Tetmhedron Vol. 47. No. 48. pp. 10101-10108.1991 caO4020/91 Priitcd in GreatBritain Pcrgamon Rcss plc $3.00+.00 The Structure of Laurobtusol, a New Rearranged Sesquiterpenoid from the Mediterranean Red Alga Luurenciu obtusa Salvatore Caccamesea, Vincenxo Amicoa, Placid0 Nerib and Mario Fotib *D@iimento di Sci~ kituto Chimiche, Universits di Calania, VJe Doria 6.95125 CNR Sostanze Natmli, Vii de1 Santuario 110.95028 Valh, fhtauh, Cata&, Italy Italy (Received in UK 18 October 1991) zyxwvutsrqponmlkjihgfedcbaZYXWVU AbstraeC An alcohol witha new m.cy clic hwnulaneskeleton, laurobtusol.was isolatedfrom the M editerranean Red Alga Laurencia obtusa. The structure was establishedmabdy by 2 D NM R medwds and the relative con&uadon was assigned by a quantitativecomputer &udadon ofthe lanthani& induced shaftsin the ‘H NM R spectrum and Molecular Mechanics calculation (M M ?). zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGF zyxwvutsrqponmlkjihgfedcbaZYXWVU The red alga Lauren&u obtwa continues to be a prolific source of metabolites of diverse nature, particularly terpenes and acetogenins. *z This is due to the dominance of one major metabolite associated with several minor components and to a marked variability of the constituents in the collections from different sites.3 We recently reported the isolation of several new representatives of the rare brasilane zyxwvutsrqp (1) clas8.4’5 These minor compounds were present with a major acetylenic cyclic ether laurencienyne (2). previously characterixed.6 We now report the isolation and structural elucidation of another very minor component, laurobtusol 3 (3), which possesses an unreported carbon ring system probably derived from the a-humulene Column chromatography and repeated preparative HPLC of selected fractions afforded compound 3 (0.07 % extract, m.p. 121-3 ‘C) and laurencienyne as Ci5HzeO (3 unsaturations) established the hydroxyl skeleton. 2. The molecular formula of 3 was established on the basis of HREIMS and 13C NMR data. IR absorptions (3300. broad) and indicated 10101 the absence of a carbonyl functionality. (cm-‘) The S. zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLK CACCAM ESEet al. 10102 13C NMR spectrum of 3 confii the presence of a carbon bearing an hydroxyl (S 72.3) and the absence of sp2 carbons. These combined data required that 3 possesses a tricyclic skeleton. The ‘H NMR spectrum (CDC13, Table 1) indicated the presence of a cyclopropane ring (g 0.10, 1 H and 6 0.79, 1 H). In addition, it showed a hydrogen on a carbon bearing oxygen (S 3.90, dd, J=2.5, 11.4 Hz), three methyl singlets (6 0.80, 0.96, 1.23) and a methyl doublet (S 1.08. J=6.5 Hz). The remaining 11 protons gave multipiet D COSY experiment showed only a few unequivocal proton-proton 1. The total assignment and by straightforward some connectivities signals congested within the region from 1.00 to 1.90 ppm and also a 2 of the spectrum correlations, as reported in Table was made resorting to clear one-bond HETCOR correlations 13C NMR DEPT signals. HRTCOR long-range experiments (Table 1) indicated regarding two and three-bond correlations. Firstly, the cyclopropane H-2 protons (6 0.79 and 0.10, d), appear to be only geminally coupled (J-J15 Hz): this methylene is thus bonded to two quaternary carbons. One of them (the C-l signal at 6 28.1) correlates to both H-2 protons and the other one (the C-3 signal at 6 22.1) correlates with 3H-12. as observed in the long-range HETCOR experiment. Methylene C-11 (6 35.9) shows a Table 1. ‘H and % Pos. 6c DEPT 1 28.1 c 22.1 CHz NM R data for compound 3.’ 6~ mult. (J, Hz) COSY H/C zyxwvutsrqponmlkjihgfedcbaZY lon&range correlationb 2H-2, 3H-1 Ha 0.79 d (4.5) Hb-2 Hb 0.10 d (4.5) Ha-2 3H-12 C 3812 38.5 zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA 5 6 7 72.3 CH 47.4 CH2 30.9 2H-5, I-I.,-2 3.90 dd (11.4, 2.5) Ha 1.16 me Hb 1.05 me H-3 1.23 mc Hb 1.10 mc 3H-13, 3H-14 C 44.3 3H-13, 3H-14, h-5, 8 47.1 CH 1.18 mc 9 44.7 CH 1.44 mc 10 33.9 CH2 11 35.9 CH2 Ha I Hb 3H-13, 3H-14 Hr2. 3H-15 1.70 me 1.86 mc Hb 1.20 mc 12 20.9 CH3 1.23 s 13 30.0 CH3 0.80 s 14 29.6 CH3 0.96 s 15 19.2 CH3 1.08 d (6.5) 3H-15 3H-15 3H-15 1.25 mc Ha It-7 b-2, Ha-10 2H-2, Ha-IO, 3H-12 2H-2 H.-J, 3813 3H-14 Ha-5, Ht,-7, 3H-13 H-8, Hb-IO “The chemical shifta are given in ppm downfield from tetramethylsilane. blong-range correlations were obtained with polarization transfer optimized for M O.0 and 5.0 Hz. % hemical shift derived by inspection of the relative cross-sections in one-bond HRTCOR spectrom F1decoupled. 10103 Structure of laurobtusol three-bond correlation with 2 H-2 and therefore has to be linked to C-l or to C-3. But, since the 3 H-12 (8 1.23) are correlated with C-2 zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGF a nd with C-11, as observed by long-range HETCOR, fragment A can be assembled fragments (the positions 1 and 3 marked by an asterisk are interchangeable). B and C are built on the basis of the connectivities elucidated Similarly, through the long-range HETCOR experiments. Further structural immediate assembly fragments of these was however not by COSY experiments due J$..e,* to the crowding and partial overlapping A0 zyxwvutsrqponmlkjih +$zz 11 of the methylene and methine multiplets A B C C-5, and C-7 through C-11. Therefore, we resorted to incremental addition of the lanthanide shift reagent (LSR) Eu(fod)s to obtain a series of ‘H and 13C Nh4R spectra of compound 3. A good linear relationship between the observed induced shifts of many proton signals (LIS) and the [LSR] / [3] was obtained for the range 0.06-0.6 mole ratio. DEPT experiments were done to assign unequivocally the signals where the 13C resonances intercross each other, due to the differing velocities of their shifts, during successive additions of the LSR. On the completely resolved doped ‘H NMR spectrum ([LSR] / [3] molar ratio 0.7) one-bond HETCOR experiment COSY experiment was done to assign unambiguously afforded the observation all proton signals. In addition, a doped of the coupling pathways of all protons. These data are shown in Table 2 where INAPT data confirming the long-range undoped HETCOR experiments are also reported. Experimental evidence supporting the C-4/C-5 connectivity is provided by the COSY correlations in Table 2 between H-4 and 2 H-5 and viceversa, thus affording the extension from fragment B to fragment D. On the other hand, a long-range HETCOR undoped experiment (Table 1) shows key C-4/I&-2 correlation which affords the combination of structural fragments D and A to give substructure E. This in turn can be linked to fragment C to give substructure F consistently with the detection of a COSY correlation between H-11 and H-10 in the doped spectrum (Table 2) and a weak C-11/I&10 long-range correlation six-membered in the undoped spectrum (Table 1). At this point, formation of the two rings now gives structure 3 apart from stereochemistry. The assignment of the relative stereochemistry to structure 3 was a challenging task. The signals of H-8 and H-9 in the NMR spectrum were buried in the 6 1.1-1.4 envelope, thus a NOESY experiment was not attempted. We resorted instead to a detailed study of the Eu(fod)g doped ‘H S. CACCAMESE et zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQ al. 10104 NM R factors zyxwvutsrqponml (GF) spectra to obtain the observed LIS and to calculate them from the geometrical of eleven protons that give identifiable signals during the incremental GF are related to relative stereochemistry among the eight diastereomeric addition of the LSR. Since the it was possible to select the most probable configuration structures (of course the cyclopropane attachment sites 1 and 3 are cis each other and the enantiomeric pairs are not considered). procedure The “‘q;;” considers M:%i,2 M:@;;,2 both the angle 0-Eu-H (x) and the Eu-0 OH2 OH2 E F D distance (r) according to the McConnell-Robertson equation Av=K[(3 cos’x-1) i3] where K is the pseudocontact Av is the observed chemical shift and GF is in brackets. For all eight configurations constant, the position of the Europium atom with respect to the lone pair of the coordinating Oxygen was optimized analogously to a brasilane sequiterpenoid whose structure was elucidated by us,’ Thus we defined the angle $ Table 2. ‘H and 13C NMR data for compound 3 iu the presence of Eu(fod)x Pas. 6c DEFT 1 33.0 C 2 28.1 CHz 3 41.6 zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA C 4 104.0 CH2 6 34.4 C 7 51.4 CH2 a 51.1 9 46.7 CH CH 10 35.4 CH2 11 37.7 CH2 12 14 15 INAPTb C-l 8.11 d (4.0) Hb-2 Hb 5.06 d (4.0) Ha-2 Ho 56.7 COSY & CH 5 13 8~ mult. (J, Hz) Hb HE Hb 20.44 dd (11.0, c) 2H-5 12.89 dd (13.0, c) 10.61 dd (13.0, 11.0) H-4, HI,-5 H-4, Ha-5 4.74 dd (13.2, 11.4) 3.25 dd (13.2, 6.6) 4.89 ddd (11.4, 6.6, c) 3.30 m H-8, 2H-10, 3H-15 3.09 m H-9, 2H-11 Hb-7, H-8 Ha-7, H-8 2H-7, H-9 4.43 3.86 ddd (11.7. (11.7, 3.0, 8.0, 7.6) c) 2H-10, Ha-1 1 29.3 32.1 CH3 CH3 7.74 s 2.66 s C-6, C-7 31.6 20.6 CH3 2.24 s 2.37 d (6.5) C-6, C-7 H-9 CH3 ‘[Eu(fod)31/[31= 0.8 molar ratio. bCarbons observed upon irradiation of the indicated proton. %uesolved by broadening. C-l, C-3, C-6, C-7, C-9, C-15 c-15 2H-10, Hb-11 I Ha Hb C-6 C-l. c-2, c-3, c-9 C-8, C-9 Structure 10105 of laurobtusol Eu-C&O 80’. the dihedral angle 0 Eu-O-C&II4 55’ at the distance Eu-0 3.5 A. The constant K was derived by a least squares minimization of the observed Av the GF for eleven protons. Hence, the theoretical shift for each proton was calculated. The difference between observed and calculated LIS for the zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA i protons was then expressed via the Hamilton agreement factor (AFj7 where the L&bed was the slope of the relationship Av vs [LSR] / [3]. For F( LIShd AF= y i - L&u i )2 calculation to optimize $ and 0 we applied a simulation ” computer method successfully used by one of us to study L&bad D2 a number of structurally rigid and flexible a&unsaturated carbonyl systems. *,’ This method uses as input data the atomic coordinates and the LIScbsd for each studied proton. The procedure is repeated for the molecular geometry corresponding to each diastereomer. The minimum value of AF defines the correct relative configuration. The observed and simulated LIS and the AF for the eight possible diastereomeric structures are given in Table 3. The flexibility of both six-membered rings introduced a complexity in the calculation of the Table 3. Measured and simulated LIS for the eight possible diastereomeric structm~ of compound sash ttc ld Proton tee zd ctc3= ccc 4’ tttse LI&kd LIsc.lca L&&d 7.47 6.74 tctQ cct 7h cttsb LIS&d L&&d LIS&d LIS&d Ha-2 5.87 a-2 4.13 H-4 Ha-5 (0.11) 6.90 LiSti 7.05 2.65 LISCdkd 2.64 2.50 2.50 3.06 (0.06) 3.49 3.56 3.96 3.54 3.31 3.23 2.90 16.20 (0.90) 16.01 16.03 15.50 14.07 17.72 17.66 16.44 15.99 9.42 (0.09) 9.13 9.14 9.82 I.96 3.60 3.56 3.02 2.94 Hb-5 7.98 (0.09) 7.92 7.88 7.95 12..13 6.88 7.04 8.78 8.93 H-8 2.90 (0.02) 2.65 2.75 3.46 3.21 3.47 3.51 6.78 6.22 H-9 2.00 (0.04) 1.84 1.15 2..13 1.46 4.07 2.58 2..27 2.41 Ha-11 2.20 (O-04) 2.30 2.27 2.14 1.65 3.41 3.14 2.50 3.90 Hb-11 2.13 (0.03) 2..31 2..25 1.77 1.46 2.31 2.20 2.08 2.46 H-12 5.58 (0.08) 5.77 5.66 4.33 3.12 4.87 4.59 3.78 4.72 H-15 1.14 (0.01) 0.76 1.30 0.96 0.98 1.86 3.93 1.89 1.46 AF’ KC 0.0603 721(14) 0.0713 720(16) 0.101 0.406 0.326 0.337 0.375 0.368 711(23) 634(89) 727(88) 794WJ) 734(94) 721(30) *Data reported in the 1st column indicate the observed molar induced shifts (LISOIXI; data in the following columns indicate calculated molar induced shifts (L&I.& for each diastavomeric structure. 4 or each diastereomer (l-8) fmt letter indicates a trans (t) or cis (c) relationshiphehveen hydroxyl(equatorial“up”)and H-8, second letter indicatesthe relationship between hydroxyl and Me-15, third letter indicate the relationship between hydroxyl and C-2. ‘In parenthesis estimated standard deviation. ‘king A chair, ring B half-chair, OH eq “up*. ’ Ring A hoat, ring B boat, OH cq “up”. fRing A chair, ring B half-chair, OH ax “up”. king A boat, ring B half-chair, OH ax “up”. bRing A chair, ring B boat, OH ax “up”. ‘Agreement factor. 10106 S. CACCAMESE et al. protons atomic coordinates. Thus, it was necessary to calculate the most preferred conformation the rings. To this end we carried out a computational analysis using the MacroModel program IoJ1 (MM2). The rigidity of the bridge C-3-C-8 involves some restraint to the conformations built; in fact, when the H-8 is tram to the cyclopropane group can assume only chair conformation of that can be C-2, the ring A which bears the hydroxyl with equatorial OH. To have chair conformation with axial OH, H-8 and C-2 must be cis each other. The preferred conformation for each possible diastereomer is given in Table 3. An examination and calculated of the results reveals that the largest deviations between observed LIS appear for the stereogenic centers of the molecule or for protons connected them. The AF resulting for the t,t,c configuration (0.0603, see Table 2 for the symbols lower with respect to the resulting AF for the seven other stereochemisties. possesses the configuration to used) is On this basis, laurobtusol depicted in formula 3. In particular, from Table 3 we find that the ratio R of the AF between the t,t,c and the t,c,c configurations is 1.18. Since our monodimensional hypothesis has nine degrees of freedom (11 observed protons-2 varying angles Q and 0 of the Eu position) the significance Tables in ref. 7 give us the experimental which the second hypothesis conformation (t,c,c configuration) for the correct diastereomer respect to other conformations are thermodynamically unfavored chair, cannot be rejected. Moreover, t,t,c has a much lower energy the most stable (38.5 Kcal mol-‘) with that strongly (A boat, B boat 41.4; A B boat half-chair R value at an 8 % level, that is the level at 41.1; A boat, B 40.6, in Kcal mol-‘). A computer-obtained stereoview of it is Fig. 1 shown in Figure 1. The AF calculated zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA for configurations possessing rings in higher energy conformations Laurobtusol with the a-humulyl resulted in any case significantly is presumably higher than those reported in Table 3. derived from humulene according to the following cation which undergoes consecutive 13-hydride Scheme starting shifts to give, after a proton elimination, a stabilized diene. This undergoes hydration and further double bond delocalization a plausible cyclopropane formation. Detailed biosynthetic studies on 1,3 hydride shifts to give tricyclic humulane were reported. l2 The 8 to 7 methyl migration is similar to that depicted in the trasformation poitanes.’ with For this reason, the numbering with that used for humulene. 13 of aromadendranes in system shown for compound 3 was chosen to coincide 10107 Structure of laurobtusol 1-H+ Recently, humulene and a monocyclic dihydroxyhumulene Laurencia obtusa.14 The presence of humulane geographical derivative derivative were isolated from japanese from this species collected areas stimulates the search for other representatives in different of this class in this prolific alga. EXPERITMENTAL General. Low and high-resolution instrument. EIMS (probe) 18 eV were obtained on a VG ZAB 2SE IR and UV spectra were recorded on a Perkin-Elmer mod. 330 spectrophotometers, AC-250 instrument respectively. mod. 684 and on a Perkin-Elmer ‘H and 13C NMR spectra were measured on a Bruker operating at 250.13 MHz and 62.9 MHz using CDC13 as solvent. Chemical shifts are quoted in ppm (8) relative to TMS. DEPT, COSY and HETCOR using standard Bruker microprograms. were performed Long-Range HETCOR was performed using the pulse sequence described in the literature.15 INAPT experiments were performed using the pulse sequence described 16.17 with standard Bruker microprograms using delays Al= A~=36 ms reports zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA in well known corresponding experiments to J=7 Hz. Optical rotation was determined with a Perkin-Elmer 141 polarimeter. Preparative Liquid Chromatography (PLC) was carried out on a Jobin-Yvon LC Miniprep instrument, using LiChroprep Si-60, 15-40 pm, (Merck) as stationary phase. Molecular Modeling and Force Field calculations (MM2) were performed on a Digital Vax-Station 3100/38 computer using MacroModel Version 2.5 program. Simulation using a computational Vax-Station of LIS spectra was performed program written in Fortran 77 language.5 The program runned on a Digital 2000 computer, using as atomic input coordinates the data obtained from MacroModel program. Collection, extraction and chromatographic separation. Laurencia obtusa was collected at S.CACCAMFSE et al. zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPON 10108 Castelluccio, 40 Kms South of Catania, eastern Sicily, in October 1988, in littoral zones. The fresh alga (ca. 15 kg wet wt.) was immediately soaked in iso-PrOH and left steeping for three months. The material was then filtered and the algal residue was repeatedly homogenized with iso-PrOH in a Waring blender and filtered. The exhausted dried powder residue weighed 720 g. The iso-PrOWwater solution was concentrated dried with NazS04, open column Fractions in and partitioned vacuum with En0 and NaNO3. The ether layer was and evaporated to give 30 g. of dark green oil. The extract was applied to an (3x100 cm) of Silica gel and eluted with increasing concentrations of 250 ml were collected and those exhibiting Fraction of Et20 in petrol. similar TLC profiles combined. 13 (440 mg) was subjected to PLC (gradient of CH2C!k?K!6Hi4 from 1:l to 3:2) to give laurencienyne (2) and impure laurobtusol (3) which was further purified (CH~CltiC6H14 3:2) to give pure laurobtusol Laurobtusol (21 mg, 0.07% extract). (3) was obtained as a white optically active powder; m.p. 121-123 ‘C, [al25 @) 28.7 (589). 28.9 (578), 33.1 (546), 57.2 (436), 90.6 (365) (c=O.9, EtOH); IR Vmax (liquid film) cm-l: 3300, 1455, 1360, 1010; HRMS [Ml: 222.1986 (calcd for Ci5H260 222.1983); MS 222 (lo), 209 (21), 207 (34), 204 (28), 189 (42), 180 (63), m/z (rel. int.) 165 (23), 162 (22). 161 (30), 153 (33), 150 (37), 139 (49), 136 (58). 125 (60), 109 (74), 95 (100); ‘H and 13C NMR see Table 1. ACKNOWLEDGEMENTS: Thanks are due to Mr. Agatino Renda for excellent technical assistance in the word processing of the manuscript. REFERENCES 1. Erickson, K.L. In Marine Narurul Producrs Vol. 5, Scheuer, P.J. Ed: Academic Press, New York, 1983; pp 131-257. 2. Faulkner, D.J. Nat. Prod. Rep. 1990, 7, 269. 3. Caccamese, S.; Azzolina, 4. Caccamese, S.; Amico, 5. Amico, 6. Caccamese, I. Hamilton, W.C., Acra Crystallogr. 8. Montaudo, G.; Caccamese, 9. 10. V.; Caccamese, R.; Toscano, R.M.; Rinehart, K.L. Biochem. System. Ecol. 1981, 9. 241. V.; Neri, P. J. Nur. Prod. 1990, 53, 1287. S.: Neri, P.; Russo, G.; Foti, M. Phytochemistry, S.; Azzolina, R.; Duesler, E. N.; Paul, l.c.; 1991, 30, 1921. Rinehart, K.L. Tetruhedron Lett. 1980, 21. 2299. 1965, 18, 502. S.; Librando. V.; Maravigna, P. Terrahedron 1973, 29, 3915. Montaudo, G.; Librando, V.; Caccamese, S.: Maravigna, P. /. Am. Chem. Sot. 1973, 95, 6365. Burkert, U.; Allinger, N.L. Molecular Mechanics ACS Monograph 177, American Chemical Society, Washington DC, 1982. 11. Still, W.C.; Mohamadi. F.: Richards, N.G.J.; Guida, W.C.; Lipton, M.; Liskamp, R.; Chang, G.; Hendrickson, 12. T.; DeGunst, F.; Hasel, W. MacroModel V 2.5, Dept. Chem. Columbia University, Arigoni, D. Pure Appl. Chem. 1975, 41, 219. 13. The Merck Index, 14. Takeda S.: Iimura, Y.; Tanaka, K.; Kurosawa, E.; Suzuki, T. Chem. 15. 16. Salazar, M.; Zektzer, A.S.; Martin, G.E. Magn. Res. Chem. 1988, 26, 28. Bax, A.: Ferretti, J.A.; Nashed, N.; Jerina, D.N. J. Org. Chem. 1985, 50, 3029. 17. Bax, A. J. Magn. Reson. 1984, 57, New York, 1989. tenth Ed. 1983, 687. 314. Lelt. 1990, 155.