Antineoplastic agents. 168. Isolation and structure of axinohydantoinl GEORGE R . PET TIT,^ CHERRY L. HERALD, JOHN E. LEET, RAJESH GUPTA, DANIEL E. SCHAUFELBERGER, ROBERT B . BATES,~ PAUL J. CLEWLOW, DENNIS L. DOUBEK, KIRK P. MANFREDI, KLAUS RUTZLER,~ JEAN M . SCHMIDT, LARRY P. TACKETT, FRANKLIN B. WARD, MICHAEL B R U C K , ~ AND FERNANDO CAM0U3 Cancer Research Institute and Department of Chemistry, Arizona Stnte University, Tempe. AZ 85287, U.S.A. Received July 4, 19895 GEORGE R. PETTIT, CHERRY L. HERALD, JOHN E. LEET, RAJESH GUPTA, DAN~EL . SCHAUFELBERGER, ROBERT B. BATES, PAUL J. CLEWLOW, DENNIS L. DOUBEK, KIRK P. MANFREDI, KLAUS R~~TzLER, JEAN M. SCHMIDT, LARRY P. TACKETT, FRANKLIN B. WARD, M~CHAEL BRUCK, and FERNANDO CAMOU. Can. J. Chem. 68, 1621 (1990). Western (Palau) and Eastern (State of Truk) Caroline Islands and Papua New Guinea sponges of the genera Axinella and Hymeniacidon were found to contain the cytostatic (PS EDso 2.5 and 2.0 pg/rnL) and antineoplastic (PS TIC 143 at 3.6 mg/kg and T/C 138 at 3.6 rng/kg) pyrrologuanidines l a and 16. The related hydantoin 2, designated axinohydantoin, was also isolated from an Axinella sp. and its structure was assigned by X-ray crystallographic techniques. Present experience with sponges in the Axinella and Hymeniacidon genera suggests that the previously known hymenialdisine (1 6) and analogous imidazole derivatives may be widely distributed among these and related orange colored Porifera. Key words: axinohydantoin, hymenialdisine, Axinella, Hymeniacidon, cystostatic. GEORGE R. PETTIT, CHERRY L. HERALD, JOHN E. LEET, RAJESH GUPTA, DAN~EL . SCHAUFELBERGER, ROBERT B. BATES, PAUL J. CLEWLOW, DENN~S L. DOUBEK, K ~ R K P. MANFREDI, KLAUS RUTZLER, JEAN M. SCHMIDT, LARRY P. TACKETT, FRANKL~N B. WARD, M~CHAEL BRUCK et FERNANDO CAMOU. Can. J. Chem. 68, 1621 (1990). On a trouvt que les Cponges du genera Axinella et du Hymeniacidon des iles Caroline occidentale (Palau) et orientale or tat de Truk) ainsi que de la Nouvelle GuinCe contiennent des pyrrologuanidines l a et 16 qui sent des cytostatiques (PS EDSo 2,5 et 2,O pg/mL) et des antinCoplasiques (PS TIC 143 ?I 3,6 rng/kg et TIC 138 ?I 3,6 rng/kg). A partir d'un Axinella sp., on a aussi is016 l'hydanto'ine apparentCe 2, appelte axinohydantoi'ne, et on a dCteminC sa structure ?I l'aide de la diffraction des rayons-X. L'expCrience acquise avec les Cponges de 1'Axinella et de 1'Hymeniacidon genera suggkre que 1'hymCnialdisine ( lb) , qui Ctait connue antkrieurement, ainsi que les dCrivCs irnidazoles analogues sont peut-&tre trks rCpandus dans ces Cponges et dans les Porifera apparent& de couleur orange. Mots clis : axinohydantoi'ne, hyrnknialdisine, Axinella, Hymeniacidon, cytostatique. [Traduit par la revue] Introduction Early (2) in our evaluation of marine animals as new sources of potentially useful anticancer drugs, good leads were uncov- ered among the Porifera, and this initial (1966- 1968) promise is now being amply realized (3 ,4) . In 1979 in Palau we collected a Hymeniacidon species (at -40m) and an Axinella sp. that provided extracts with confirmed levels of activity against the U.S. National Cancer Institute's (NCI) murine P388 lympho- cytic leukemia (PS system). Other PS active sponge collections were completed in 1981 (Papua New Guinea) and 1985 (Truk, Federated States of Micronesia) that included Axinella carteri (Dendy) and a Hymeniacidon species. Initial extracts of each sponge were found to provide a confirmed level of activity against the PS system. By means of PS (in vitro) bioassay guided separation procedures, these sponge species led to two cytostatic and antineoplastic alkaloids ( l a , b ) accompanied in the case of Axinella sp. by a closely related, but marginally (PS ED50 18 @g/mL) inactive, com- ponent (2). The PS active marine alkaloids proved to be identical6 with the known (5-7) hymenialdisine ( l b , ref. 6 , PS or contribution 167 refer to ref. 1. 2 ~ u t h o r to whom correspondence may be addressed. 3~epartment of Chemistry, University of Arizona, Tucson, AZ 85721, U.S.A. 4~a t iona l Museum of Natural History, Smithsonian Institution, Washington, DC 20560, U.S.A. '~evision received March 27, 1990. 6 ~ y comparison with authentic specimens provided by Dr. I. Kita- gawa (see ref. 6). 2.0 p,g/mL and T / C 138 at 3 .6 mg/kg)7 and its debromo derivative l a (refs. 5-7, PS 2.5 pg /mL and T I C 143 at 3.6 mg/kg). 0 l a , R = H debromohymenialdisine 2 1 h, R = Br hyrnenialdisine axinohydantoin The unequivocal X-ray crystal structure of hymenialdisine was nicely established by Cimino e t a l . (5) and reconfirmed in the following year by the Kitagawa group (6). In turn, these advances simplified characterization of the companion sub- stance from Axinella sp. , herein named axinohydantoin (2), as a closely related compound. But establishing the exact geometri- cal configuration for its hydantoin-lactam sp2 bond required the following crystal structure determination. Axinohydantoin (2) crystallized from methanol as yellow prisms, which corresponded to CllH9BrN403 (by hreims) and with one mole of methanol. The structure was solved using 71nterestingly, hymenialdisine was previously active in the KB cell line, but inactive employing the P388 leukemia: cf. ref. 5. Perhaps the initial negative results were due to the sparingly soluble properties of this pyrrologuanidine. Prinled In Canada i Imprim6 au Canada 1622 CAN. J. CHEM. I 1 FIG. 1. ORTEP view of a single molecule of 2, with 50% thermal ellipsoids. 1 I MULTAN (8) and refined to R = 0.054 using anisotropic I temperature factors for Br and oxygens other than 0-3, and isotropic temperature factors for the other non-hydrogens and I for hydrogens (unrefined) in calculated petitions. HN- 1, HN- 1 1, and HN-9 were calculated to be 0.95 A along a line to the respective oxygen. These positions differed very little from those calculated assuming bonding to trigonal atoms. HN-4 has been shown at the calculated position assuming a trigonal N-4 that is too far (2.4 A) from the closest 0-10 for significant hydrogen bonding; it may actually bend somewhat toward this 0- 10. The structure deduced for axinohydantoin (2, Fig. 1) was found to be closely related to that of hymenialdisine ( lb) , with reversal of configuration at the C7-C8 double bond being the most interesting difference. In turn, this suggested that axino- hydantoin was not simply a hydrolysis product of guanidine l b . The most prominent bond length difference !etween the two structures occurs at S10-010, with 1.23 A in hydantoin 2 compared to 1.33 A for C I 1 -N 1 1 in 1 b. No significant differences in bond angles were observed. An angle of 36" was observed between the least-squares planes of the two nearly planar five-membered rings in hydantoin 2, compared to 43.8' in guanidine l b . In both cases, the seven-membered ring has adopted a boat conformation with C-5 at the prow, and similar torsion angles except for C2-C3-N4-C5 expanding from - 10.5" in 1 b to - 15" in 2, and C2-C13-C7-C6 contracting from 41. 1" in l b to 31" in 2. The twist angle C13-C7-C8- C12 about the carbon-carbon double bond increases from 0.5" in l b to 10" in 2, presumably to relieve the steric interaction between 0-12 and HC-14. The arrangement of intermolecular hydrogen bonds govern- ing the packing in hydantoin 2 (see supplementary material)8 was found to be completely different than that in guanidine 1 b (6). The hydantoin ring in each molecule was found linked to the hydantoin rings of two other molecules via base-pairing interactions across centers of symmetry. The pyrrole NH proved to be hydrogen bonded to the methanol solvate oxygen and in turn to 0-3. Only a few substances (9-12) with a hydantoin system have been isolated from sponges, and one of these, midpacamide, found by Scheuer and co-workers (9) in an unidentified Marshall Island sponge, may be biogenetically related to axinohydantoin. From evidence now in hand, pyrroles 1 and 2 and related substances may prove to be ubiquitous Porifera biosynthetic products. Experimental General methods Marine sponge taxonomic identification was performed in the Smithsonian Institution where voucher specimens are deposited in the collections of the Department of Invertebrate Zoology, National Museum of Natural History. All solvents employed were redistilled. Size exclusion chromatography was accomplished with Sephadex LH-20 (particle size: 25-100 km) suppled by Pharmacia Fine Chemi- cals, Uppsala, Sweden. Thin-layer chromatography was carried out with silica gel GHLF Uniplates (Analtech Inc.) and with RP-8 precoated plates (layer thickness: 0.25 rnrn) from E. Merck, Darm- stadt, Germany. High-speed countercurrent chromatography was accomplished with an Ito multilayer coil extractor-separator (P.C. Inc., Potomac, MD) using 2.6 mm i.d. tubing, and an FMI Lab Pump. Melting points are uncorrected and were determined on a Kofler-type hot-stage apparatus. Ultraviolet spectra were recorded employing a Hewlett-Packard model 8450 uv/vis spectrophotometer and ir spectra with a Nicolet ft-ir model MX-1 instrument. The nrnr spectra were measured in DMSO-d6 using a Bruker AM-400 instrument and are recorded in ppm downfield to TMS (assignments bearing the same superscript may be reversed). The "C nmr multiplicities were deter- mined with APT experiments based on an average coupling constant of 135 Hz. The eims spectra were recorded with a Kratos AEI 5076 spectrometer at the NSF Regional Facility, University of Nebraska, Lincoln, Nebraska. Palau Porifera (Axinella sp. and Hymeniacidon sp.) collection and extraction. The initial collection of Axinella sp. (Demospongiae class, Axinel- lida order, Axinellidae family) in Palau, Western Caroline Islands, was conducted in May, 1979. The sponge displayed a brownish-yellow ex- terior of irregular mass. A 2-propanol-CHCI3 extract gave confirma- tory in vivo activity with PS T/C 201 at 100 mg/kg, PS ED,o = 2.5 pg/rnL. A scale-up re-collection (220 kg wet wt.) of this sponge was completed in March, 1985, and preserved in 2-propanol. The preserving solution was separated from the sponge, concentrated to an aqueous slurry, and extracted with CHzClz ( 13). The remaining sponge material was re-extracted with 2-propanol-CHzClz (1:l); the extract was separated, solvent removed, and the residue partitioned between CH2C12 and HzO. At this early stage a solid precipitate appeared at the CH2C12-H20 interface. The precipitate was separated and amounted to 8Tables of observed and calculated structure factor amplitudes, cal- culated hydrogen coordinates, isotropic temperature factors, bond lengths and angles, torsion angles, and a packing diagram may be pur- chased from the Depository of Unpublished Data, CISTI, National Research Council of Canada, Ottawa, Ont., Canada KIA OS2. Tables of positional parameters and bond distances have also been deposited with the Cambridge Crystallographic Data Centre, and can be obtained on request from The Director, Cambridge Crystallographic Data Centre, University Chemical Laboratory, Lensfield Road, Cam- bridge CB2 IEW, U.K. ET AL. 1623 1.8 kg (PS T/C 227 at 294 mg/kg, EDs0 2.8 pg/mL). Analogous solid fractions were also obtained at this initial CH2C12 step during separations of the sponge extracts summarized below. The orange sponge Hymeniacidon sp. (Demospongiae class, Hali- chondrida order, Hymeniacidonidae family) was collected (1979), re-collected (in 1985, 218 kg wet wt.), and extracted (2-propanol extract showed PS T/C 130 at 5.5 mg/kg and 27 kg/mL) as summarized above. Removal of solvent from the initial 2-propanol extract led to an aqueous concentrate that contained 1.2 kg of a solid fraction with PS = 3.7 pg/mL. Isolation of hymenialdisine (Ib) and axinohydantoin (2 ) A 10 g aliquot from the 1.8 kg of solid precipitate noted above was dissolved in CH30H (400mL) and separated by size exclusion chromatography on a column of Sephadex LH-20 (100 X IOcm) to yield two major marine alkaloid fractions. When fraction 1 (elution volume: 12.0-12.6 L) was allowed to stand for 24 h at room tempera- ture, axinohydantoin (2) slowly crystallized as yellow needles (30 mg): mp >350?C; tlc on silica gel Rf = 0.83, I-BuOH-AcOH 50% (955); tlc on RP-8 R f = 0.51, CH30H-AcOH 5% (I:]); uv (CH30H) A,,: 264sh (log E = 3.88), 345 (log E = 4.16) nm; ir (KBr) v,,,: 1740, 1702, 1638, 1480,1425, 1407 cm-I; 'H nrnr (DMSO-d6) 6: 2.67 (2H, m, H-6), 3.22 (2H, m, H-5), 6.66 (IH, s, H-14), 7.89 (IH, t, HN-4), 9.83, 10.91 (2 X IH, s, HN-9,HN-Il), 12.35(1H,s,HN-1); 13Cnmr (DMSO-d6) 6: 36.2 (t, C-6), 38.5 (t, C-5), 101.6 (s, C-15), 113.9 (d, C-14), 120.0 (s, C-13), 121.2 (s, C-7), 125.5 (s, C-2), 126.5 (s, C-8), 153.8 (s, C-lo), 162.7 (s, C-3), 163.3 (s, C-12); hreims mlz: 325.9842 and 323.9834 (CllH9N403Br equires 325.9836). Fraction 2 (elution volume: 12.9-13.5L) yielded a crystalline precipitate (100 mg), which was identified as hymenialdisine ( lb ) by comparison (uv, ir, 'H nmr, eims) with an authentic sample (6). Truk Porifera (Axinella carteri) collection and extraction In May 1985, approximately l kg of an orange-yellow sponge subsequently identified as Axinella carteri (Dendy), was collected in the Truk Lagoon, Federated States of Micronesia, at - 3 to -24 m. The preserving solution (2-propanol) was removed and this extract proved toxic down to 50mg/kg against the PS leukemia. The 2-propanol extract was partitioned between CHzC12 and H20 and the resulting CHzC12 extract was successively partitioned (13) between 9: 1- 4: 1- 1: 1 MeOH:H20 with hexane-CC14-CH2C12. The final CH2Cl2 extract showed PS T/C 135 at 100 mg/kg and PS cell line EDs0 = 1.2 pg/mL. In October 1985, approximately 148 kg (wet wt.) of the sponge was recollected and preserved in MeOH. The MeOH solution was decanted, and the sponge was ground and extracted with MeOH:CH2C12 (1:l). The original MeOH solution was concentrated to an aqueous phase and extracted with CH2C12 (3X) followed by 1-BuOH. Study of this 1-BuOH fraction was discontinued when PS results showed minimal activity. When the ambient temperature extraction of the sponge with MeOH:CH2C12 was completed, the aqueous MeOH phase was sepa- rated and concentrated to an aqueous phase, which was extracted with 1-BuOH (15 L). The 1-BuOH phase was concentrated, redissolved in MeOH (1.5 L), and dried to give a 232 g fraction (PS EDs0 1.4 pg/mL). A 97 g aliquot of the MeOH soluble fraction was treated with I-BuOH (800 mL, 50?C, 12 h) and the relatively insoluble part (50g, PS EDs0 1.5 pg/mL) was collected. The MeOH (600 mL) sparingly soluble portion weighed 4.26 g (PS EDs0 0.11 pg/mL). Papua New Guinea Porifera (Hymeniacidon sp.) collection and extraction The collection (May 1981, near Motapure Island, Papua New Guinea) and recollection (October 1983, 44 kg wet wt.) of an orange Hytneniacidotl sp. as well as the large scale extraction (crude extract PS T/C 136 at 100 mg/kg and EDs0 24 pg/mL) and solvent partitioning were performed as described above for A . carteri. In this case, when the 934 g initial CH2C12 fraction was subjected to further separation by MeOH:H20 with the hexane + CC14 + CH2C12 sequence, a total TABLE 1. Positional and thermal parameters Atom x Y z B (A') *Starred atoms were refined isotropically. Anisotropically refined atoms are given in the form of the isotropic equivalent thermal parameter, defined as 8.rr2(Ull + U22 + U33)/3. of 135 g (PS EDs0 14 p g / r n ~ ) of a solid interfacial fraction was col- lected and used to isolate hymenialdisines l a and l b . Isolation of hymenialdisines l a atzd Ib-Procedure A An aliquot (250mL) of the preceding Axinella carteri MeOH (600 mL) solution was applied to a column of Sephadex LH-20 (1.9 kg in MeOH). A total of 460 fractions of 20 mL each were collected and a fraction weighing 0.73 g (PS EDSo 2.2 pg/mL) was further separated using high speed countercurrent distribution with an Ito coil. A 50 mg aliquot was applied (6 mL) in I-BuOH:HOAc:H20 (4:1:5) to the coil with the I-BuOH phase as stationary (upper) and the aqueous part as mobile (lower) phase. Fractions (120) of 6.5 mL each were collected; fractionation was monitored with ultraviolet detection (254 nm). The fractions were neutralized (pH 7) with aqueous NaOH and refrigerated. Debromohymenialdisine l a , 9 mg, PS EDSo = 3.0 pg/mL, crystal- lized from fractions 28-33 and was identical (tlc, ms, nmr) with an authentic sample (6). The MeOH less soluble fraction (4.268) described above was extracted with MeOH (5 X 25 mL) at 40?C and the solution filtered to give 3.73 g of residue. A 0.90g portion was triturated with DMSO (1OmL). The soluble portion (0.25 g) was chromatographed on a column of Sephadex LH-20 in MeOH to provide 0.13 g of hymenialdi- sine ( lb ) as yellow crystals (PS EDs0 0.62 pg/mL), identical (tlc and ms comparisons) with an authentic sample (6). Procedure B The 1.2 kg fraction (see above) from the Palau Hymetliacidon sp. was further separated by successive Soxhlet extraction (20 g aliquot) with CH2CI2 (6 X 5 L), EtOH (6 x 5 L), and I-BuOH (6 X 5 L) to give respectively 35 g (PS EDso 26 pg/mL), 500 g (PS EDs0 8.6 pg/mL), and 106 g (PS EDSo 2.6 pg/mL) fractions. A 10 g sample of the 1-BuOH fraction in MeOH was subjected to chromatography on a column of Sephadex LH-20 (500 g) to give 26 individual (by tlc comparisons) fractions using 4:1 CH2C12:MeOH. Of these, 56 mg proved to be largely debromohymenialdisine l a (PS EDs0 1.4 pg/mL) and hymenialdisine (5.5 mg, 1 b, PS EDso 7.5 pg/mL) by comparison nmr and tlc. Procedure C The 135 g fraction from the Papua New Guinea Hymetliacidon sp. 1624 CAN. 1. CHEM. \ was extracted (Soxhlet procedure with two stainless steel 1-gallon extractors) with EtOH to yield a 22 g alcohol soluble fraction. Treatment of this fraction with CH2C12:CH30H (1:l) yielded a precipitate (4.3 g, PS EDso 8.5 p,g/mL), which was extracted with hot 1-BuOH. The 1.5 g 1-BuOH soluble fraction was preabsorbed onto silica gel and separated by chromatography on a column (3 X 62 cm) of silica gel (180 g). Gradient elution with 95:s CH2C12:CH30H with increments of MeOH provided fractions that yielded (0.17 g and 0.06 g respectively) debromohymenialdisine ( l a , PS T/C 143 at 3.6mg/kg and EDso 2.5p,g/mL) and hymenialdisine ( lb , PS T/C 138 at 3.6 mg/kg and EDso 2.7 p,g/mL). Both l a and l b were identified by direct compari- son with authentic samples (6) employing tlc, I3c and 'H nrnr, ms, uv, and ir spectral data. X-ray structure of arinohydantoin (2) Monoclinic, C2/c, a = 19.558(2), b = 7.505(1), c = 19.092(3) A, P = 103.78(1)", V = 2754.3 A3, Z = 8, p, = 1.72; Nicolet P2, diffractometer, crystal 0.17 X 0.13 X 0.08 mm, 23"C, MoK,, , X = 0.71073 A, 20/0 scans, 20,,, 50"; 836 of 2446 reflections with F: > 30 (F,,') used; solved by direct methods; R = 0.053, R, = 0.055 excluding unobserved reflections. Coordinates and isotropic tem- perature factors of non-hydrogens are given in Table Acknowledgments We are pleased to acknowledge the very necessary financial assistance provided by the Fannie E. Rippel Foundation, the Arizona Disease Control Research Commission, Outstanding Investigator Grant CA 44344-0lAl and PHs Grants CA-16049- 07-12 awarded by the National Cancer Institute, DHHS, the Robert B. Dalton Endowment Fund, Virginia Piper, Polly Trautman, Eleanor W. Libby, the Donald Ware Waddell Foundation, Herbert and Dianne Cummings (The Nathan Cummings Foundation, Inc.) Mary Dell Pritzlaff, the Olin Foundation (Spencer T. and Ann W .), and by the U.S. Army Medical Research and Development Command under Grant No. DAMD17-89-2-9021. Other helpful assistance was provided by the Governments of Papua New Guinea (N. Kwapena, A. Richards, and Drs. John L. Munro, J. M. Lock, and N. Polunin), the Federated States of Micronesia (Truk, D. E. Aten, R. Killion, and A. Amaraich), and Palau (Dr. T. Paulis and K. B. Batcheller), Singapore Airlines Ltd. (M. Theam Kong), Blue Lagoon Dive Shop (MOEN, Truk), Drs. C. G. Bass, C. Dufresne, Mr. G . R. Pettit 111, Mrs. D. N. Tackett, S . Taylor, and T. N. Trautman, the Smithsonian Institution Oceanographic Sorting Center, The U.S. National Science Foundation (Grant CHE-8409644), The Swiss National Science Foundation (to D.E.S.), and the NSF Regional Instrumentation Facility in Nebraska (Grant CHE-8620 177). 1. J. A. MCBAIN, G. R. PETTIT, and G. C. MUELLER. Cell Growth and Differentiation, 1, 281 (1990). 2. G. R. PETTIT, J. F. DAY, J. L. HARTWELL, and H. B. WOOD. Nature, 227, 962 (1970). 3. Y. KATO, N. FUSETANI, S. MATSUNAGA, and K. HASHIMOTO. J. Org. Chem. 53, 3930 (1988). 4. Y. HIRATA and D. UEMURA. Pure Appl. Chem. 58,701 (1986). 5. G. CIMINO, S. DEROSA, S. DESTEFANO, L. MAZZARELLA, R. PULITI, and G. SODANO. Tetrahedron Lett. 23, 767 (1982). 6. I. KITAGAWA, M. KOBAYASHI, K. KITANAKA, M. KIDO, and Y. KYOGOKU. Chem. Pharm. Bull. 31, 2321 (1983). 7. F. J. SCHMITZ, S. P. GUNASEKERA, V. LAKSHMI, and L. M. V. TILLEKERATNE. J. Nat. Prod. 48, 47 (1985). 8. P. MAIN, S. J. FISKE, S. E. HULL, L. LESSINGER, G. GERMAIN, J. P. DECLERQ, and M. M. WOOLFSON. MULTAN 80. A system of computer programs for the automatic solution of crystal structures from X-ray diffraction data. University of York, England, and Louvain, Belgium. 1980. 9. L. CHEVOLOT, S. PADUA, B. N. RAVI, P. C. BLYTH, and P. J. SCHEUER. Heterocycles, 7, 891 (1977). 10. R. KAZLAUSKAS, P. T. MURPHY, R. J. QUINN, and R. J. WELLS. Tetrahedron Lett. 1, 61 (1977). 11. P. DJURA, D. B. STIERLE, B. SULLIVAN, D. J. FAULKNER, E. ARNOLD and J. CLARDY. J Org. Chem. 45, 1435 (1980). 12. D. J. FAULKNER. Nat. Prod. Rep. 3, 1 (1986). 13. G. R. PETTIT, Y. KAMANO, R. AOYAGI, C. L. HERALD, D. L. DOUBEK, J. M. SCHMIDT, and J. J. RUDLOE. Tetrahedron, 41,985 (1985).