeOH) A B C D Silica gel EtOAc/MeOH (1:9) K L M N O R.P. HPLC MeOH/H20/base X Staurosporine 136 137 3.2.2: The origin of two semisynthetic analogues: We also isolated two closely related compounds identified by MS and NMR as the oxazolidone 17 and the carbamate 49 derivatives of staurosporine (10). The compounds were not present in the original extract. While carrying out 'H NMR analysis on one of the samples of staurosporine in C D C I 3 , it was observed that over time a series of new minor signals appeared in the spectrum. Analysis of this aged NMR sample by T L C revealed the presence of a new compound that was not present in the original sample when it was added to the NMR solvent. This transformation product was isolated by reversed phase HPLC (eluent: H20/MeOH/diisopropylamine 30:70:0.005) and shown by MS and NMR analysis to be the oxazolidone derivative 17. It appeared likely that the CDC1 3 used in the NMR experiment had partially decomposed to generate phosgene, which in turn had reacted with staurosporine to give the new derivative 17. In order to confirm this supposition, pure staurosporine (50 mg) was dissolved in a mixed solvent of CHC1 3 (20 ml) and MeOH (5 ml) and phosgene gas was bubbled through the solution for 2 minutes. The reaction mixture was purified using silica gel flash chromatography (eluent: EtOAc/MeOH 90:10) and reversed phase HPLC (eluent #1: MeOH/H 2 0 7:3; eluent #2: MeOH/H 20/(iPr) 2NH 55:45:0.005) to give the pure oxazolidone derivative 17 (40% yield), the carbamate 49 (6% yield), the aglycone 50 (12% yield), and unreacted staurosporine (10). The structure of the new carbamate and the known aglycone 50 were identified by analysis of their MS and NMR data and in the case of aglycone 50 comparison with literature values." Reaction of staurosporine with phosgene to give the 137 oxazolidone derivative represents a new type of transformation of the staurosporine structure to give bioactive derivatives. 138 139 3.2.3. Structure elucidation of two semisynthetic analogues of staurosporine: Oxazolidone derivative 17 The oxazolidone derivative 17 was isolated as a white amorphous solid, with the molecular formula determined from the parent ion in the HREIMS (Figure 3.2). This compound gave a very intense M + peak at m/z 478.16362 corresponding to a molecular formula of C 2 8 H 2 2 O 4 N 4 (AM l.Oppm), requiring 20 degrees of unsaturation. It contained two more degrees of unsaturation than staurosporine (10), resulting from the loss of four hydrogen atoms and the addition of one oxygen atom. The UV spectrum of 17, which was indistinguishable from that of staurosporine, suggested the presence of an indolo[2,3-a]pyrrolo[3,4-c]carbazole-5(6H)-one chromophore.2"5 The 'H and 1 3 C NMR spectra obtained for the oxazolidone derivative were also similar to those of staurosporine. Comparison of NMR data revealed the presence of the same indolocarbozole core. Some differences were observed in the chemical shifts for the sugar moiety signals. In the 'H NMR spectrum of the oxazolidone derivative, the H3'(S 5.31), H4'(8 4.34), H6'(8 6.96) and N - C H 3 (8 2.58) signals were observed at lower field H o 17 139 o o 141 than those of staurosporine. Two geminal H5' protons (8 2.91, 1.97) were split away from each other by 0.94 ppm while those geminal protons in staurosporine had almost the same chemical shifts at 2.50 ppm. One additional carbonyl signal was observed at 8 155.6 (C7'), and the 3'-OMe signal was lost in the 1 3 C spectrum. A HMBC correlation was observed between Me-9' (8 2.58) and C7'(8 155.6), which indicated that an amide bond was formed between a nitrogen and the carbonyl (C7') group. The 'H chemical shift of H3' (8 5.31) suggested that an ester bond was connected to C3'. Therefore, an oxazolidone ring existed between C3' (8 75.4) and C4' (8 52.0) (Figure 3.3). One degree of unsaturation for the oxazolidone ring and one for the carbonyl group in the ring accounted for the two additional degrees of unsaturation relative to staurosporine. Figure 3.3: Selected NOE (left) and HMBC (right) correlations. Detailed analysis of the COSY, HMQC and HMBC data confirmed that the rest of the structure was identical with that of staurosporine. The COSY correlations observed between H3' (8 5.31) and H4' (8 4.34), H4' and H5' (8 1.97, 2.91), H5' and H6'(8 6.96), indicated that the sugar moiety still existed. Connectivity of the sugar moiety and aglcone was confirmed by NOE difference spectra. Irradiation of Me-8' (8 2.03) enhanced the intensity of H3' (8 5.31) and HI 1 (8 8.06). In addition, a weak NOE between HI (8 7.78) O H M B C Correlation 141 142 and H6'(8 6.96) /H2(5 7.52) was observed. These NOE data confirmed the attachment of C2' to N12 and C6' to N13, as in staurosporine. 142 H (ppm) 8 7 6 5 4 3 2 1 Figure 3.7 2D H M Q C spectrum of oxazolidone derivative 17 (500MHz, DMSO-d^). 146 148 Table 3.1:NMR data for oxazolidone derivative 17 recorded in DMSO-d 6 500MHz ('H). c# 5 l3C(ppm) 5 'H(ppm) a C O S Y Correlation b H M B C Correlation a N O E - D I F F 1 108.6 7.78(d,8.2Hz) H2 . . . H2,H6' 2 125.4 7.52 (dd, 7.6,8.2Hz) H 1 , H 3 H4 — 3 119.5 7.30 (dd,7.6,7.9Hz) H 2 , H 4 HI — 4 125.4 9.22(d,7.9Hz) H3 H2 . . . 4a 122.3 — . . . H1.H3 . . . 4b 115.4 — . . . H4 . . . 4c 132.9 — — H 6 . . . 5 171.5 — . . . H 6 . . . NH(6) — 8.63(s) H7 . . . . . . 7 45.2 4.98(d,6.3Hz) H 6 . . . . . . 7a 120.2 — . . . H6 . . . 7b 1 15.6 . . . — . . . . . . 7c 124.9 . . . — H8 _ _ _ 8 121.2 8.03(d, 7.6Hz) H9 . . . _ _ _ 9 120.9 7.38(dd, 7.6,7.6Hz) H8 ,H10 H l l . . . 10 124.6 7.51(dd, 7.6, 8.4Hz) H9.H11 . . . 1 1 1 16.6 8.06(d, 8.4Hz) H10 H9.H10 . . . 1 la 140.3 . . . — H8 . . . 12a 125.7 . . . . . . . . . . . . 12b 128.6 . . . — . . . . . . 13a 136.4 . . . — H4 . . . 2' 92.5 — — H 8 ' . . . 3' 75.4 5.3l(d, 8.7Hz) H 4 ' H 8 ' H8 ' 4' 52.0 4.34(m) H3\H5'a / | j H 9 ' H 3 ' 5'oc 45.2 1.97(m) H 4 ' , H 6 ' , H 5 ' p . . . . . . 5'P 2.91(m) H4' ,H6 ' ,H5 'a — . . . 6' 79.1 6.96 (dd, 6.4, 9.8Hz) H5'a/(3 . . . HI T 155.6 . . . . . . H 9 ' . . . 8' 29.5 2.03(s) . . . . . . H 3 \ H 1 1 9' 28.7 2.58(s) . . . . . . — a. Correlated to proton resonance in 5 H column. b. Correlated to carbon resonance in 8 I 3 C column. 148 149 Carbamate derivative 49 H 49 O — The carbamate derivative 49 was isolated as a white amorphous solid. The molecular formula of the carbamate derivative was determined from the parent ion in the HRFABMS (Figure 3.9). This compound gave a very intense peak [M+H]+ at m/z 525.21235 corresponding to a molecular formula of C 3 0 H 2 8 O 5 N 4 (AM -2.75ppm), requiring 19 degrees of unsaturation. It contained one more degree of unsaturation than staurosporine (10) resulting from the addition of a C2H2O2 unit. The UV spectrum of the carbamate derivative, which was indistinguishable from that of staurosporine, suggested the presence of an indolo[2,3-a]pyrrolo[3,4-c]carbazole-5(6H)-one chromophore.2"5 The H and " C NMR spectra obtained for the carbamate were also similar to those of staurosporine. Comparison of the NMR data revealed the presence of the same indolocarbozole core. Some differences were observed in the chemical shifts assigned to protons in the sugar moiety. In the 'H NMR spectrum of the carbamate, the H3'(§ 4.28), H4'(6 4.65), H6'(6 7.01), and N - C H 3 (5 2.69) signals were observed at lower field than the corresponding resonances in the staurosporine. Two geminal H5' protons (8 2.63, 2.25) were split away from each other by 0.38 ppm while those geminal protons in staurosporine had almost the same chemical shifts at ~ 2.50 ppm. An additional -OMe 149 o </-> <r, * 9-CN C7\ (0 I oo °o -a g 2 - 8 « - S o ? i < 3 • ? » <—i > CN < in — J S M 5 52 GO g §g CO + c cu . 5 ctf J2 J •a o • o L S Ls 9 o . o cs =a IN i - 8 L 8 00 U _> CO > u <u 3 c3 X ) O o 1) a. 3 S C o 2 o O N s 6X1 o CO 151 signal was observed at 8 3.71 ppm and 52.8 ppm in the 'H and 1 3 C NMR spectrum, respectively. One additional carbonyl signal (C7\ 8 155.6) was present in the l 3 C NMR spectrum. A H M B C correlation was observed between Me-9' (8 2.69) and C7' (8 156.8). It indicated that an amide bond existed between the nitrogen and the carbonyl (C7') group. The O C H 3 (10') still existed because a three bond correlation was observed between H3'(8 4.28) and O C H 3 (10', 60.3 ppm). Therefore, one extra O C H 3 (11', 55.6 ppm) must be connected to the carbonyl (C7') group, indicating that a carbamate was attached to the amino group of the sugar moiety (Fig. 3.10). The degree of unsaturation in the carbamate group satisfied the requirement of one additional degree of unsaturation than that of staurosporine. Figure 3.10: Selected NOE (left) and HMBC (right) correlations. Detailed analysis of the COSY, HMQC and HMBC data confirmed that the rest of the structure was consistent with that of staurosporine. The HMBC correlations observed between H3' (8 4.28) and C4'(8 50.3) / C5' (8 29.1); H5'(8 2.25, 2.63) and C47C6' (8 82.2); H6' (8 7.01) and C2' (8 94.6), indicated the sugar moiety still existed. Connectivity of the sugar moiety and aglcone was confirmed by NOE difference and H M B C spectra. 151 152 Irradiation of Me-8' (5 2.35) enhanced the intensities of H3' (5 4.28) and HI 1 (5 8.00). Also H M B C correlation between H6'(S 7.01) and C12b (8 125.4) was observed. The NOE and H M B C data confirmed the attachment of C2' to N12 and C6' to Nl 3. The proposed mechanisms for the formation of the oxazolidone derivative 17 and the carbamate 49 are shown below. The nucleophilic amino group of 10 underwent addition at the carbonyl group of phosgene, followed by elimination of the chloride to give the intermediate chloroformamide. The methoxyl group of 10 reacted as a second nucleophile to attack the intermediate chloroformamide, followed by elimination of the methyl chloride to give 17. When the solvent methanol reacted as a second nucleophile to attack the intermediate chloroformamide, followed by elimination of the hydrogen chloride, the carbamate 49 was formed. 152 153 157 158 159 Table 3.2: NMR data for carbamate derivative 49 recorded in DMSO-d 6 500MHz ('H). c# 5 1 3 C(ppm) 5 'H(ppm) a C O S Y Correlation b H M B C Correlation " N O E - D I F F 1 109.0 7.64(d,8.2Hz) H2 H3 . . . 2 125.4 7.48 (dd,7.5, 8.2Hz) H 1 . H 3 H4 . . . 3 119.5 7.29 (dd,7.5, 8.0Hz) H 2 , H 4 HI — 4 125.7 9.27(d, 8.0Hz) H3 H2 . . . 4a 122.6 . . . . . . H 1 , H 3 . . . 4b 1 15.1 . . . . . . H4 . . . 4c 132.6 . . . — H 6 , H 7 . . . 5 171.9 _._ . . . H6,H7 ___ NH(6) . . . 8.55(s) H7 . . . . . . 7 45.4 4.98(s) H 6 H 6 — 7a 119.4 . . . . . . H 6 , H 7 , H 7 ' ___ 7b 114.1 . . . — H8 . . . 7c 123.8 — — H 9 . H U . . . 8 121.5 8.05(d, 7.6Hz) H 9 H10 9 120.4 7.35(dd, 7.6,7.5Hz) H8,H10 HI 1 . . . 10 125.0 7.49( dd, 7.5, 8.4Hz) H9.H11 H8 — 11 113.5 8.00(d, 8.4Hz) H10 H9 . . . 11a 138.7 . . . H8,H10 . . . 12a 129.2 . . . — . . . ___ 12b 125.4 . . . — H6' 13a 136.2 . . . — H2, H4 . . . 2' 94.6 . . . — H 6 ' , H 8 ' . . . 3' 83.6 4.28(s) . . . H8' ,H10 ' . . . 4' 50.3 4.65(bm) . . . H 3 ' , H 5 ' a / p \ H 9 ' . . . 5'cx 29.1 2.25(m) H 5 ' P H 3 ' . _ _ 5'(3 2.63(m) H 5 ' a . . . 6' 82.2 7.01(bm) H5'oc/p H 5 ' a . . . 7' 156.8 . . . . . . H 9 ' . . . 8' 29.1 2.35(s) . . . . . . H3',H11 9' 29.9 2.69(s) . . . . . . 10' 60.3 2.69(s) . . . H 3 ' 11' 52.8 3.71(s) . . . . . . a. Correlated to proton resonance in 5 H column. b. Correlated to carbon resonance in 5 I 3 C column. 159 160 3.2.4 Biological activity: The G2 checkpoint inhibition and the cytotoxicity of staurosporine and related compounds were measured. The known metabolites staurosporine and UCN-01 showed strong G2 checkpoint inhibition in Michel Roberge's assay, with the most effective concentrations of 0.2± 0.2 nM and 6±2 nM, respectively (Fig. 3.16). The oxazolidone derivative had the most effective concentration of 360+.20 nM, and showed less activity at its optimal concentration (Figure 3.16). Interestingly, the carbamate derivatives showed no detectable G2 checkpoint inhibition, even at the highest concentration tested (0.2 mM, Figure 3.16), yet retained cytotoxicity (IC5 0= 14±1 uM). Staurosporine, UCN-01 and the oxazolidone derivative were also significantly cytotoxic when used alone (IC50 of 60±20 nM, 80±10 nM and 40+.10 uM, respectively). \ Q \° »P \° Drug(M)_ Figure 3.16: G2 checkpoint inhibition by staurosporine and derivatives. G2 checkpoint inhibition by staurosporine (O), UCN-01 (•), the oxazolidone derivative of staurosporine (A) or the carbamate derivative of staurosporine (•) is expressed as the ELISA signal, where A405=l corresponds to about 60% mitotic cells. 160 161 3.2.5. Experimental NMR data were collected on either a Bruker AMX500, a Bruker WH400 or a Bruker AM400 spectrometer each equipped with a 5mm probe. All spectra were obtained in DMSO-dg. Proton spectra were referenced using internal residual DMSO-d 6 (§ 2.49) and carbon spectra were referenced to the DMSO methyl carbon resonance (8 39.5). FABMS data were collected on a Kratos Concept IIHQ hybrid mass spectrometer with cesium ion secondary ionization and a magnetic sector mass analyzer. Samples were dissolved in a MeOH-thioglycerol matrix and spectra were obtained using a source voltage of 8 K V and a cesium ion gun voltage of 12 KV. Low and High resolution EI mass spectra were recorded on Kratos MS50/DS55SM mass spectrometer. Reversed-phase and Normal-phase thin layer chromatography (TLC) was performed using Whatman MKC18F and Kieselgel 60 F254 plates. Visualization was detected by U V (X=254 nm) and/or heating after spraying with vanillin reagent. Normal phase column chromatography was carried out either on Merck silica gel G60 (230-400 mesh) or Sigma silica gel (size: 10-40 u). Reversed phase chromatography was performed using reversed-phase silica prepared according to the literature6. Sephadex LH-20 (bead size 25-100 |i) was used for size exclusion chromatography. High performance liquid chromatography (HPLC) separations were done on one of two possible systems using either a Whatman Partisil 10 ODS-3 Magnum column or Rainin Partisil 10-ODS column. The first system consisted of a Waters 600E HPLC pump/system controller with a Waters 486 tunable absorbance detector. The second system consisted of a Waters 600E HPLC pump/system controller equipped either with a Waters 996 photodiode array detector or Rl (Perkin Elmer LC-25) detector. Both systems 161 162 were interfaced with a personal computer using Millenium 2010 chromatography software. The solvents used for extraction and for column chromatographies were Fisher reagent grade. HPLC solvents were Fisher HPLC grade which were filtered and degassed prior to use. All other solvents, reagents and standards were reagent or commercial grade and were used without further purification. Oxazolidone derivative 17: White amorphous solid; HREIMS: [M] + m/z 478.16362 (C28H22N4O4, AM 1 .Oppm); ! H NMR (DMSO-d 6, 500MHz) and l 3 C NMR (DMSO-d 6, 125MHz), and 2D NMR data listed in Table 3.1. Carbamate derivative 49: White amorphous solid; HRFABMS: [M+H]+ m/z 525.21235 (C 3 oH29N 4 0 5 , AM -2.75ppm); *H NMR (DMSO-d 5, 500MHz) and 1 3 C NMR (DMSO-d 6, 125MHz), and 2D NMR data listed in Table 3.2. Staurosporine (10): White amorphous solid; HREIMS: [M] + m/z 466.20145 ( C 2 8 H 2 6 N 4 O 3 , AM 0.45ppm); 'H NMR (DMSO-d 6, 500MHz) §7.55, d, J=8.0Hz(H 1), 7.46, dd, J=7.4, 8.0Hz (H2), 7.28, dd, J=7.4, 8.2Hz (H3), 9.32, d, J=8.2Hz (H4), 8,52, s (H6), 4.95, s (H7), 7.96, d, 7.4 Hz (H8), 7.28, dd, J=7.4, 7.6Hz (H9), 7.42, dd, J=7.6,8.6Hz (H10), 7.98, d, J=8.6Hz (HI 1), 4.03, d, J=6.5Hz (H3'), 3.26, m, (H4'), 2.50, m (H5'), 6.70, dd, J=6.4,6.4Hz (H6'), 2.30, s, (2'Me), 3.34, s, (3'0CH 3), 1.44, s, (N-Me); and 1 3 C NMR (DMSO-d 6, 125MHz)8 108.2(C1), 124.6 (C2), 118.9 (C3), 125.7 (C4), 122.5 (C4a), 113.5 (C4b), 118.8 (C4c), 172.2 (C5), 45.4 (C7), 131.9 (C7a), 114.2 (C7b), 123.8 (C7c), 120.6 (C8), 119.6 (C9), 124.2 (C10), 115.2 (CI 1), 139.6 (CI la), 130.0 (C12a), 126.6 (C12b), 136.4 (C13a), 91.0 (C2'), 82.7 (C3'), 50.0 (C4'), 29.4 (C5'), 79.8 (C6'), 29.8 (2'Me), 57.4 (3'OMe). 162 164 3.3 The Indonesian Sponge Aaptos suberitoides Specimens of Aaptos suberitoides (Bn|)ndsted, 1934) were collected by our group as part of a general collecting expedition to Indonesia. The crude extracts of sample RJA-96-57, later identified as A. suberitoides, were found to have strong G2 cell cycle checkpoint inhibition in preliminary bioassay screening. The sponge was identified by Dr. R. van Soest and a voucher sample has been deposited at the Zoologisch Museum, University of Amsterdam. Sponges of the genus Aaptos belong to the family Suberitidae in the order Hadromerida and class Demospongiae according to Dr. R.van Soest. 3.3.1. Isolation of active metabolites from sponge extract RJA-96-57 Specimens of Aaptos suberitoides were collected by hand using SCUBA in Indonesia in July 1996. Freshly collected sponge was chopped, immersed in MeOH, and soaked at room temperature for 2 days on site in Indonesia. The MeOH was decanted and concentrated in vacuo to give a dark green liquid that was transported to Vancouver over dry ice and stored frozen. In Vancouver, the dark green extract was filtered and concentrated to dryness in vacuo to give a gummy residue. A small amount of the extract was dissolved in DMSO for the G2 checkpoint inhibition assay. The active MeOH extract was also fractionated by a modified Kupchan partitioning scheme. This involved partitioning the crude extract between organic solvents of increasing polarity and an aqueous methanol solution of varying concentrations.1 The extract was dissolved in 500 ml of MeOH/H 2 0 (1:4) and sequentially extracted with hexane (3x200 ml), chloroform (3x200 ml) and ethyl acetate (3x200 ml). The fractions obtained after this solvent partitioning were tested again by the 164 165 G2 assay. The chloroform soluble fraction, which was found to be active, was further fractionated by Sephadex LH-20 (eluent: MeOH) chromatography to give several fractions. Two fractions (3M, 3N) showed G2 activity. The 3M fraction was chromatographed first on Sephadex LH-20 (eluent: EtOAc/MeOH/H 2 0 = 40:10:4) and then on reversed phase HPLC (eluent: MeOH/H 2 0 95:5) to yield the known alkaloids aaptamine (12) and isoaaptamine (50). Fraction 3N was directly chromatographed by normal phase HPLC (eluent: EtOAc/Hex 95:5) to yield the known metabolite demethyl(oxy)aaptamine (51) and the new compound 52. 165 167 Figure 3.18: Isolation scheme of aaptamine and its derivatives from RJA-96-57. MeOH extract 1. Concentrated 2. Diluted with water 3. Partitioned between aq. and org. solvents Hexane fraction CHC1 3 fraction EtOAc fraction Water fraction Sephadex LH-20 (MeOH) 3 L •3M 3 N 3 0 LH20 (EtOAc:MeOH:H 2 Q 40:10:4) NP HPLC EtOAc/Hex (95:5) 18A 18B 18C 18D 18E R.P. HPLC I M e O H / H 2 0 (95:5) Demethyl(oxy) aaptamine Analog 52 < aaptamine isoaaptamine 167 3.3.2. Structure elucidation of aaptamine and its derivatives: Aaptamine 12 Aaptamine (12) was isolated as a yellow solid. The molecular formula of aaptamine was determined from the parent ion in the HRFABMS, which was a very intense [M+H]+peak at m/z 229.09690 corresponding to a molecular formula of C i 3 H i 2 N 2 0 2 (AM, -1.35 ppm). Aaptamine (12) was identified by extensive NMR spectroscopy and mass spectrometric analysis. Its structure was ultimately confirmed by comparison with published data.7'8 The 'H and l 3 C NMR spectra are shown on the next page. 168 ON v o co co vo VO I o 00 O 00 Q Q o o m CN" o o CN C PH cd cd o u PH C/3 ON CO CU s-3 C O O H C O " ' Z D CD C iS PH cd cd o PH Pi ON CO CU •-fl DC ON V D CO CO 1 7 0 Isoaaptamine 10 O C H 3 H C K 9 H 3 C 9aN 16a 50 Isoaaptamine (50) was isolated as a yellow solid. The molecular formula of isoaaptamiene was determined from the parent ion in the HRFABMS, where it gave a very intense [M+H]+ peak at m/z 229.09760 corresponding to a molecular formula of C 1 3 H 1 2 N 2 O 2 (AM, -0.47 ppm). Isoaaptamine (50) is an isomer of aaptamine (12). The methyl group from the C9-OMe substituent is shifted to N1. It was also identified by extensive NMR spectroscopy and mass spectrometric analysis. Its structure was also ultimately confirmed by comparison with published data.7'8 The 'H and l 3 C NMR spectra are shown on the next page. 1 7 0 172 Demethyl(oxy)aaptamine 51 Demethyl(oxy)aaptamine (51) was isolated as a yellow solid. The molecular formula of demethyl(oxy)aaptamine (51) was determined from the parent ion in the HREIMS, where it gave a very intense [M] + peak at m/z 212.05872 corresponding to a molecular formula of C 1 2 H 8 N 2 O 2 (AM, -0.7 ppm). It was also identified by extensive NMR spectroscopy and mass spectrometric analysis. Its structure was also ultimately confirmed by comparison with published data.7'8 The 'H and l 3 C NMR spectra are shown on the next page. 172 o to O co Q N o o o o < n s a CO «£), <L> C E CJ X -g E to T 3 4 -o E 3 cx 3 t7> to CO to to CM LO O H cJ X £ T D <+-O P 3 O C X CO u L . s CM L O 174 Demethyl(oxy)aaptamine analog 52 10 ii O C H 2 O C H 3 4 52 Demethyl(oxy)aaptamine analog 52 was isolated as a yellow solid. The molecular formula of 52 was determined from the parent ion in the HREIMS, where it gave a very intense [M] + peak at m/z 242.06914 corresponding to a molecular formula of C 1 3 H 1 0 N 2 O 3 (AM, 0.1 ppm). This molecular formula differed from that of demethyl(oxy)aaptamine (51) by the addition of C H 2 0 . The 'Hand l 3 C NMR spect rum of 52 were similar to those of 51 except that the 'H NMR spectra of 52 contained an extra methylene proton resonance at 8 5.37 ppm. This methylene resonance was correlated to a 1 3 C signal at 8 94.6 ppm in the HMQC spectrum. It suggested that this methylene was a methylenedioxyl (OCH 20) group. The observed H M B C between this methylene resonance H10 (8 5.37) and C8 (8 153.3) / CI 1(8 56.8) confirmed that methylenedioxyl group was connected to both C8 and Me-11. Detailed analysis of the COSY, HMQC and HMBC data confirmed that the rest of structure was consistent with the demethyl-(oxy)aaptamine (51). This is the only new analogue isolated from this species. It contains an acetal protecting group which has never been found in a natural product. This compound 52 could be an artifact of the original extraction process in Indonesia. 174 m c--177 10 11 O C H 2 O C H 3 J _ r - - o - . . r et 0 y m C • • L 5 ( f i b . . mv • • • 0 O • • as*, r * *. : 0 0 e «•» 0 i-! t —i 1 ' 1 1 1 1 . 1 —] . • 1 i i . ! , . . 1 , 1 1 1 8.00 7.00 6.00 5.00 4.00 3.00 (ppm; Figure 3.24: 2D H M B C spectrum of 52 (500MHz, DMS0-d 6). 177 178 10 11 O C H 2 O C H 3 J I L 52 (ppm) - 1 6 0 (ppm) Figure 3.25: 2D HMQC spectrum of 52 (500MHz, DMSO-d 6). 178 179 Figure 3.26. 2D COSY spectrum of 52 (500MHz, DMSO-d 6). 179 180 Table 3.3 NMR data for 52 recorded in DMSO-d6 500MHz ('H). N O . ' H ppm (J,Hz) 51 52 l 3 C ppm (J,Hz) 51 52 52 a C O S Y Correlation 52 b H M B C Correlation 2 9.10 9.12 (d, 4.44) (d, 5.64) 148.82 149.155 H3 H3 3 8.21 8.25 (d,4.44) (d, 5.64) 126.41 126.79 H2 H2 3a 147.41 147.81 5 9.13 9.17 (d,3.32) (d, 4.40) 157.31 157.72 H 6 6 7.74 7.81 (d,3.32) (d, 4.40) 122.18 123.01 H5 H7 6a . . . . 136.38 136.41 H5 7 7.16(s) 7.28(s) 108.87 112.79 . . . H 6 8 155.68 153.33 — H 7 , H10 9 176.98 177.12 . . . H7 9a . . . . 148.80 149.24 — H2 9b 117.67 119.50 . . . H3, H7, H 6 10 3.90(s) 5.37(s) 56.02 94.66 — H I 1 11 3.45(s) 56.75 — H10 a. Correlated to proton resonance in 5 H column. b. Correlated to carbon resonance in 8 I 3 C column. 180 181 3.3.3 Biological activity: Aaptamine alkaloids possess many potent pharmacological activities. They were originally found in a screen for a-adrenoceptor blocking activity on vascular smooth muscle.9 In addition, they exhibited cytotoxicity to HeLa cells, human KB 16, A549, and HT-29 tumor cells,10 and they also showed potent antimicrobial activity against Gram-positive and Gram-negative bacteria such as Staphylococcus aureus and Proteus vulgaris.1 It has now been found that those alkaloids have G2 cell cycle checkpoint inhibition activity. It was measured by Lynette Lim in the Roberge's lab. Aaptamine (12) showed the most powerful G2 checkpoint inhibition with a I C 5 0 of 0.44 u M . Demethyl(oxy)aaptamine (51) had a higher IC50 of 4.71 u M . However, isoaaptamine (50) and demethyl(oxy)aaptamine analog 52 showed no detectable G2 checkpoint inhibition. Unfortunately, their numerous pharmacological effects preclude their effectiveness in chemotherapy as is the case for staurosporine. One of their drawbacks is that they induce premature entry into mitosis in both normal and tumor cells causing non-selective cell death. 181 182 3.3.4. Experimental NMR data were collected on either a Bruker AMX500, a Bruker WH400 or a Bruker AM400 spectrometer each equipped with a 5mm probe. All spectra were obtained in DMSO -d6. Proton spectra were referenced using internal residual DMSO -d6 (§ 2.49) and carbon spectra were referenced to the DMSO methyl carbon resonance (5 39.5). FABMS data were collected on a Kratos Concept IIHQ hybrid mass spectrometer with cesium ion secondary ionization and a magnetic sector mass analyzer. Samples were dissolved in a MeOH-thioglycerol matrix and spectra were obtained using a source voltage of 8 K V and a cesium ion gun voltage of 12 KV. Low and High resolution EI mass spectra were recorded on Kratos MS50/DS55SM mass spectrometer. Reversed-phase and Normal-phase thin layer chromatography (TLC) was performed using Whatman MKC18F and Kieselgel 60 F254 plates. Visualization was detected by UV (^ =254 nm) and/or heating after spraying with vanillin reagent. Normal phase column chromatography was carried out either on Merck silica gel G60 (230-400 mesh) or Sigma silica gel (size: 10-40 Reversed phase chromatography was performed using reversed-phase silica prepared according to literature.6 Sephadex LH-20 (bead size 25-100 p„) was used for size exclusion chromatography. High performance liquid chromatography (HPLC) separations were done on one of two possible systems using either a Whatman Partisil 10 ODS-3 Magnum column or Rainin Partisil 10-ODS column. The first system consisted of a Waters 600E HPLC pump/system controller with a Waters 486 tunable absorbance detector. The second system consisted of a Waters 600E HPLC pump/system controller equipped either with a Waters 996 photodiode array detector or RI (Perkin Elmer LC-25) detector. Both systems 182 183 were interfaced with a personal computer using Millenium 2010 chromatography software. The solvents used for extraction and for column chromatographies were Fisher reagent grade. HPLC solvents were Fisher HPLC grade which were filtered and degassed prior to use. All other solvents, reagents and standards were reagent or commercial grade and were used without further purification. Demethyl(oxy)aaptamine analog 52: yellow amorphous solid; HREIMS: [M] + m/z 242.06914 (C l 3H,oN 203, AM O.lppm); 'H NMR (DMSO-d 6, 500MHz) and 1 3 C NMR (DMSO-d 6, 125MHz), and 2D NMR data listed in Table 3.3. 183 184 3.4 Diterpene lactones from South African plant Parinari curatellifolia More than 80,000 extracts from the National Cancer Institute's open repository of natural product extracts were screened in the Roberge's lab by the G2 checkpoint inhibitor assay resulting in several hits. One of active extracts was from the Chrysobalanaceae species Parinari curatellifolia. The roots of this tropical bush are used by traditional healers in Zimbabwe to treat delirium.11 Therefore, this tropical plant was chosen as a candidate for chemical studies in order to identify the bioactive metabolites. 3.4.1 Isolation of active metabolites from plant Parinari curatellifolia. A concentrated dark brown methanol extract of the roots of P. curatellifolia was received from NCI. This extract was initially fractionated by a modified Kupchan partitioning scheme. The extract was dissolved in 500 ml of MeOH/H 2 0 (1:4) and sequentially extracted with hexane (3x200 ml), chloroform (3x200 ml), and EtOAc (3x200 ml). The fractions obtained after this solvent partitioning were tested in the G2 assay. Both the hexane and chloroform soluble fractions were active. The hexane soluble material was further fractionated by sequential application of silica gel flash chromatography (eluent: EtOAc/Hex, 20:80) and reversed phase HPLC (eluent: H 2 0/MeOH 20:80). This bioassay guided isolation procedure yielded two pure active compounds. One was the known diterpene lactone 19 and the other was a new derivative, the diterpene ester 53. The chloroform fraction was chromatographed by silica gel flash normal phase column (eluent: EtOAc/Hex, 20:80) to yield two known derivatives, the diterpene lactones 54 and 55. Both of these compounds were also active in the G2 bioassay. 184 185 MeOH extract 1. Concentrated 2. Diluted with water 3. Partitioned between aq. and org. solvents Hexane fraction CHC13 fraction EtOAc fraction Water fraction 1. NP silica gel (EtOAc/Hex 20:80] 2. R.P. HPLC MeOH/H 20 (80:20) NP silica (EtOAc/Hex: 20:80) 19 53 54 55 Figure 3.27: Isolation scheme of diterpene lactone derivatives from P. curatellifolia. 185 186 3.4.2 Structure elucidation of the diterpene lactone derivatives: The structures of diterpene lactone 19 and its derivatives 54 and 55 were solved I 13 by analysis of their H and C NMR and mass spectrometric data, and were confirmed by comparison of their NMR data with literature values.12'13 The new derivative 53 was identified by the extensive analysis of its 2D NMR data, mass spectrometric data, and comparison of its NMR data with the NMR data for the known diterpene lactone 19. The details of the structure elucidation of 53 are presented below. Diterpene ester 53 Diterpene ester 53 was isolated as a white solid. The molecular formula of diterpene ester 53 was determined from the parent ion in the HREIMS. It gave a very intense peak [M] + peak at m/z 346.01256 corresponding to a molecular formula of C 2 1 H 3 0 O 4 (AM, 0.47). This molecular formula differed from that of diterpene lactone 54 by the addition of two hydrogen atoms. The 'H and 1 3 C NMR spectra obtained for 53 were similar to those of 54. However, the observed HMBC correlation between a methyoxyl resonance (H21, 5 • 186 187 3.63) and a carbonyl resonance (C18, 8 178.5) indicated that the methyoxyl group was not connected to CI 3 (8 37.2). Instead, the correlation was consistent with a methyl ester at C4 (8 44.0). This was confirmed by the HMBC correlation between Me-19 (8 1.18) and CI 8 (8 178.5). The 1 3 C chemical shift of the aliphatic quaternary CIO (8 77.0) suggested that CIO was tertiary carbinol. The rest of the structure was the same as that of diterpene lactone 19 as confirmed by detailed analysis of the COSY, HMQC, and H M B C data. Diterpene ester 53 is possibly an artifact formed by hydrolysis of a lactone ring and esterification by the methanol extraction solvent. We attempted to hydrolyze 54 under both acidic and basic conditions, however, lactone opening did not occur. Instead, Michael addition occurred at the a, (3 unsaturated ketone. Therefore, diterpene ester 53 appears not to be an artifact of the extraction process. HO Figure 3. 28: Selected HMBC correlation in compound 53. 187 oo oo 1 ? • Ul 5 N D 2 6 2 • S O > < U J " 5 2 t I S M , . _ Is 3 » R § r5 s §> s s s s ro IT) l— <D t/3 CO U B CD & CJ s s o CO CL, Vi t/3 Vi ta c o o ON ro U Si fa 0 0 oo 190 Table 3.4. NMR data for 53 recorded in CDC1 3 at 500 MHz ( ]H). c# 8 1 3C(ppm) 5 'H(ppm) a COSY Correlation b H M B C Correlation 1 31.5 2.12 HI',H2,H2' . . . r 1.80 H1,H2,H2' 2 19.9 1.92 H1,HI',H2',H3 HI 2' 1.78 H1,H1',H2,H3,H3' 3 37.8 2.15 H2,H2',H3' HI, H19 3' 1.02 H2,H2',H3 4 44.0 — — H19.H2' 5 49.2 1.80 H6,H6' H6, H19 6 19.0 1.80 H5,H6' 6' 1.48 H5,H6 7 29.3 2.20 H6,H6',H7' 7' 1.10 H6,H6',H7 8 56.0 — . . . H7,H13 9 44.0 . . . . . . H20 10 77.1 — — H1,H20, 11 33.3 1.85 Hir ,H12,H12' H13 11' 1.78 HI 1.H12 12 37.5 1.58 HI 1,H12',H13 H14.H13 12' 3.02 HI 1.H12 13 37.2 2.60 HI2 H14 14 29.0 1.40 H14' 14' 1.20 H14 15 212.5 — . . . H13,H17,H17' 16 148.5 — . . . H13.H17 17 1 15.4 5.96 H17' 17' 5.30 H17 18 178.5 — . . . H19.H21 19 29.4 1.18 . . . 20 17.2 1.00 . . . 21 50.5 3.63 . . . a. Correlated to proton resonance in 5'H column. b. Correlated to carbon resonance in 8 I 3 C column. 192 193 3.4.3 Biological activity The Parinari. diterpene lactones seem to have a broad spectrum of biological activities, both against microorganisms such as Cladosporium cucumerinum, and against This is the first time that kaurene diterpenoids have been found to exhibit strong G2 checkpoint inhibition. Diterpene lactone 19 and its derivatives (53-55) showed comparable G2 checkpoint inhibition. Natalie Rundae in Roberge's group explored the relationship between the level of inhibition of G2 arrest and the concentration of lactone 55 (Figure 3.33). Cells overcame G2 arrest when the compound was present at concentrations between 5 and 15 uM. At its most effective concentration (10 uM), it caused 17% of the cells to enter mitosis. When arrested cells were exposed to concentrations higher than 10 uM, the level of entry into mitosis was progressively reduced because most cells died. Lactone 55 causes two distinct effects on the cell cycle: it relieves cells from G2 arrest, and causes an accumulation of cells in mitosis. It acts in the low micromolar range, a potency that compares favorably with debromohymenialdisine (DBH, 15) and isogranulatimide (IGR, 16). a panel of cultured human cell cancer lines. 13 LO o LO CO Cone, of 55 (uM) Figure 3.33: Effects of lactone 55 on cells arrested in G2. 193 194 Inhibitors of the DNA damage checkpoint are sought because they force cells into mitosis before DNA damage can be repaired, eventually resulting in death of the cell. In the clinical setting, checkpoint inhibitors could enhance the selective killing of cancer cells by radiation, while leaving non-radiated normal cells unharmed. The cytotoxicity of lactone 55 to non-irradiated and irradiated cells was measured using the M T T assay after 1-2 days incubation with lactone 55. It was cytotoxic to MCF-7 psY cells with an IC50 of 0.73 uM, and was cytotoxic to y-irradiated MCF-7 p53" cells with an I C 5 0 of 0.40 u M . 1 4 Table 3.5: Interactions between the checkpoint inhibiting effects of lactone 55 and IGR (16), DBH (15), UCN-01 (11), and caffeine (8). 1 4 Entry into Mitosis (%) 55 (lOuM) IGR (20uM) D B H (40uJVI) UCN-01 (0.1 uM) Caffeine (2mM) No 55 17.5 28 32 67 75 +55 (10 nM) 34.8 48 55 68 73 Interactive effects were also studied between lactone 55 and the other checkpoint inhibitors caffeine (8), UCN-01 (11), DBH (15), and IGR (16). Irradiated MCF-7 p53" cells were treated with caffeine, UCN-01, DBH, or IGR, at the most effective concentrations for each compound. The number of cells that were in mitosis after 8h of the treatment was counted and compared to the responses of cells that received lactone 55 in addition to each individual compound (Table 3.5). The most effective inhibitors were caffeine and UCN-01, which released 60-70% of cells from arrest. DBH (15) and IGR (16) each induced a more moderate response, between 20 and 30% of cells exited from G2. The effects of caffeine or UCN-01 were unchanged by lactone 55. DBH (15) and IGR (16), however, caused a much stronger checkpoint inhibition in the presence of lactone 55. The increases in checkpoint inhibition of DBH and IGR were equal to that caused by lactone . 194 195 55 on its own, indicating that effects of lactone 55 are additive with those of DBH (15) and IGR (16). The checkpoint inhibitors discovered so far fall into two main groups: kinase inhibitors (e.g. cafeine and UCN-01) and phosphatase inhibitors (e.g. okadaic acid). Lactone 55 belongs to neither of these groups. The activity of two kinases in the G2 checkpoint pathway, Chkl and A T M , were not affected by the lactone 53. Roberge's group has also tested compound 55's ability to inhibit dephosphorylation of a phosphoserine-phosphothreonine peptide, and found that phosphatase activity was not inhibited. The DNA damaged cells escaped G2 arrest and became blocked in mitosis phase when treated with lactone 55. That indicated the compound also blocks or delays exit from mitosis in cells. Therefore, lactone 55 is both a G2 checkpoint inhibitor and an agent-that causes mitotic arrest. Other known checkpoint inhibitors do not have this feature, indicating that lactone 55 could act on a different target of G2 checkpoint pathway. OH OH ,OH 'SH 55A A reactive structural feature of lactone 55 is the presence of an unsaturated a,p -unsaturated carbonyl group at CI6. This functionality makes it easily to have a Michael 195 196 addition reaction, resulting in a covalent bond between 55 and its target. To test whether a Michael addition is required for the action of 55, it was reacted with mercaptoethanol (HSCH 2 CH 2 OH) to form the adduct 55A. Adduct 55A's ability to inhibit the G2 checkpoint was tested and it was found that 55A had lost all activity, suggesting that the biological activity of lactone 55 is attributed to the structure component: the unsaturated a,(3 - unsaturated carbonyl group. The opening of lactone ring in the structure to give the analog 53, does not affect the activity. This observation indicated that the lactone ring is not crucial for the molecule's activity. By making analogs of the diterpenoids that retain the a,(3 - unsaturated carbonyl group, attached to modified ring structures, it may be possible to widen its therapeutic window. The target of lactone 55 that is relevant to the G2 checkpoint is not yet known. However, this target seems that to be a novel component in the G2 checkpoint pathway for several reasons. The structure of lactone 55 does not resemble that of other checkpoint inhibitors. There is no evidence that it inhibits kinases or phosphatases, which are the typical targets for checkpoint inhibitors. The interactive effects of 55 with other checkpoint inhibitors suggests how the target of this drug may fit into the known checkpoint signaling pathway. The additive increases seen when 55 was combined with DBH (15) or IGR (16) suggest that it acts on a target or pathway separate from these inhibitors. The fact that 55 did not increase the effects of UCN-01 or caffeine indicates that 55 acts downstream of A T M . The lack of checkpoint inhibition by the (3-mercaptoethanol adduct 55A implies that the target of the 55 has a cysteine in its active 196 197 site. In order to characterize the mechanism of action of 55 and reveal new components of G2 checkpoint pathway, we made an isotope-labeled lactone derivative. Compound 55 was reacted with sodium hydride to yield the sodium alkoxide, which was treated with tritium-labeled methyl iodide to yield compound 54A with a tritium labeled methyloxyl group at the C13 position. Because the final product was also the natural product, the isotope labeling reaction did not change its original biological function against the G2 checkpoint. [3H]-54A was fed to irradiated cell extracts and it was found to bind to several proteins. The identification of the target proteins is being investigated. 197 198 3.4.4 Experimental NMR data were collected on either a Bruker AMX500, a Bruker WH400 or a Bruker AM400 spectrometer each equipped with a 5mm probe. All spectra were obtained in CDC1 3. Proton spectra were referenced using internal residual C D C I 3 (8 7.27) and carbon spectra were referenced to the C D C I 3 carbon resonance (8 77.3). Low and High resolution EI mass spectra were recorded on Kratos MS50/DS55SM mass spectrometer. Reversed-phase and Normal-phase thin layer chromatography (TLC) was performed using Whatman MKC18F and Kieselgel 60 F254 plates. Visualization was detected by UV (k=254 nm) and/or heating after spraying with vanillin reagent. Normal phase column chromatography was carried out either on Merck silica gel G60 (230-400 mesh) or Sigma silica gel (size: 10-40 p.). Reversed phase chromatography was performed using reversed-phase silica prepared according to the literature.6 High performance liquid chromatography (HPLC) separations were done on one of two possible systems using either a Whatman Partisil 10 ODS-3 Magnum column or Rainin Partisil 10-ODS column. The first system consisted of a Waters 600E HPLC pump/system controller with a Waters 486 tunable absorbance detector. The second system consisted of a Waters 600E HPLC pump/system controller equipped either with a Waters 996 photodiode array detector or RI (Perkin Elmer LC-25) detector. Both systems were interfaced with a personal computer using Millenium™ 2010 chromatography software. The solvents used for extraction and for column chromatographies were Fisher reagent grade. HPLC solvents were Fisher HPLC grade which were filtered and degassed 198 199 prior to use. All other solvents, reagents and standards were reagent or commercial grade and were used without further purification. Isotope labeling experiment: Five milligram diterpene lactone (55) was dissolved in 5 ml anhydrous tetrahydrofuran. The 0.5 mg sodium hydride was added and stirred for 2 minutes. Then 1 ml of the normal methyl iodide was added and reacted at the room temperature for 16 hours. The samples were then evaporated and redissolved in 500 ul EtOAc. The product (3.5 mg) was compound 54 and was confirmed by comparison of its 'H NMR spectrum and mass spectrum with natural compound 54. When the reaction condition was known, we used 1 ml of tritium-labeled methyl iodide under the same conditions. The tritium-labeled compound was detected by the T L C , where the tritium-labeled products gave the same Rf as the natural compound 54. Diterpene ester 53: a white amorphous solid; HREIMS: [M]+ m/z 346.01256 (C21H30O4, AM 0.47ppm); 'H NMR (CDC13, 500MHz) and 1 3 C NMR (CDC13, 125MHz), and 2D NMR data listed in Table 3.4. 199 200 Reference: 1. Kupchan, S. M. ; Britton, R. W.; Laccadie, J.; Ziegler, M . F.; Sigel, C. W. J. Org. Chem. 1975, 40(5), 648-654. 2. Omura, S.; Iwai, Y.; Hirano, A.; Nakagawa, A.; et al. A new alkaloid AM-2282 of Streptomyces origin. Taxonomy, fermentation, isolation and preliminary characterization. J. Antibiotics. 1997, 30, 275-282. 3. Yasuzawa,, T.; Iida, T.; Yoshida, M.; Hirayama, N.; et. al. The structures of the novel protein kinase c inhibitors K-252a, b, c and d. / . Antibiotics. 1986, 39, 1072-1078. 4. Tanida, S.; Takizawa, M. ; Takahashi, T.; Tsubotani, S.; Harada, S. TAN-999 and TAN-1030A, new indolocarbazole alkaloids with macrophage-activating properties. J. Antibiotics, 1989, 42, 1619-1630. 5. Takahashi, H.; Osada, PL; Uramoto, M. ; Isono, K. A new inhibitor of protein kinase C, RK-286C (4'-demethylamino-4'-hydroxystaurosporine). II. Isolation, physico-chemical properties and structure. J. Antibiotics. 1990, 43, 168-173. 6. Kiisher, T. C ; Lindsten, G. R. J. Org. Chem. 1983, 48, 3589-3591. 7. Nakamura, H..; Kobayashi, J-L; Ohizumi, Y. / . Chem. Soc. Perkin trans. I, 1987, 173-176. 8. Shen, Y.C.; Chein, C.C.; Hsieh, P.W.; Duh, C.Y. / . Fish. Soc. Taiwan. 1997, 24, 117-125. 9. Ohizumi, Y.; Kajuwara, A.; Nakamura, H.; Kobayashi, J. J.Pharm. Pharmacol. 1984, 36, 785-786. 10. Shen, Y.; Lin, T.; Sheu, J.; Duh, C. / . Nat. Prod. 1999, 62, 1264-1267. 11. Gelfand, S; Mavi, R.; Drummond, B.; Ndemera, B. ' The traditional Medical Practitioner in Zimbabwe', Mambo Press, Gweru, 1985. 12. Garo, E.; Maillard, M. ; Hostettmann, K. Helv. Chimi. Acta. 1997, 80, 538-544. 13. Lee, L.A.; Shamon, A.D.; et. al. Chem-Biol. Interact. 1996, 99, 193-195. 14. Natalie Rundae's MS thesis (May, 2000) in Roberge's group. All of biological data was measured by Natalie Rundae. 200 201 Chapter IV: Ingenamine alkaloids . 4.1 Introduction: Recently drug discovery efforts from natural sources have emphasized investigations of the marine environment. This has resulted in the identification of numerous, often highly complex chemical structures.1'2 The widespread application of cytotoxicity directed screening assays has yielded a tremendous diversity of highly toxic compounds which affect numerous targets involved in the eukaryotic cell processes. Potent cytotoxic metabolites may serve as molecular probes for active sites in enzymes. Information on active site interactions between cytotoxic metabolites and enzymes will provide useful knowledge for rational drug design. This portion of my thesis research involved the structure elucidation of new cytotoxic chemical structures from the marine sponge Xestospongia ingens. Isolation of these compounds was guided by a cell-based cytotoxicity bioassay using mouse lymphocytic leukemia cell cultures (P388). Specimens of Xestospongia ingens (Thiele, 1899) were collected as part of a general collecting expedition to Papua New Guinea. The crude extracts of sample PNG95-202, later identified as X. ingens, were found to be cytotoxic against murine leukemia P388 in preliminary bioassay screening. The isolation and structure elucidation of the secondary metabolites obtained from the X. ingens crude extracts are described in this chapter. Sponges of the genus Xestospongia belong to the family Petrosiidae in the order Haplosclerida and class Demospongie. Most Xestospongia species have a similar 202 morphology with large and thick spicules. However, Xestospongia species with small spicules are also known. There are apporoximately 30 species of Xestospongia found throughput the world in both tropical and cold water habitats.3 The specimens of X. ingens examined in this study were collected on tropical reefs near Sek Point off Madang, Papua New Guinea. The sponge was obtained by hand using SCUBA at depths of -15 to -20 meters in areas where there was little or no surge. The dark brown sponge had a relatively soft texture with fairly small spicules. It was also identified by Dr. R. Van Soest and a voucher sample has been deposited at the Zoologisch Museum, University of Amsterdam. m O CN o 204 4.2 Isolation of cytotoxic metabolites from Xestospongia ingens. Frozen specimens (220g, wet) of Xestospongia ingens were thawed and extracted exhaustively with MeOH. The MeOH extract was filtered and concentrated in vacuo to give a dark brown aqueous suspension which showed cytotoxic activity against murine leukemia P388. The aqueous slurry was diluted with MeOH/water (1:3) to 500 ml and partitioned sequentially against hexanes (3x300 ml), chloroform (3x300 ml) and ethyl acetate (3x300 ml). T L C and 'H NMR analysis suggested the presence of ingenamine alkaloids in the active chloroform fraction.3 The chloroform fraction (850 mg) was chromatographed twice on a sephadex LH-20 column, first with a MeOH eluent, and secondly with a EtOAc:MeOH:H 20 (40:10:4) solution to afford a complex mixture of ingenamine-type compounds. This major fraction was subjected to silica-gel flash chromatography (eluent: EtOAc: Hexane: (iPr)2NH 80:20:2) to give six fractions: fraction 1, consisting of mainly madangamines A (61), B (62) and C (63); fraction 2, pure ingamine A (64); fraction 3, mainly ingamine A (64) and ingenamine I (57); fraction 4, mainly ingenamine E (65) and G (66); fraction 5, mainly the new ingenamines L (60), J (58) and H (56); fraction 6, almost pure ingenamine K (59). Each fraction was further purified on reserved-phase HPLC using an eluent of MeOH/H 2 0 (82:18) to give each of the following pure components: fraction 1 (150mg), consisting of mainly madangamine A (22mg), B (50mg) and C (33mg); fraction 2 (70mg), pure ingamine A (52mg); fraction 3, mainly ingamine A (36mg), keramaphidin B (27mg) and new ingenamine I (23mg); fraction 4 (24mg), mainly ingenamine E (4mg) and G (2.2mg); fraction 5 (70mg), mainly new ingenamine H (12mg), J (7mg) and L (1 lmg); fraction 6 (lOmg), containing pure ingenamine K (4mg). 205 MeOH extract 1. Concentrated 2. Diluted with water 3. Partitioned between aq. and org. solvents Hexane fraction CHC1 3 fraction EtOAc fraction Water fraction Sephadex LH-20 (l.MeOH; 2. EtOAc/MeOH/H 2 0 40:10:4) mixture of alkaloids Silica gel flash chromatography (EtOAc: Hex: (iPr)2NH 80:20:2) Madangamine Ingamine A A. B. C. R . P . H P L C ( M e O H / H 2 0 ) | Ingenamine E. G. Ingamine A Keramaphidin B ingenamine I Ingenamine J. H. L. Ingenamine K Figure 4.1: Isolation of ingenamine and madagamine alkaloids from X. ingens. 206 4.3 Structure elucidation of ingenamine alkaloids from Xestospongia ingens. The structures of the new ingenamines H (56), I (57), J (58), K (59), and L (60) were solved by extensive analysis of ID and 2D NMR, and mass spectrometric data. Proton-carbon attachments were determined by HMQC experiments and proton spin systems were identified from COSY data. The assignments of the quaternary carbons and nitrogen atom interrupted systems were mainly dependent on H M B C results. The identification of the known compounds madangamine A (61), B (62), C (63), ingamine A (64) and the other ingenamine derivatives E (65) and G (66), were confirmed by comparison of their 'H and l 3 C NMR and mass spectroscopic data with the reported values. 3,4 i s 21 13 56 6 4 57 58 59 60 207 I K 208 Ingenamine H 18 21 13 4 56 Ingenamine H (56) was isolated as an optically active, white amorphous solid. The molecular formula of ingenamine H (56) was determined from the parent ion in the EIHRMS (Figure 4.2). Compound 56 gave a very intense M + peak at m/z 426.3582 corresponding to a molecular formula of C 2 8 H 4 6 N 2 0 (AM 2.80ppm), requiring seven sites of unsaturation. Table 4.1 provides a summary of the NMR data acquired for 56. The 'H NMR spectrum of 56 was very similar to those of the known ingenamine alkaloids. It showed three olefinic protons and complex multiplets of aliphatic methylene and methine resonances between 8 0.8 and 3.5 ppm which were not well resolved, even at 500 MHz (Figure 4.3). A broad band at 3422 cm"1 in the IR spectrum, a 1 3 C NMR resonance at 8 70.1 (CHOH: C9) and a 'H NMR resonance at 8 3.31 (CHOH) ppm lead to the assignment of a secondary alcohol, which accounted for the one proton not attached to carbon. All twenty eight carbons were observed in the l 3 C NMR spectrum (Figure 4.3). Among them, four deshielded carbon resonances could be assigned to two double bonds and no additional unsaturated functional groups were apparent from the 1 3 C NMR data, therefore, ingenamine H (56) was pentacyclic. 0 \ o CN -i a a <u OB C •o / ~~3 = t CM eo J O 4 §3 w c o 8 ^8 o O • • CN •<* D U 3 OD £ 1 •o V 1 8 . 1 . . . 11'1 o o 4 - 8 211 6:0C 5.00 4.00 3.00 2.0C i.be Figure 4.4: 2D COSY spectrum of ingenamine H (56) (500MHz, C 6 D 6 ) . 212 A C 8HioO fragment, that included C2 to C6, C8 to CIO, the secondary alcohol, and trisubstituted alkene functionalities, was routinely identified from the COSY, HMQC and H M B C data (Table 4.1). COSY correlations (Figure 4.4) were observed between H4 (8 5.74) and H5 (8 2.48) and H2 (8 2.76, allylic coupling); between H5 (8 2.48) and H6 (8 2.76), H6'(8 1.69) and H8 (8 0.92); between H8 and H9 (8 3.31); and between H9 and both H10 (8 2.63) and H10' (8 2.21) for this contiguous spin system. H M B C correlations were observed between H4 (8 5.74) and C2 (8 65.8); between H6 (8 2.76) and C5 (8 35.6); and between H6' (8 2.76) and C8 (8 53.8). Figure 4.5: Substructure of fragment C 8 H i 0 O and selected HMBC correlations. The aliphatic quaternary carbon at 8 46.6 (C7) , that was indicated by a APT experiment, could be positioned between C2 (8 65.8) and C8 (8 53.8) on this fragment by a series of H M B C correlations. Thus, H8 (8 0.92) showed two bond H M B C correlations to C7 (8 46.6). In addition, H2 (8 2.76) and H6' (8 1.69) were correlated to C8 (8 53.8) via a three bond coupling. A network of HMBC correlation identified C12 (8 54.4) and C20 (8 45.1) as the remaining two substituents flanking the quaternary carbon. A pair of geminal methylene protons HI2 (8 2.23) and HI2' (8 2.08) showed two bond HMBC 213 correlations to the quaternary carbon C7 (8 46.6). In addition, HI2 and HI2' also showed a three bond correlation to C2 (8 65.8) and C20 (8 45.1), and H12' showed a three bond correlation to C8 (8 53.8). The proton H20' (8 1.78) in turn showed two bond HMBC correlations to the quaternary carbon C7 (8 46.6). Furthermore, H20 (8 1.92) showed a three bond correlation to C12 (8 54.4), and H2 (8 2.76) showed a three bond correlation to C20 (8 45.1). Figure 4.6: The tricyclic core structure and four appendages with selected HMBC correlations Since ingenamine H has only one exchangeable proton, it was apparent that there were no hydrogen atoms directly bonded to the two nitrogen atoms. COSY and HMBC data provided evidence that both nitrogens were present as tertiary amines and they were attached to the carbon atoms at C2/C6/C13 and C10/C12/C21, respectively. The chemical shifts of these carbons (C2, 8 63.8; C6, 53.8; CIO, 52.5; CI 2, 50.8) were consistent with attachment to nitrogen. H M B C correlations between H2 (8 2.76) and C6 (8 53.7) were consistent with a nitrogen bridge between C2 and C6. Further HMBC correlations between H2 (8 2.76) and H6 (8 2.76) and a methylene carbon at C13 (8 54.0), and between the geminal protons HI3 (8 2.70) and HI3' (8 2.18) and both C2 (8 65.8) and C6 214 (5 53.7) identified the methylene carbon C13 as the third substituent of the Nl tertiary amine. Similarly, H M B C correlations between HI2 (8 2.23) and CIO (5 57.1) provided evidence for a nitrogen bridge between CIO (8 57.1) and C12 (8 54.4), and correlations between proton H21(8 2.43) and CIO (8 57.1) and CI2 (8 54.4) identified another methylene carbon C21 as the third substituent of the second tertiary amine N i l . The location of the two tertiary amines completed the structure of the tricyclic central core of ingenamine H together with three methylene appendages at N1, C7, and N i l . Confirming that ingenamine H had the typical tricyclic center core found in the family of ingenamine alkaloids, we then assumed this tricyclic center has the same relative stereostructure as that of the known ingenamine compounds. The fourth methylene attachment to the core structure of ingenamine H at C3 was simply identified by the COSY correlation between H4 (8 5.74) and H30 (8 1.96) and H30' (8 1.95), which was due to allylic couplings. The presence of the allylic methylene C30 (8 34.8) substituent at C3 was also confirmed by two bond H M B C correlations from the methylene protons H30 and H30' to the olefinic quaternary carbon C3 (8 144.3), and by three bond H M B C correlation to another olefinic carbon C4 (8 121.1). The remaining portion of ingenamine H had to account for twelve aliphatic methylene and two olefinic methine carbons. Detailed analysis of the COSY, HMQC and HMBC data (Table 4.1) showed these atoms comprised two linear aliphatic chains (C14 to C19 and C22 to C29). The six carbon chain (C14 to C19) linked the C13 and C20 methylene appendages. Thus, COSY correlations observed between H13 (8 2.70) and H14(8 1.45) /HI4' (8 1.18); between HI4/H14' andH15 (8 1.62); between HI5' (8 1.22) 215 and H16 (8 2.38); between H16' (8 1.61) and H17 (8 5.83); between H17 and HI 8 (8 5.82); between H18 and H19' (8 1.76); and between H19 (8 2.28) and H20 (8 1.92) completed the assignment of the first linkage from N l to the aliphatic quaternary carbon C7 (8 46.6). Two bond HMBC correlation from H19' (8 1.76) to the olefinic carbon CI8 (8 130.8) further confirmed the location of the A 1 7 ' 1 8 double bond. Comparison of the NMR data for ingenamine H with that reported for the known ingenmines confirmed the Z geometry for the A 1 7 ' 1 8 alkene.5 The coupling constant between the two olefinic protons H17 and H18 was not measured because of a non 1st order coupling pattern. Instead, The IR spectrum showed the medium resonance at 750 cm"1, which is typical absorbance for the cis disubstituted double bond. The remaining eight aliphatic carbons were saturated methylenes indicated by 1 3 their "C chemical shifts and the APT experiment. Therefore, they comprised the second aliphatic chain bridging the C21 and C30 appendages. Figure 4.7: Substructures of the two side chains and selected HMBC correlations. 217 Table 4.1. N M R data for ingenamine H (56) recorded in benzene-d f i at 500 M H z ( ! H). c# 5 l 3 C ( p p m ) 5 'H(ppm) C O S Y Correlation b H M B C Correlation 2 65.83 2.76 H4 H4,H12,H12'H13,H13' 3 144.34 . . . . . . H2,H30,H30' 4 121.11 5.74 H2,H5,H30,H30' H2,H8,H30,H30' 5 35.57 2.48 H4,H6,H6' ,H8 H 6 6 53.69 2.76 H5,H6' H2,H13,H13' 6' — 1.69 H 5 , H 6 . . . 7 46.63 . . . H2 H2,H12,H12' ,H19,H19'H20 I 8 53.79 0.92 H 5 , H 6 , H 9 H2,H6',H12' 9 70.12 3.31 H8,H10,H10' H8,H10,H10' 10 57.10 2.63 H9,H10' H12,H12',H21 10' — 2.21 H9 ,H10 . . . 12 54.38 2.23 H12' H20,H21 12' 2.08 H12 . . . 13 54.01 2.70 H13',H14,H14' H2 ,H6,H6' 13" — 2.18 H13,H14 . . . 14 26.22 1.45 H13,H13',H14',H15,H15' H13 14' — 1.18 H13,H14,H15,H15' . . . 15 26.97 1.62 H14,H14',H15',H16,H16' HI 3' 15' — 1.22 H14,H14',H15,H16' . . . 16 23.27 2.38 H15,H16',H17 —.-16' — 1.61 H15,H15',H16,H17 . . . 17 131.57 5.83 H16,H16',H18 . . . 18 130.81 5.82 H17,H19,H19' H19' 19 21.55 2.28 H18,H19',H20,H20' H20,H20' 19' — 1.76 H18,H19,H20,H20' . . . 20 45.09 1.92 H19,H19',H20' H2,H12,H12' ,H19 20' . . . 1.78 H19,H19',H20 . . . 21 57.45 2.43 H21',H22,H22' H22,H22' 21' . . . 2.42 H21,H22,H22' . . . 22 25.68 1.36 H21,H21',H23,H23' H21,H21',H23' 22' . . . 1.34 H21,H21',H23,H23' . . . 23 25.52 1.42 H22,H22',H23',H24,H24' H21.H21' 23' — 1.34 H22,H22',H23,H24,H24' . . . 24 26.89 1.40 H23,H23',H24',H25,H25' H23',H25' 24' . . . 1.28 H23,H23',H24,H25,H25' . . . 25 27.47 1.46 H24,H24',H25',H26,H26' . . . 25' . . . 1.32 H24,H24',H25,H26,H26' . . . 26 27.70 1.48 H25,H25' ,H26\H27,H27' . . . 26' . . . 1.32 H25,H25' 1 H26,H27,H27' . . . 27 25.96 1.31 H26,H26',H27',H28,H28' — 27' . . . 1.30 H26,H26',H27,H28,H28' . . . 28 25.86 1.35 H27,H27',H28',H29' H26' 28' — 1.22 H27,H27' ,H28 ! H29,H29' . . . 29 24.11 1.60 H28 , .H29',H30,H30' H30 29' . . . 1.48 H28,H28 , ,H29,H30,H30' . . . 30 34.81 1.96 H4,H29,H29',H30' H4 30' — 1.95 H4,H29,H29' ,H30 — '' Correlated to proton resonance in 5 H column; ' Correlated to carbon resonance in 8 l 3 C column. 218 Ingenamine I IB 21 18 24 31 13 6 4 57 57 Ingenamine I (57) was obtained as a white powder. In the HRFABMS spectrum it gave a parent ion [M+H]+ at m/z 441.3846 corresponding to a molecular formula of C29H48N2O (AM 0.2ppm), which differed from the molecular formula of ingenamine H (56) by an additional C H 2 . The structure of 57 was solved by analysis of a combination of ID and 2D 'H and l 3 C NMR data (Table 4.2). Examination of the NMR data revealed that ingenamine I also had a hydroxylated tricyclic core (Nl to C12) and an eight carbon bridge (Nl to C7) that were identical to those previously identified ingenamine H. For example, the 'H NMR spectrum (Figure 4.10) recorded in benzene-d6 clearly showed two overlapping olefinic protons at a chemical shift of 5.82 ppm ( H17 and HI8) that were very diagnostic for the ingenamine type compounds with an eight carbon chain bridged from Nl to C7 and the Z-double bond at the C17/C18 position. Subtraction of the atoms present in the already identified tricyclic core and the Nl to C7 bridge (C!8H26N 20) from the molecular formula of ingenamine I indicated that the C3 to Nl 1 bridge contained eleven aliphatic carbons. The C APT NMR experiment showed all of the 29 resolved carbon resonances, including five methine carbons and a methyl carbon. Subtraction of the methine carbons C2 (5 65.8), C5 (5 35.7), C8 (5 54.1), 219 C9 (8 71.0) present in the already identified tricyclic core indicated that the C3 to N l 1 bridge contained a methyl branch. Interpretation of COSY (Figure 4.11), HMQC (Figure 4.12) and H M B C data identified the 10-membered aliphatic chain with a methyl branch, which links the C3 and N i l . COSY correlations were observed between H21 (8 2.48) and H22 (8 1.40); between H22 and H23 (8 1.42); between H28 (8 1.46) and H29 (8 1.62); between H29 and H30 (8 2.06) for the straight aliphatic methylene chain. HMBC correlations were observed between H2 (8 2.78) / H4 (8 5.75) and C30 (8 35.2); between H29 (8 1.62) / H30 (8 2.06) and C3 (8 143.9); between H21 (8 2.48) and C12 (8 55.0); between H10 (8 2.63) / H12 (8 2.31) and C21 (8 57.3). They suggested C21 to C23 methylene chain was connected to Nl 1 and the other methylene chain C28 to C30 was connected to C3. The doublet methyl resonance (H31, 0.95) in *H NMR spectr um was coupled to a proton H-b (8 1.60) which was assigned to a methine carbon C-b (8 31.5) by HMQC. The strong H M B C correlations were observed between the methyl group (H31, 0.95) and the methine carbon C-b (8 31.5); between the methyl group H31 and two methylene carbons C-a (8 34.6) and C-c (8 32.8). These correlations indicated that the methine carbon C-b was connected to the two methylene carbons C-a and C-c, and a methyl group C31. However, it was difficult by 220 NMR data to find which carbon was connected to C23 and which carbon was connected to C28. Therefore, there are two possible structure as shown below. In the first one, the methyl group (C31) is connected to C25. In the second one, the methyl group is connected to C26. Freshly isolated ingenamine I (57) was a white solid powder. It (57) was not stable at normal conditions, slowly changing colors from yellowish to deep red-brown in deuterated benzene NMR solvents, and finally decomposing to form a yellowish precipitate. Therefore, the NMR spectra contained the signals of decomposed portions. CN CN s-5 84-•8 o • in o 8 o • o CO o • to Cvl 8 S3 o • tr> o • o o • o cn o -8 C D O l O l O l IT) Q u X o o m 03 c c o CU z ; i -s 0* 223 Figure 4.11: 2D COSY spectrum of ingenamine I (57) (500MHz, C 6 D 6 ) . 224 225 Table 4.2 N M R data for ingenamine I (57) recorded in benzene-d6 at 500 MHz ('H). c# S 13C(ppm) S 'H(ppm) . a COSY Correlation b HMBC Correlation 2 65.81 2.78 H4 H4,H6,H12,H12',H13,H20,H20' 3 143.92 . . . . . . H2,H5,H29,H29',H30,H30' 4 121.35 5.75 H2.H5 H2,H5,H6,H6',H8,H30,H30' 5 35.71 2.55 H4,H6,H6',H8 H4,H6,H6',H8,H9 6 53.48 2.80 H5,H6' H2,H5,H8,H13,H13' 6' 1.65 H5.H6 7 46.45 . . . . . . H2,H5,H8,H12!H12',H20,H20' 8 54.12 0.95 H5,H9 H2,H5,H6,H6',H10,H12,H12',H20' 9 71.00 3.38 H8,H10,H10' H8,H10,H10' 10 56.87 2.63 H9.H10' H8,H12,H12',H21,H21' 10' 2.10 H9,H10 12 55.04 2.31 H12' H2,H10,H10',H20,H20',H21,H2r 12' 2.05 H12 13 53.94 2.68 H13',H14,H14' H2,H6,H6' 13' 2.14 H13,H14,H14' 14 26.23 1.41 H13,H13',H14',H15,H15' H13,H13',H16 14' 1.18 H13,H13',H14,H15,H15' 15 27.01 1.62 H14,H14',H15',H16,H16' H13',H16' 15' 1.49 H14,H14',H15,H16,H16' 16 23.26 2.39 HI5,H15',H16',H17, H15',H17,H18, 16' 1.62 H15,H15',H16,H17 17 131.62 5.83 H16,H16',H18 H16,H16',H19,H19' 18 130.78 5.81 H17,H19,H19' H16,H16',H19,H19' 19 21.55 2.31 H18,H19',H20,H20' H17,H18,H20,H20' 19' 1.80 H18,H19,H20,H20' 20 45.76 1.89 H19,H19',H20' H8,H12,H12' 20' 1.78 H19,H19',H20 21 57.32 2.48 H21',H22,H22' H10,H10',H12,H12',H22,H22' 21' 2.38 H21,H22,H22' 22 25.86 1.40 H21,H21',H22',H23,H23' H21,H21' 22' 1.23 H21,H21\H22,H23,H23' 23 25.01 1.42 H22,H22',H23',H24,H24' H21,H21',H22,H22',H24,H24' 23' 1.08 H22,H22',H23,H24,H24' 24c 34.56 1.25 H23,H23',H24',H25 H31 24' 1.04 H23,H23',H24,H25 25c 31.50 1.60 H24,H24',H26,H26',H31 H24', H31 26c 32.85 1.48 H25,H26',H27,H27' H31 26' 1.22 H25,H26,H27,H27' 2T 21.57 1.31 H26,H26',H27',H28,H28' 27' 1.25 H26,H26',H27,H28,H28' 28 27.85 1.46 H27,H27',H28',H29,H29' H27',H29,H29',H30,H30' 28' 1.29 H27,H27',H28,H29,H29' 29 24.53 1.62 H28,H28',H29',H30,H30' H30,H30' 29' 1.42 H28,H28',H29,H30,H30' 30 35.22 2.06 H29,H29',H30' H2,H4,H29,H29' 30' 1.92 H29,H29',H30 31 21.29 0.95 (d, 8.7Hz) H25 — a. Correlated to proton resonance in 5'H column. b. Correlated to carbon resonance in 8 I 3C column. C. Data for the first proposed structure. 226 Ingenamine J 58 Ingenamine J (58) was obtained as a white powder. The HRFABMS spectrum gave a parent ion [M+H]+ at m/z 413.35273 corresponding to a molecular formula of C27H44N2O (AM - 1.12ppm), which differed from the molecular formula of ingenamine J by the loss of one methylene unit. The structure of 58 was solved by analysis of a combination of ID and 2D 'H and 1 3 C NMR data (Table 4.3). Examination of the NMR data revealed that ingenamine J also had a hydroxylated tricyclic core (Nl to C12) and an eight carbon bridge (Nl to C7) that were identical to those previously identified in the ingenamine H. For example, the 'H NMR spectrum (Figure 4.14) recorded in benzene-d6 clearly showed two overlapping olefinic protons at a chemical shift of 5.62 ppm (H17 and HI8) that were very diagnostic for the ingenamine type compounds with an eight carbon chain bridged from Nl to C7 and the Z-double bond at the C17/C18 position. The l 3 C NMR experiment was also in complete agreement with the proposed structure 58, which showed all of the 27 resolved resonances. Subtraction of the atoms present in the already identified tricyclic core and the Nl to C7 bridge (Ci 8H 26N 20) from the molecular formula of ingenamine J indicated that the C3 to Nl 1 bridge contained nine aliphatic carbons. p -<N CN 3 s s 8 5 00 cn <D cn — o P S i in > -co oo _ i CO CO 1 CO o • LO S3 •9 o • in co o • o CO o - in CJ o - in o • o r--8 ^ a--8 cn o •a o • o co • S o • o o • m o - o o - in in o • o in CM 1°. •= CC o o o o 229 L ' 00 (ppm) Figure 4.16: 2D HMQC spectrum of ingenamine J (58) (500MHz, CD 3 0D). 231 Tab le 4.3. N M R data for ingenamine J(58) recorded in methanol-d 4 at 500 M H z ('H). c# 5 l 3 C(ppm) 5 'H(ppm) a C O S Y Correlation b H M B C Correlation 2 65.74 3.04 H4 H 6 , H ! 2 ' , H I 3 , H 2 0 3 144.70 . . . . . . H 2 ) H 5 , H 2 8 ' , H 2 9 , 4 121.58 5.92 H2 ,H5 H 2 , H 6 , H 6 ' , H 8 , H 2 9 5 35.24 2.69 H 4 , H 6 , H 6 ' , H 8 H 4 , H 6 , H 6 ' , H 9 , 6 54.78 2.86 H 5 , H 6 ' H2 ,H13 ,H13' 6' 1.78 H 5 , H 6 7 47.71 . . . . . . H2,H5,H12' ,H20 ,H20' 8 53.43 0.78 H 5 , H 9 H 2 , H 5 , H 6 , H 6 ' , H 1 0 ' 9 64.90 3.38 H 8 , H I 0 , H 1 0 ' H8.H10 .H10' 10 55.90 2.65 H9,H10' H 2 I ' 10' 2.53 H9 ,H10 12 51.22 2.36 H12' H10,H20' ,H21 12' 2.16 H12 13 55.19 2.92 HI3 ' ,H14 ,H14' H 2 . H 6 ' 13' 2.22 H13.H14 14 27.53 1.48 H13,H13',H14\H15,H15' H13,H13' ,H15 14' 1.25 H13,H14,H15,H15' 15 27.10 1.52 H14.H14' .H15' ,H16,H16' H13 15' 1.35 H14,H14\H15,H16,H16' 16 23.96 2.36 H15,H15 ' ,H16' ,H17 H17.H18, 16' 1.55 H 1 5 , H 1 5 \ H 1 6 , H I 7 17 132.45 5.63 H16,H16 ' ,H18 Hie .Hie ' .Hig .Hw 18 131.32 5.62 H17,H19,H19' H16,H16' ,H19,H19' 19 21.68 2.31 H 1 8 , H 1 9 \ H 2 0 , H 2 0 ' H17,H18,H20,H20' 19' 1.71 H18,H19,H20,H20' 20 43.11 1.82 H19,H19',H20' H2,H8,H12' ,H19 ,H19' 20' 1.72 H19,H19' ,H20 21 58.02 2.87 H21' ,H22,H22' H10,H10' ,H12' 21' 2.45 H21,H22,H22' 22 23.86 1.54 H21 ,H2r ,H22' ,H23,H23' 22" 1.48 H 2 1 , H 2 1 \ H 2 2 , H 2 3 , H 2 3 ' 23 26.71 1.56 H22,H22' ,H23' ,H24,H24' 23' 1.38 H22,H22' ,H23,H24,H24' 24 26.71 1.55 H23,H23' ,H24' ,H25,H25' 24' 1.38 H23,H23' ,H24,H25,H25' 25 27.03 1.35 H 2 4 , H 2 4 \ H 2 5 \ H 2 6 , H 2 6 ' H24' ,H26' 25' 1.33 H24,H24' ,H25,H26,H26' 26 27.34 1.52 H 2 5 , H 2 5 ' , H 2 6 \ H 2 7 , H 2 7 ' H28.H28' 26' 1.35 H25,H25' ,H26,H27,H27' 27 27.40 1.52 H 2 6 , H 2 6 ' , H 2 7 \ H 2 8 , H 2 8 ' H28,H28' ,H29' 27' 1.31 H26,H26' ,H27,H28,H28' 28 25.05 1.77 H27,H27' ,H28' ,H29,H29' H27,H27' ,H29' 28' 1.52 H27,H27' ,H28,H29,H29' 29 34.81 2.19 H28,H28' ,H29' H2,H4,H27',H28' 29' 2.13 H28,H28' ,H29 11. Correlated to proton resonance in 8 ' H column. b . Correlated to carbon resonance in 8 1 3 C column. 232 Ingenamine K 59 Ingenamine K (59) was obtained as a white powder. The HRFABMS spectrum gave a parent ion [M+H]+ at m/z 411.32949 corresponding to a molecular formula of C27H42N2O (AM 0.5 ppm), which differed from the molecular formula of ingenamine J (58) by the loss of two protons. The ID and 2D 'H and 1 3 C NMR data obtained for ingenamine K showed a close correspondence to the data obtained for ingenamine J. However, the 'H NMR spectrum (Figure 4.18) showed one unique resonance at the chemical shift 6.20 ppm (H29), which is not common in the ingenamine family compounds. The COSY correlation showed this olefinic proton H29 (8 6.20) was coupled to another olefin proton H28 (8 5.58) (Figure 4.19). The observation of 16 Hz coupling constant between them indicated that this double bond was trans. The H M B C correlation between H29 (8 6.20) and C2(8 60.4), C3 (8 141.0), and between H28 (8 5.58) and C3 (8 141.0) indicated that the C29 was connected to the tricyclic core appendages of C3 to have a conjugated diene. A broad band at 3422 cm"1 in the IR spectrum and NMR resonances at 8 75.5 (CHOH: C27) and 3.90 (CHOH) ppm were assigned to a secondary alcohol, which accounted for the one proton not attached to carbon. The COSY m CM 235 correlation between olefinic H28 (5 5.58) and H27 (3.90) indicated that the C27 (8 75.5) carbinol methine was directly connected to C28 (8 129.8), which was one end of the conjugated diene. It also suggested that the tricyclic core was not hydroxylated. A detailed analysis of the COSY, HMQC and HMBC data obtained for ingenamine K confirmed the structure. HMQC/COSY correlations identified the chemical shifts of the carbon resonances assigned to C5 (8 39.3), C8 (8 43.5), C9 (8 26.2) and CIO (8 56.3) and their attached protons (Table 4.4). An examination of COSY and HMBC spectroscopic data established the connectivity. Thus, COSY correlations were observed between H5 (8 2.03) and H8 (8 1.05), between H8 and H9 (8 1.42) /H9' (8 1.22), and between H9' and HI0 (8 2.38)/HI0' (8 2.10). A two bond H M B C correlation was observed between H10/H10' and C9 (8 26.2). The l 3 C chemical shifts assigned to C8 (8 43.5) and C9 (8 26.2) of ingenamine K, which differed from those of ingenamine J, could be simply attributed to substituent effects. The absence of an oxygen atom at the C9 position dramatically changed its chemical shift from 64.9 ppm in ingenamine J to 26.2 ppm in ingenamine K. The difference of carbon A A Figure 4.19: Tricyclic core structure with selected HMBC (left) and COSY (right) correlation. 236 chemical swift of 10 ppm for C8 was due to the (3-effect of the oxygen atom. The remaining skeleton of the tricyclic core was the same as that of ingenamine J. An eight carbon bridge extending from Nl to C7 was found to be identical to that previously identified in ingenamine J. COSY correlations observed between H13 (5 2.82) and H14 (5 1.22) /HI4' (8 1.15); between H14/H14' and H15 (8 1.52); between HI5'(8 1.22) andH16 (8 2.37); between H16' (8 1.62) andH17 (8 5.82); between H17 andH18 (8 5.81); between H18 and HI9'(8 1.78); and between HI9 (8 2.32) and H20 (8 1.81) completed the bridge assignment from the first linkage Nl to the aliphatic quaternary carbon C7. A two bond HMBC correlation from HI9' to the olefinic carbon CI 8 further confirmed the location of the A 1 7 ' 1 8 double bond. Similarly, interpretation of COSY, HMQC and H M B C data identified the second five membered aliphatic chain bridging the C21 and C27 appendages. Therefore, there is one ten member carbon linkage from Nl 1 to the olefinic carbon C3. One double bond was connected to C3 and one allylic alcohol was located at C27. Unfortunately, the absolute configuration of C27 is still unknown due to insufficient material. 237 Figure 4.20: 2D COSY spectrum of ingenamine K (59) (500MHz, C 6 D 6 ) . 238 239 Table 4.4: N M R data for ingenamine K (59) recorded in benzene-d 6 at 500 M H z ( 'H). c# 5 1 3 C(ppm) 5 'H(ppm) a C O S Y Correlation b H M B C Correlation 2 60.38 3.529 H 4 H4 ,H6 ,H29 3 140.99 . . . . . . H2,H28,H29 4 128.77 5.90 (dd,6.48,1.00) c H2,H5 H2 ,H6 ,H29 5 39.30 2.03 H4,H6,H6 ' ,H8 H4 ,H6 ,H6 ' 6 52.57 1.70 H5,H6 ' H2 ,H13,H13 ' 6' 2.78 H 5 , H 6 7 49.01 — . . . H12, H19 8 43.47 1.05 H5 , H9 ,H9 ' H2 ,H6 ,H6 ' 9 26.25 1.42 H 8 , H 9 ' , H 1 0 , H 1 0 ' H10.H10 ' 9' 1.22 H 8 , H 9 , H10,H10 ' 10 56.34 2.38 H9,H9 ' ,H10 ' 10' 2.10 H9,H9 ' ,H10 12 56.36 2.45 H12' 12' 2.30 H12 13 53.63 2.82 H 1 3 ' , H 1 4 , H14' H6 ' 13' 2.28 H13, H14, H14' 14 26.02 1.22 H13, H13' , H14' ,H15,H15' H13 14' 1.15 H13,H13 ' ,H14,H15,H15 ' 15 27.19 1.52 HH.HM' .H lS ' .H ie .Hie 1 HI 3' 15' 1.22 H l 4 , H 1 4 ' , H 1 5 , H 1 6 , H 1 6 ' 16 23.29 2.37 H15,H15' ,H16' ,H17, 16' 1.62 H15,H15' ,H16,H17 17 131.67 5.82 H16,H16' ,H18 H16,H16 ' ,H19,H19 ' 18 130.81 5.81 H17,H19,H19 ' H19' 19 21.45 2.32 H18,H19' ,H20,H20' H17,H18,H20,H20 ' 19' 1.78 H18,H19,H20,H20 ' 20 44.95 1.81 H19,H19' ,H20' H2.H19 ' 20' 1.78 H19,H19 ' ,H20 21 52.58 2.83 H21' ,H22,H22' H10' 21' 1.80 H21,H22,H22 ' 22 27.72 1.45 H21 ,H2r ,H22 , ,H23,H23' 22' 1.22 H21,H21 ' ,H22,H23,H23 ' 23 28.31 1.38 H22,H22' ,H23' ,H24,H24' 23' 1.15 H22,H22' ,H23,H24,H24' 24 27.20 1.30 H23,H23' ,H24' ,H25,H25' 24' 1.18 H23,H23 ' ,H24,H25,H25 ' 25 23.47 1.06 H24,H24' ,H25' ,H26,H26' H27 25' 1.02 H24,H24' ,H25,H26,H26' 26 36.97 1.80 H25,H25' ,H26' ,H27 26' 1.38 H25,H25' ,H26,H27 27 75.47 3.90 H26,H26' ,H28 H29 28 129.84 5.58 (dd, 15.82,8.20) H27,H29 29 131.97 6.20(d, 15.83) H28 H 2 , H 4 a. Correlated to proton resonance in 5 ! H column. b. Correlated to carbon resonance in 8 1 3 C column. 240 Ingenamine L 18 Ingenamine L (60) was obtained as a white powder. In the HRFABMS spectrum it gave a parent ion [M+H]+ at m/z 423.33870 corresponding to a molecular formula of C28H42N2O (AM 2.73 ppm), which differed from the molecular formula of ingenamine H (56) by the loss of four protons. The structure of 60 was solved by analysis of a combination of ID and 2D ! H and 1 3 C NMR data (Table 4.5). Examination of the NMR data revealed that ingenamine L also had a hydroxylated tricyclic core (Nl to C12) and an eight carbon bridge (Nl to CI) that were identical to those previously identified in the ingenamine H. For example, the 'H NMR spectrum (Fig. 4.23) recorded in methanol^ clearly showed two overlapping olefinic protons at a chemical shift of 5.63 ppm (HI7 and HI8) that were very diagnostic for the ingenamine type compounds with an eight carbon chain bridged from Nl to C7 and the Z-double bond at the C17/C18 position. The 1 3 C NMR data, which showed all of the 28 resolved resonances, was also in complete agreement with the proposed structure 60. However, it contained four more olefinic carbon resonances in the 8 129 to 130 ppm region and lost four aliphatic carbon resonances in the 8 28 to 30 ppm region compared with that of ingenamine H. Subtraction of the atoms present in the already identified tricyclic core and the N l to C7 co O CO CO o oo fe c? BE f l C M co •*"" i 3 co 5 + 8 o i n co o • o CO o i n C M 8 Si-ts • ts C M 8-S3-o • o 9-8 " o • i n o • o ts • t n i n o - o i n o m o - o o o o o 243 . . . . . . . *^s- . - . t f . . * . . . , ; . . 9: . . . i i . .** v . ft § a 8 ! & ''-FA'*- e • » OOD to C 3 g t • a * • < « > & —i . 1 1 r 0 c5 (ppm) Figure 4.24: 2D COSY spectrum of ingenamine L (60) (500MHz, Methanol-d4). 244 bridge ( C i 8 H 2 6 N 2 0 ) from the molecular formula of ingenamine L indicated that the C3 to N i l bridge had to be also ten carbons long and had to contain two olefins. The absence of a UV chromophore in ingenamine L indicated that there was no conjugation between the two double bonds. In addition, the 'H NMR spectrum contained one set of geminal methylene protons H25 (8 2.93) / H25' (8 2.53) that had chemical shifts typical of protons on doubly allylic carbons in polyunsaturated fatty acids. These two geminal protons were also correlated into the cluster of olefinic protons between 5.3 and 5.4 ppm in the COSY spectrum (Figure 4.24). This evidence, along with the HMBC data were consistent with a methylene interrupted diene substructure (C23 to C27). Furthermore, the COSY and HMBC data (Table 4.5) demonstrated that there were two methylene carbons between Nl 1 and the respective ends (C23) of the diene substructure, and three methylene carbons existed between C3 and another end (C27) of the diene substructure. Thus, COSY correlations were observed between H21 (8 2.78) and H22 (8 2.24), between H22 and H23 (8 5.31); and between H30 (8 2.30) and H29 (8 2.08), H29 and H28 (8 2.38), H28 and H27 (8 5.31). H M B C correlations were observed between H21(8 2.78) /H23 (8 5.31) Figure 4.25: Selected COSY and HMBC correlations of one side chain (C21-C30). 245 and C22 (5 23.2), and between H26 (8 5.31) /H27 (5 5.31) /H30 (5 2.30) and C28 (8 27.5), H27/H30 and C29 (8 27.6) (Figure 4.25). Comparison of the N M R data for ingenamine L with its geometric isomer ingenamine D (67)~ confirmed the Z geometry for A " and A" . Also, there was no absorbance at 950-980 cm"1 in the IR spectrum, which is typical resonance for the trans disubstituted double bond. Instead, the IR spectrum only showed the medium resonance at 750 cm"', which is typical absorbance for the cis disubstituted double bond. Therefore, 1 - 7 1 0 9 ^ 9 4 . 9fS 97 three disubstituted A ' , A , and A ' olefins had the Z geometry. The complete structure of ingenamine L was as shown. 246 \- 1 0 0 Figure 4.26: 2D H M Q C spectrum of ingenamine L (60) (500MHz, Methanol-d4). 247 Table 4. 5: N M R data for ingenamine L (60) recorded in methanol-d 4 at 500 M H z ( 'H). c# 8 l 3 C(ppm) 5 'H(ppm) a C O S Y Correlation b H M B C Correlation 2 65.65 2.98 H4 H 4 , H 6 , H 1 3 , H 1 2 „ H 1 2 ' , H 2 0 3 144.60 — . . . H 2 , H 5 , H30,H30' , 4 121.97 5.92 H2,H5 H2,H5,H6,H6 ' ,H8 ,H30 ,H30 ' 5 35.28 2.69 H4,H6,H6 ' ,H8 H2 ,H4 ,H6 ,H6 ' ,H8 ,H9 , 6 55.07 2.87 H5,H6 ' H2 ,H8 ,H13 ,H13 ' 6' 1.75 H 5 . H 6 7 47.70 . . . . . . H 2 , H 5 , H 8 , H 12.H 12',H20,H20' 8 53.37 0.78 H 5 , H 9 H 2 , H 4 , H 6 , H 6 ' , H 10,H 10',H 12 9 69.17 3.30 H8,H10,H10 ' H8,H10,H10 ' 10 55.42 2.65 H9.H10 ' H12.H21 10' 2.55 H9 ,H10 12 50.06 2.27 H12' H 2 , H 10,H10',H22,H22',H20' 12' 2.12 H12 13 55.41 2.94 H13' ,H14,H14' H2 .H6 ' 13' 2.22 H13,H14,H14 ' 14 27.18 1.49 H I S . H I S ' . H U ' . H I S . H I S 1 H n . H H ' ^ i s ' . H i e 14' 1.28 H13,H13 ' ,H14,H15,H15 ' 15 27.15 1.79 H14,H14' ,H15' ,H16,H16' H13,H13 ' ,H16 15' 1.59 H14,H14' ,H15,H16,H16' 16 23.91 2.38 H15,H15' ,H16' ,H17 H17.H18 16' 1.56 H15,H15' ,H16,H17 17 131.39 5.63 H 1 6 , H 1 6 \ H 1 8 H16.H19.H19 ' 18 132.40 5.63 H17,H19,H19 ' H16,H19,H19 ' 19 21.67 2.31 H18,H19' ,H20,H20' H17,H18,H20,H20 ' 19' 1.74 H18,H19,H20,H20 ' 20 42.93 1.85 H19,H19' ,H20' H2,H8,H12,H12 ' ,H19 ' 20' 1.74 H19,H19 ' ,H20 21 58.27 2.78 H21' ,H22,H22' H10,H10' ,H12' ,H22' 21' 2.41 H21,H22,H22 ' 22 23.16 2.24 H21,H22,H22 ' ,H23, H21.H23.H24, 22' 2.21 H21,H21' ,H22,H23 23 129.01 5.31 H22,H22' ,H24, H22 ,H22 ' ; H25,H25 ' 24 129.86 5.59 H23,H25,H25 ' H22,H22 ' ,H25,H25 ' 25 27.20 2.93 H24,H25' ,H26 H23,H24,H26,H27 25' 2.53 H24,H25,H26 26 129.69 5.31 H25,H25' ,H27 H25,H25 'H28,H28 ' 27 129.19 5.31 H26,H28,H28 ' H25,H25 ' ,H28,H28 ' 28 27.45 2.38 H27,H28' ,H29,H29' H26,H27,H30 28' 2.32 H27,H28,H29,H29 ' 29 27.57 2.08 H28,H28' ,H29' ,H30,H30' H27 ,H30 29' 2.00 H28,H28 ' ,H29 > H30,H30 ' 30 36.19 2.30 H29,H29' ,H30' H2 ,H4,H28 30' 2.15 H29,H29' ,H30 a. Correlated to proton resonance in 8 H column. b. Correlated to carbon resonance in 8 I 3 C column. 248 4.4 Biogenic consideration. Sponges of the Order Haplosxlerida are a rich source of biogenetically related 3-alkylpiperidine and 3-alkylpyridine alkaloids that possibly constitute one of the best chemotaxonomic indicators for marine sponges.6 The ingenamine and madangamine alkaloids isolated from Xestospongia ingens also belong to this class of alkaloids. It has been suggested that these alkaloids are all derived from reduced bis-3-alkylpyridine macrocycles according to Baldwin and Whitehead's first proposed biogenesis for the manzamines.7 Their proposal suggested that the pentacyclic skeleton 68 of the ingenamine alkaloids arose from a biological intra molecular [4+2] cycloaddition reaction between the tautomeric forms of the two dihydropyridine rings in a bis-3-alkyldihydropyridine macrocycle 69 (Figure 2.27). The full connectivity and relative stereochemistries observed in the ingenamine alkaloids follows from the expected endo and regiochemical preferences of the [4+2] cycloaddition reaction. Recently, Baldwin reported the biomimetic synthesis of ingenamine derivative keramaphidin B (70) from the macrocycle 69, the first in vitro chemical evidence for this proposal.8 18 70 Figure 4.27: Proposed biogenesis for the ingenamine skeleton. 250 4.5 Biological activity and discussion All five new isolated ingenamine derivatives were tested in vitro against murine leukemia P388, giving the results in the Table 4.6. These activities can thus be rationalized that these alkaloids share in common two important features: the presence of one or more relatively delocalized positive charges (protonated dehydropyridine ring) and hydrophobic alkyl chains. These features are known to be sufficient for significant toxicity against mammalian, bacterial, and fungal cells.9 Such properties are likely to be attributed to disruption of membranes which cause lysis of cell. But it could be that these compounds have also the ability, for the same reasons, to cross membranes and further interact with other targets in the cell, thus reinforcing their activity. Table 4.6: cytotoxic activity against murine leukemia P388 cells in vitro.* Ingenamine H I J K L ED50(ug/ml) 1.48 1.05 1.45 0.89 0.71 The data were measure by Dr. Theresa M . Allen group in University of Alberta. Ingenamine H (56) and J (58) have different length saturated aliphatic chains between N i l and C3 compared to known ingenamines. Ingenamine L (60) is a geometric 23 4^ C^ "^7 isomer of ingenamine D (67). L (60) contained two disubstituted A " a n d A" '" olefins while D (67) has two olefins at A" ' ' and A"" . The more interesting derivatives are ingenamine I (57) and K (59). 57 was found to have a methyl branch in the aliphatic chain between Nl 1 and C3. Methyl branched 3-alkylpiperidine alkaloids have previously only been found in many of the monomeric 3-alkylpiperidines (such as ikimine A (71)6'10 ) and bis-quinolizadine alkaloids (such as petrosin (72)6'" ). The methyl group in these molecules is usually 2-4 bonds removal from the nitrogen atom. Ingenamine I (57) is the 251 first ingenamine alkaloid that has a methyl branch. Moreover, the methyl group is at the center of a 10-membered aliphatic chain. Its (57) structure raises interesting biosynthetic questions. Ingenamine K (59) has a conjugated diene and an allylic alcohol in the aliphatic chain. It is also first compound in the ingenamine family which has this structural functionality. 252 4.6 Experimental NMR data were collected on either a Bruker AMX500, a Bruker WH400 or a Bruker AM400 spectrometer each equipped with a 5mm probe. The chemical shifts are reported in ppm downfield from the tetramethylsilane resonance with the solvent residual peaks as the references ('H: CD 3 OD 3.30 ppm, C 6 D 6 7.15 ppm; 1 3 C : C D 3 O D 49.0 ppm, CeD6 128.0 ppm). The coupling constants (J) are given in Hz. The low and high resolution EI mass spectra were recorded on Kratos MS50/DS55SM mass spectrometer. The low and high resolution FAB mass spectra were recorded on a Kratos Concept II HQ mass spectrometer. UV, IR spectra were taken on Bausch-Lomb Sepctronic-2000 spectrophotometer and Perkin-Elmer 1600 FT spectrometer, respectively. Optical rotations were measured with a JSACO J-700 spectropolarimeter and the [OC]D values are 1 ^ 1 given in 10" degcm'g" . Reversed phase and Normal-phase thin layer chromatography (TLC) was performed using Whatman MKC18F and Kieselgel 60 F254 plates. Visualization was detected by U V (X,=254 nm) and/or heating after spraying with vanillin reagent. Normal phase column chromatography was carried out either on Merck silica gel G60 (230-400 mesh) or Sigma silica gel (size: 10-40 u). Reversed phase chromatography was performed using reversed phase silica prepared according to the literature.12 High performance liquid chromatography (HPLC) separations were done on one of two possible systems using either a Whatman Partisil 10 ODS-3 Magnum column or Rainin Partisil 10-ODS column. The first system consisted of a Waters 600E HPLC pump/system controller with a Waters 486 tunable absorbance detector. The second system consisted of a Waters 600E HPLC pump/system controller equipped either with a 253 Waters 996 photodiode array detector or RI (Perkin Elmer LC-25) detector. Both systems were interfaced with a personal computer using Millenium™ 2010 chromatography software. The solvents used for extraction and for column chromatographies were Fisher reagent grade. HPLC solvents were Fisher HPLC grade which were filtered and degassed prior to use. All other solvents, reagents and standards were reagent or commercial grade and were used without further purification. Ingenamine H (56): a white amorphous solid; [a] 2 2 D = -30° (10"' degcm2g"', MeOH); HREIMS: [M] + m/z 426.3582 ( C 2 8 H 4 6 N 2 0 , AM 2.8ppm); 'H NMR (C 6 D 6 , 500 MHz) and 1 3 C NMR (C 6 D 6 , 100 MHz), and 2D NMR data listed in Table 4.1. Ingenamine I (57): a white amorphous solid; [a]22o = -150° (10"' degcm2g"', MeOH); HRFABMS: [M+H]+ m/z 441.3846 ( C 2 9 H 4 9 N 2 0 , AM 0.2ppm); 'H NMR (C 6 D 6 , 500 MHz) and 1 3 C NMR (C 6 D 6 , 100 MHz), and 2D NMR data listed in Table 4.2. 22 0 1 ^ 1 Ingenamine J (58): a white amorphous solid; [a] D = -12 (10" degcnTg" , MeOH); HRFABMS: [M+H]+ m/z 413.3527 ( C 2 7 H 4 S N 2 0 , AM -l.lppm); 'H NMR (CD 3OD, 500 MHz) and l 3 C NMR (CD 3 OD, 100 MHz), and 2D NMR data listed in Table 4.3. Ingenamine K (59): a white amorphous solid; [a] 2 2 D = -83° (10"' degcirfg"1, MeOH); HRFABMS: [M+H]+ m/z 411.3295 ( C 2 7 H 4 2 N 2 0 , AM 0.5ppm); 'H NMR (C 6 D 6 , 500 MHz) and 1 3 C NMR (C 6 D 6 , 100 MHz), and 2D NMR data listed in Table 4.4. Ingenamine L (60): a white amorphous solid; [OC] 2 2D = -18° (10"1 degcnrg"1, MeOH); HRFABMS: [M+H]+ m/z 423.3387 ( C 2 8 H 4 3 N 2 0 , AM 2.73ppm); 'H NMR (CD 3 OD, 500 MHz) and 1 3 C NMR (CD 3 OD, 100 MHz), and 2D NMR data listed in Table 4.5. 254 Reference: 1. Attaway DH, Zaborsky OR (1993). Marine biotechnology, Vol 1, pharmaceutical and bioactive natural products, Plenum Press, New York. 2. Konig G M , Wright AD (1996). Planta Media 62:193. 3. Kong, Fangming thesis, The University of British Columbia, 1994. 4. a) Kobayashi, J.; Tsuda, M. ; Kawasaki, N.; Matsumoto, K.; Adachi, T. Tetrahedron Lett. 1994, 35, 4383-86. b) Kong F.; Andersen, R.J.; Allen, T .M. Tetrahedron 1994, 50, 6137-44. 5. a) Kong, F.; Andersen, R.J.; Tetrahedron 1995, 57, 2895-2906. b) Fusetani, N.; Asai, N.; Matsunaga, S.; Honda, K.; Yasumoro, K. Tetrahedron Lett. 1994, 35, 3967. c) Kondo, K.; Shigemori, H.; Kikuchi, Y. Ishibashi, M. ; Sasaki, T.; Kobayashi, J. J. Org. Chem. 1992, 57, 2480. 6. (a). Andersen, R.J.; Van Soest, R. W. M ; Kong, F. Alkaloids: Chemical and Biological Perspectives; Pelletier, S. W., Ed.; Pergamon Press, Elsevier Science: Oxford, UK, 1996; p301-355. (b) Tsuda, .; Kobayashi, J. Heterocycles 1997, 46, 765-794. 7. Baldwin, J.E.; Whitehead, R.C. Tetrahedron Lett. 1992, 33, 2059-2062.. 8. Baldwin, J.E.; Claridge, T.D.W; Culshaw, A.J.; Heupel, F.A. Lee, V.; Spring, D.R.; Whitehead, R . C ; Boughtflower, R.J.; Mutton, I.M.; Upton, R.J. Angew. Chem. Int. Ed. 1998, 37, No. 19, 2661-2663.. 9. Rose, I.C.; Sharpe, B.A.; Lee, R . C ; Griffin, J.H.; Capobianco, J.O.; Zakula, D.; Goldman, R.C. Bioorg. Med. Chem. 1996, 4, 97-103. 10. Kobayashi, J.; Zeng, C ; Ishibashi, M.; et, al. J. Chem. Soc. Perkin Trans. 1992, 7, 1291-1294. 11. Braekman, J . C ; Daloze, D.; et, al. Tetrahedron Lett. 1982, 23, 4277-4282. 12. Kiisher, T . C ; Lindsten, G.R. J. Org. Chem. 1983, 48, 3589-3591. 255 Chapter V: Antimitotic agents from the Southern American plant Ilex macrophylla More than 80,000 extracts from the National Cancer Institute's open repository were screened by the antimitotic assay. One of the active extracts was from the South American plant Ilex macrophylla. This plant genus is widely distributed in South America and Eastern Asia. The barks, leaves, and roots have been used as a traditional Chinese remedy for scalds and burns, to control bleeding, and as a binding medicine;1 It is also used for the treatment of cardiovascular disease, hypercholesteremia, etc." However, no antimitotic activity has ever been reported. The crude Ilex extract was found to contain antimitotic activity which acts like paclitaxel by stabilizing microtubules. Because paclitaxel is the only known microtuble stabilizer isolated from a terrestrial plant, this tropical plant became a promising candidate to provide novel antimitotic agents. Further chemical studies were performed in order to identify the antimitotic metabolites. 5.1 Isolation of antimitotic metabolites from plant Ilex macrophylla . The concentrated dark green methanol extract was received from NCI. This extract was fractionated by a modified Kupchan partitioning scheme. The 2g extract was dissolved in 500ml of MeOH/H 2 0 (1:4) and sequentially extracted with hexane (3x 200 ml), chloroform (3x 200 ml), and EtOAc (3x 200 ml). The fractions obtained after this solvent partitioning were tested again by the antimitotic assay. Only the chloroform soluble fraction was active. The chloroform fraction was further separated by Sephadex LH-20 chromatography (eluent: MeOH). The eluent from the column was collected in 100 test tubes. Each tube contained 8 ml of solution. The bioassay results showed samples in 256 tubes 10 to 15 were active. A small portion of the active component was fractionated by gradient reversed-phase HPLC. General gradient rates of 3%/min from 0 to 100% acetonitrile were employed. The fractions were directly collected in 96 well plates and bioassayed for antimitotic activity. It was found that two peaks (E and F) in the HPLC trace gave a positive response. These two peaks were eluted from HPLC with the solvent MeCN/HaO (57:43) and the mass spectrum of both peaks was directly obtained by L C -MS (Fig. 5.1). 700 800 900 1000 Figure 5.1: HPLC traces on UV spectrum (228 nm) with flow rate of l.Oml/min (Top), and corresponding MS of peaks E and F (Bottom). In the next step, the large amounts of fractions E and F were prepared from the previous active fraction by the repeated reversed-phase Sep-pack chromatography using a gradient solvent from 0% to 100% acetonitrile in water. The mixture of fractions E and F were eluted with a solvent composed of 60% acetonitrile. In the final step, this mixture was purified by HPLC to give pure fraction F and two compounds in fraction E (Figure 5.2). The pure fraction F (lOmg) was obtained from the 2g crude extract, which is the major metabolite in the extract. Fraction E gave two pure compounds A (1.3 mg) and B (3.6 mg) from the total 18g crude extract. NCI extract Hexane Chloroform Water LH-20 (eluent: MeOH) Tube 10-15 of 100 MeCN: H 2 0 57 : 43 LC-UV-MS R.P. Sep-pack Fraction E R.P. HPLC (eluent: MeCN/H 2 0) MeCN : H 2 0 MeCN : H 2 0 MeCN : H 2 0 MeCN : H 2 0 20 : 80 40 : 60 60:40 80 : 20 R.P. HPLC (eluent: MeCN/H 2 0) Fraction F ( Sapnin ) Compound A Compound B Figure 5.2: Isolation scheme of antimitotic agents from Ilex macrophylla. 258 5.2 Results: Fraction F was isolated as a white solid. The molecular weight was determined from the parent ion [M+H]+ at m/z 789.44039 in the FABMS, corresponding to a molecular formula of C^H^OnNa. It was identified as the sodium salt of ilexsaponin B1 (73) by extensive ID and 2D NMR spectroscopy and mass spectrometric analysis. Its structure and stereochemistry were ultimately confirmed by comparison with published data.3 The 'H and l 3 C NMR spectra are shown on the next page. This triterpene glycoside is a typical component and marker of the genus Ilex. It belongs to the family of saponins, which are present in many organisms, particularly plants. Because saponins have surface-active properties, they may be recognized by shaking an aqueous solution of the sample and observing the production of foam, which is stable for approximately 15 minutes.4 260 Fraction F gave two pure compounds A and B. Compound A was also a white solid. The molecular weight was determined from the parent ion [M+H]+ at m/z 806.30215 in the HRFABMS, corresponding to a molecular formula of C 4 4 H 5 5 N O 1 3 . It was identified as 10-deacetyltaxuyunnanine A (74) by the ID and 2D NMR spectroscopy and mass spectrometric analysis. Its structure was confirmed by comparison with published data.5 Compound B was identified as 7-(p-xylosyl)-10-deacetyltaxol C (75) with the molecular formula of C 4 9 H 6 3 N O 1 7 . Its structure was also confirmed by comparison with published data.6 Both compounds A and B are analogues of paclitaxel (76). They are modified at the amide bond on the side chain. The benzoyl group was replaced by an n-hexanoyl group. Moreover, compound B contains a (3-xylosyl moiety, attached via an acetal bond with the 7-hydroxyl group. The *H and l 3 C NMR spectra are shown on the next page. 263 5.3 Antimitotic activity of paclitaxel analogs7 The extent to which the antimitotic compounds arrested cells at mitosis was measured by the amount of mitosis-specific antigen as determined by ELICA assays. MCF-7 mp53 cells were exposed to paclitaxel (76), 10-deacetyltaxuyunnanine A (74), and 7-((3-xylosyl)-10-deacetyltaxol C (75) and the result of ELICA assays is shown in Figure 5.6. Whereas paclitaxel (76) shows an IC50 between 1-10 ng/ml (1.2-12 11M), 10-deacetyl-taxuyunnanine A (74), and 7-(p-xylosyl)-10-deacetyltaxol C (75) have much higher ICso's at 50 - 100 nM respectively. The results showed that a 100-fold higher concentration of 10-deacetyl-taxuyunnanine A and 7-((3-xylosyl)-10-deacetyltaxol C must be used as compared to paclitaxel to obtain a similar degree of mitotic arrest. Whereas paclitaxel arrests cells in mitosis in the nM range, the paclitaxel analogs isolated from Ilex work in the pJVI range. Ilexsaponin B 1 (73) does not show antimitotic activity. The reason why fraction E was active could be that fraction contained a little portion of fraction F when collected. However, it has ability to lyse red blood cells as well as the detergent ability due to its surface-active properties.8 5.4 Cytotoxicity of paclitaxel analogs7 Cytotoxicity assays using crystal violet staining were used to determine the percentage of survival when cells were exposed to antimitotic agents. When MCF-7 mp53 cells were used (Fig. 5.7), paclitaxel (76) showed a lower IC50 of 1-5 nM while 10-deacetyltaxuyunnanine A (74) and 7-((3-xylosyl)-10-deacetyltaxol C (75) possessed higher IC 5 0's of 1 - 100 nJVI and 0.1-1 pJVI, respectively. 264 2 R 1.5 -in 9 1 < 0.5 h 0.001 0.01 Concentration ipM for Ilex analogs; ng/ml for paclitaxel) Figure 5.6: Mitotic arrest by paclitaxel analogs. MCF-7 mp53 cells were incubated with increasing concentration of 10-deacetyltaxuyunnanine A (•) and 7-(p-xylosyl)-10-deacetyltaxol C (O), and paclitaxel (A). After 16 hours, antimitotic ELICA was performed using TG3, a mitotic indicator antibody. Absorbance at 405 nm was taken, and is a measure of relative amount of cells arrested in mitosis. Cytotoxicity 265 100 r 80 60 > CO — 40 o 20 o V 0.001 0.01 0.1 Concentration (//M) 10 Figure 5.7: Cytotoxicity of paclitaxel analogs. MCF-7 mp53 and A549 cells were incubated with increasing concentration of 10-deacetyltaxuyunnanine A (A, •) and 7-(P-xylosyl)-10-deacetyltaxol C (£?, O). After 16 hours, the antimitotic agent was removed and the cells were grown until those in control wells had reached confluency. Number of cells were measured by crystal violet staining and A570 readings. Cell survival was then calculated by comparing number of cells in treated wells to those in control wells. 266 5.5 Structure activity relationship of paclitaxel analogs Although 10-deacetyltaxuyunnanine A (74) and 7-([3-xylosyl)-10-deacety] taxol C (75) differ from paclitaxel (76) at different side groups, they show the same mode of action, inducing mitotic arrest by tubulin overpolymerization. However, the paclitaxel analogs are a hundred fold less active than paclitaxel. To investigate the structure activity relationship for these analogs, a structural model was considered.9 Ojma attributed the biological activity of paclitaxel to three key structural components: the C-2 benzoate, the C-3' -N-benzoyl, and the C-3' phenyl group. It was found that the C-2 benzoate (motif B) of paclitaxel binds specifically to the tubulin-binding site found on the (3-subunit of tubulin heterodimers.10 Figure 5.8 : Structure of paclitaxel (76), 10-deacetyltaxuyunnanine A (74) and 7-(0-xylosyl)-10-deacetyltaxol C (75). Labeled boxed regions are areas of common overlap determined by 3-dimensional structural analysis. By comparing 10-deacetyltaxuyunnanine A (74) and 7-((3-xylosyl)-10-deacetyltaxol C (75) to the structure-activity model, a few observations were made. First, the C-2 benzoate (motif B for paclitaxel) was found on both paclitaxel analogs, allowing them to bind to tubulin heterodimers. The substitution at C-10 and C-7 positions play less 267 significant roles in the antimitotic activity of 10-deacetyltaxuyunnanine A (74) and 7-((3-xylosyl)-10-deacetyltaxol C (75) since they acted as the binding region with tubulin heterodimers and are not part of the motif necessary for action. Finally, the alkyl side chain substitution at the C-3' position for the phenyl group might have an effect on the activity of the paclitaxel analogs. However, it would not result in total loss of activity since the long alkyl chain would still allow for hydrophobic interaction to take place. Therefore, both paclitaxel analogs contain the backbone and essential side chain that gave them their antimitotic activity. 5.6 Effect of combining antimitotic compounds and saponin metabolites7 Isolation of pure active components from a natural extract which demonstrates antimitotic activity often leads to a pure compound that has a lower biological activity than the original crude extract. To see if other components present in the Ilex extract would enhance the ability of the antimitotic agents, 10-deacetyltaxuyunnanine A (74) and 7-((3-xylosyl)-10-deacetyltaxol C (75) were each combined with the ilexsaponin BI (73). The results of ELICA assays are shown in Figure 5.9. The results show that the presence of ilexsaponin BI (73) enhances the activity of both paclitaxel analogs. Ilexsaponin BI (73) concentration (nM) Figure 5.9: Effect of ilexsaponin B1 (73) on anti-mitotic activity of 10-deacetyl-taxuyunnanine A (74) and 7-((3-xylosyl)-10-deacetyltaxol C (75). MCF-7 mp53 cells were incubated with one of paclitaxel analog for 16 hours with increasing concentrations of ilexsaponin BI. Antimitotic ELICA was performed using TG3, a mitotic indicator antibody. Absorbance at 405 nm was taken, and is a measure of relative amount of cells arrested in mitosis. 269 5.7 Significance of the result and future research Although it was disappointing that the active components in the extract of plant Ilex macrophylla are paclitaxel analogs and not structurally distinct new antimitotic agents, we nevertheless obtained two antimitotic compounds that work via a taxol-like mechanism. It is the first time that pacitaxel derivatives have been reported from a tree outside of the Yew tree genus Taxus. These two paclitaxel derivatives were originally isolated from the plants Taxus yunnanensis5 and baccata L6, respectively. The present finding offers the possibility of identifying other pacitaxel derivatives from other sources and shows the effectiveness of the screening method at identifying potential antimitotic compounds. Future research should continue to identify possible sources and to isolate potential antimitotic agents using the screening method. In addition, the interesting observation that ilexsaponin B l (73) enhances the antimitotic activity of 10-deacetyltaxuyunnanine A (74) and 7-(fj-xylosyl)-10-deacetyltaxol C (75) warrants further investigation. 270 References: 1. Jiangsu New Medical College. " Dictionary of Chinese Materia Medica", Shanghai Scitific Technological Publishers, Shanghai, 1977, pp. 2096. 2. Jiangsu New Medical College. " Dictionary of Chinese Materia Medica", Shanghai Scitific Technological Publishers, Shanghai, 1977, pp. 441. 3. Hidaka. K.; Ito, M. ; et al; Kagei, K. Chem. Pharm. Bull. 1987, 35(2), 524-529. 4. Isolation of Natural Products. Richard J. P. Cannell, Humana Press Inc. pp 358 5. Zhang, H.; Takeda, Y.; et. al; Sun H. Heterocycles. 1994, 38 (5), 975-980. 6. Senilh, V.; et. al.; Varenne, P. J. Nat. Prod. 1984, 47(7), 131-137. 7. Wesley Chiang's thesis (April, 2000) in Dr. Roberge group. All of bioassay data were measured by Wesley Chiang. 8. Isolation of Natural Products. Richard J. P. Cannell, Humana Press Inc. pp. 290. 9. Ojima, I; Inoue T.; et al. Proc. Natl. Acad. Sci. USA. 1999, 96, 4256-4261. 10. Han, Y.; Chaudhary A.G.; ChordiaM.D.; etal. Biochemistry. 1996, 35, 14173-14183. 271 Chapter VI: Conclusion Herein I presented the isolation and structure elucidation of secondary metabolites from several marine organisms and terrestrial plant extracts guided by novel high throughput bioassays (HTS) for G2 checkpoint inhibitors and antimitotic agents. The biological activity and chemical structure of secondary metabolites from four different species were investigated using the novel G2 checkpoint inhibitor assay. When this novel assay was first applied to screening natural product extracts, a bacterial isolate, clinl 116, obtained from the surface of a Northeastern Pacific Ocean sponge, was found to have the significant activity. This microorganism extract was found to contain the known G2 checkpoint inhibitor staurosporine (10), a result which validated the ability of the new assay to identify G2 checkpoint inhibitors. The second species, the soft coral Pachyclavularia violacea, collected near Sek Point off Madang in Papua New Guinea, was found to produce many diterpenoids. The diterpenoid pachyclavularolides A (43), B (44), C (45), D (46), E (47), F (18), and G (48) were isolated and characterized by extensive ID and 2D NMR spectroscopy. Pachyclavularolide F (18) was the only one of these diterpenoids that had significant biological activity. It showed moderate activity in assays for in vitro cytotoxicity and G2 cell cycle checkpoint inhibition. The third species, the sponge Aaptos auberitoides, was collected as part of a general collecting expedition to Indonesia. Its extract was found to contain the known metabolite aaptamine (12) and its analogs 50 to 52. Aaptamine alkaloids possess potent pharmacological activities besides the G2 checkpoint inhibition. The fourth species studied was a tropical bush Parinari curatellifolia, obtained from the National Cancer 272 Institute's (USA) open repository of natural product extracts. The enr-kaurene diterpenoids 19 and 53 to 55 were found to have a significant G2 checkpoint inhibition. The overall checkpoint inhibitors described above provide substantial opportunities for using these small molecule as tools to identify molecular targets and functionally relevant residues of proteins on the G2 checkpoint pathway in mammalian cells. Thus, the crucial protein and related factors could be elucidated and a deeper understanding of the mechanisms of G2 arrest could be gained. Meanwhile, a new protein target-directed HTS assay for anticancer drug discovery will emerge. It was found that the targets of lactone 55 are not the known targets for the typical checkpoint inhibitors. Its isotope-labeled analog 54A was used as a ligand and several proteins were found to bind to this molecule. The purification and identification of these proteins is being investigated. Using another new high throughput antimitotic assay, the paclitaxel derivatives 10-deacetyltaxuyunnanine A (74) and 7-((3-xylosyl)-10-deacetyltaxol (75) were isolated from the extract of South American plant Ilex macrophylla. This finding offers the possibility of identifying other pacitaxel derivatives from different sources and shows the effectiveness of the screening method at identifying antimitotic compounds. In addition, the interesting observation that ilexsaponin BI (73) enhances the antimitotic activity of 74 and 75 warrants further investigation. Several new ingenamine alkaloids 56 to 60 were isolated from the marine sponge Xestospongia ingens guided by a traditional cytotoxicity bioassay. Ingenamines I (57) and K (59) have novel functionality which has not been previously found in the ingenamine family. Ingenamine I (57) has a methyl branch in one of aliphatic chains. Ingenamine K (59) has a conjugated diene and an allylic alcohol on one of the aliphatic chains. The origin of these functional groups in ingenamine alkaloids will be an interesting topic i future biosynthetic studies of the 3-alkylpiperidine originated alkaloids. 

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:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0061302/manifest

Comment

Related Items