Phytochemistry 71 (2010) 1545–1557
Contents lists available at ScienceDirect
Phytochemistry
journal homepage: www.elsevier.com/locate/phytochem
Irregular sesquiterpenoids from Ligusticum grayi roots
Laurence G. Cool a,*, Karl E. Vermillion b, Gary R. Takeoka a, Rosalind Y. Wong a a United States Department of Agriculture, Agricultural Research Service, 800 Buchanan St., Albany, CA 94710, USA b United States Department of Agriculture, Agricultural Research Service, 1815 N. University St., Peoria, IL 61604, USA article info abstract
Article history: Root oil of Ligusticum grayi (Apiaceae) contains numerous irregular sesquiterpenoids. In addition to the Received 26 March 2010 known acyclic sesquilavandulol and a sesquilavandulyl aldehyde, two thapsanes, one epithapsane, and 14 Received in revised form 26 May 2010 sesquiterpenoids representing eight hitherto unknown carbon skeletons were found. These skeletons are: Available online 6 July 2010 prethapsane, i.e. 1,1,2,3a,7,7-hexamethyloctahydro-1H-indene; isothapsane, i.e. 1,2,3a,6,7,7a-hexamethyl- octahydro-1H-indene; ligustigrane, i.e. 1,1,2,7,7,7a-hexamethyloctahydro-1H-indene; isoligustigrane, i.e. Keywords: 1,1,2,6,7,7a-hexamethyloctahydro-1H-indene; preisothapsane, i.e. 1,1,2,3a,6,7-hexamethyloctahydro-1H- Ligusticum grayi indene; isoprethapsane, i.e. 1,1,2,4,7,7-hexamethyloctahydro-1H-indene; allothapsane, i.e. 1,1,2,3a,7,7a- Apiaceae hexamethyloctahydro-1H-indene; and oshalagrane, i.e. 1,1,2,4,6,6-hexamethylspiro[4.4]nonane. Gray’s lovage Root The bicyclic sesquiterpenoids are presumably biosynthesized by head-to-head coupling of geranyl Sesquiterpenoids diphosphate and dimethylallyl diphosphate, followed by a cyclization sequence leading to a hydroindane Thapsane skeleton with six one-carbon substituents. Subsequent rearrangements—primarily methyl migrations— Epithapsane account for the remarkable variety of structures represented in L. grayi root oil. Prethapsane Ó 2010 Elsevier Ltd. All rights reserved. Isothapsane Preisothapsane Ligustigrane Isoligustigrane Allothapsane Isoprethapsane Oshalagrane
1. Introduction to the usual ‘‘head-to-tail” 10–4 coupling of GPP with DMAPP or GPP that generates FPP or GGPP. Van Klink et al. (2005) have pre- The plant family Apiaceae has been a rich source of novel natu- sented evidence supporting such a biosynthetic sequence for the ral products. Of particular interest are several terpenoids whose anisotomenes. structures are incompatible with biosynthetic mechanisms involv- ing the usual acyclic substrate molecules farnesyl diphosphate (FPP, leading to most sesquiterpenoids) and geranylgeranyl diphosphate (GGPP, leading to most diterpenoids). Thus, Thapsia villosa L. produces sesquiterpenoids based on the irregular thap- sane carbon skeleton A (Pascual Theresa et al., 1985, 1986a,b) while Anisotome lyallii Hook. f. generates anisotomane diterpenoids (skeleton B; van Klink et al., 1999; Zidorn et al., 2002). Their abso- lute stereochemistry has been proven to be as indicated (Srikrishna and Anebouselvy, 2002, 2003; van Klink et al., 2004). Biosynthesis Ligusticum grayi J.M. Coult. & Rose (oshala, Gray’s lovage or licorice- of these terpenoids apparently involves 10–2 or ‘‘head-to-head” root; Apiaceae) grows in the Sierra Nevada, Klamath, and Cascade coupling of geranyl diphosphate (GPP) with dimethylallyl Mountains of California, Nevada, Oregon and Washington, and in diphosphate (DMAPP) or a second GPP molecule. This is in contrast similar habitats in Idaho, Montana and westernmost Utah. The roots of this species and the more easterly L. porteri (‘‘osha”) were valued by Native Americans for medicinal purposes. Though L. porteri and some Asian and European Ligusticum species have been chem- * Corresponding author. Present address: 2408 McKinley Ave., Berkeley, CA 94703, USA. Tel.: +1 510 549 3972. ically analyzed, composition of L. grayi root essential oil, which E-mail address: [email protected] (L.G. Cool). has a strong, distinctive fragrance, has not been reported. Our
0031-9422/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.phytochem.2010.06.003 1546 L.G. Cool et al. / Phytochemistry 71 (2010) 1545–1557 analysis of root oil from a Sierran L. grayi population by gas The non-terpenoid hydrocarbon viridene was originally found chromatography–mass spectrometry (GC–MS) found an abundance in roots of two other members of Apiaceae, Meum athamanticum of the non-terpenoid hydrocarbon viridene, as well as numer- Jacq. and L. mutellina (L.) Crantz (Kubeczka et al., 1980, 1982). ous monoterpenes and sesquiterpenoids. Most of the latter The authors noted the striking ‘‘green” odour of viridene, and we appeared to be alcohols, but of these only the irregular acyclic likewise found it to be a major contributor to the fragrance of L. sesquilavandulol could be identified from existing GC–MS data grayi root oil. Kubeczka et al. (1980, 1982) depicted the structure collections. The composition of the root oil, and in particular, of viridene as 1-[(2E)-pent-2-en-1-yl]cyclohexa-1,3-diene, but we the identity of 18 new sesquiterpenoids, is the subject of this consider this to be questionable since no NMR spectroscopic data report. was given, nor did they discuss how they deduced the location or geometry of the double bonds. Our re-examination of the structure of viridene will be published separately. 2. Results and discussion The minor known sesquiterpenes a- and b-barbatene and iso- bazzanene, originally identified from liverworts, have also been re- The composition of the Ligusticum grayi root oil, as determined ported from another member of the Apiaceae, Meum athamanticum by GC–MS and GC-FID, is given in Table 1. Unidentified compounds (König et al., 1996), where they were found to be enantiomeric to are designated by abbreviations consisting of a species and organ the corresponding sesquiterpenes from liverworts. Sufficient b- identifier (LiGrR = L. grayi root), followed by apparent molecular barbatene was isolated from L. grayi root oil to measure its optical mass and a suffix letter (e.g. 222a). rotation (ca. +30°), proving that its absolute stereochemistry is the same as that from M. athamanticum. The acyclic irregular sesquiterpenoid sesquilavandulol, i.e. (4E)-5,9-dimethyl-2-(1-methylethenyl)deca-4,8-dien-1-ol, had 1D Table 1 NMR spectroscopic data identical to that previously reported Composition of Ligusticum grayi root oil from population 2 (38.45459°N, (Faure et al., 2000). The optical rotation of the L. grayi material 119.75941°W). is +7°, but rotation data on material of known absolute stereo- Compound Percent Identification chemistry has not been reported. The depicted chirality is pro- a-Thujene 0.1 GC–MS posed because the dextrorotatory monoterpenoid homologue a-Pinene 0.2 GC–MS (+)-lavandulol has S absolute stereochemistry (Piva, 1995). Also, Sabinene 2.4 GC–MS Thapsia villosa produces thapsanes with 2S absolute stereochemis- b-Pinene 0.4 GC–MS try (Srikrishna and Anebouselvy, 2002, 2003), and it is likely that Myrcene 0.3 GC–MS a-Phellandrene 3.7 GC–MS the new thapsanes and related structures in L. grayi described be- a-Terpinene 0.1 GC–MS low share this stereochemistry. Since thapsane biosynthesis p-Cymene 6.3 GC–MS doubtless derives from a chiral sesquilavandulyl precursor (van b-Phellandrene 1.7 GC–MS Klink et al., 2000), 2S stereochemistry for sesquilavandulol in L. Limonene 0.2 GC–MS c-Terpinene 9.2 GC–MS grayi is inferred. LiGrR132a, -134a 1.4 Compound 1, a minor component, is an achiral aldehyde Terpinolene 1.2 GC–MS with the sesquilavandulane skeleton. The 1H NMR spectrum (Ta- Viridene 13.9 GC–MS ble 2) showed an aldehydic proton (d 10.02), five vinyl methyl Methyl carvacrol 0.4 GC–MS signals between d 1.49 and 1.72, two vinyl methines (d 5.13 a-Barbatene 0.1 GC–MS Isobazzanene 0.1 GC–MS and 5.16), and three allylic methylene signals (d 2.04, 2.13 (+)-Isothapsadiene 4 4.2 NMR, polarimetry and bis-allylic 3.10, two protons each); a total of 24 protons (+)-Thapsadiene 16 0.7 NMR, polarimetry was indicated by these data. The 13C spectrum (Table 4) con- (+)-b-Barbatene 0.3 GC–MS, polarimetry sisted of 15 signals, seven of which were in the olefin region. (+)-b-Acoradiene 0.1 GC–MS, polarimetry Myristicin 0.6 GC–MS One at d 189.3 was clearly the aldehydic carbon, and two others Elemicin 0.4 GC–MS at d 124.8 and 122.6 represented vinyl methines, as confirmed Dihydroagarofuran 0.2 GC–MS by DEPT experiments. Four quaternary vinyl carbons were evi- Dill apiole 0.5 GC–MS dent (singlets between d 131 and 153). Taken together, the 1D (+)-Sesquilavandulol 4.9 GC–MS, NMR, polarimetry NMR spectroscopic data indicated a molecular formula of LiGrR222a 0.2 LiGrR222b 0.1 C15H24O, which was confirmed by HRMS (Section 4.3.1). Assign- 1 13 (+)-a-Isothapsenol 6 4.5 NMR, polarimetry ment of protons to carbons was by H– C HSQC, and the LiGrR222d 0.1 1H–13C HMBC and 1H–1H NOESY spectra completely determined LiGrR222e 0.2 the structure (key correlations in Table 5). A NOESY correlation (+)-a-Preisothapsenol 11 1.5 NMR, polarimetry from H-3 to Me-12 established the E geometry of the central ( )-a-Thapsenol 17 0.4 NMR, polarimetry ( )-a-Prethapsenol 2 15.5 NMR, polarimetry double bond. Sesquilavandulal 1 0.3 NMR Compounds 2–17 proved to be irregular hydroindane sesqui- (+)-b-Isothapsenol 5 0.5 NMR, polarimetry terpenoids that had substitution patterns reminiscent of the thaps- (+)-b-Prethapsenol 3 3.3 NMR, polarimetry anes and anisotomane diterpenoids. Their atom numbering in the LiGrR222j 0.1 (+)-Isoprethapsenol 14 0.2 NMR, polarimetry tables and text follows that given to the thapsanes by Srikrishna LiGrR222l 0.1 et al. (2000), except in Sections 4.3.2–4.3.19 where IUPAC names (+)-Allothapsenol 13 0.4 NMR, polarimetry and numbering are given. The absolute stereochemistry was not (+)-Oshalagrenol 18 2.3 NMR, polarimetry determined, but is assumed to be the same as the known thapsanes (+)-a-Isoligustigrenol 10 1.8 NMR, polarimetry (Srikrishna and Anebouselvy, 2002, 2003). Since many related (+)-b-Preisothapsenol 12 0.8 NMR, polarimetry (+)-a-Ligustigrenol 8 1.8 NMR, polarimetry compounds are reported here, elucidation of structures of only (+)-b-Isoligustigrenol 9 6.9 NMR, polarimetry the first unknowns isolated, 2 and 4, is described in much detail. ( )-b-Epithapsenol 15 1.9 NMR, polarimetry For the rest, briefer accounts are given, the NMR spectroscopic data ( )-b-Ligustigrenol 7 3.3 NMR, polarimetry (Tables 2–5) being sufficient to prove each assigned structure LiGrR222t 0.1 unambiguously. L.G. Cool et al. / Phytochemistry 71 (2010) 1545–1557 1547
Table 2 1 a H NMR spectroscopic data of Ligusticum grayi sesquiterpenoids 1–9 in C6D6, d from TMS (ppm).
Position 12 3b 3-OAc 45678 9 1 10.02 s 1.36 s 1.48 s 1.44 s ––––– – 2 – – – – 1.62 qd 1.64 qd 1.50 qd – – 1.84 qd (7.2,4.0) (7.2,4.4) (7.3,4.3) (7.4,3.9) 3 2H, 3.10 d 5.21 ddd a 1.68 ddd a 1.66 ddd 1.78 br m 1.80 m 1.34 m a 1.25 br m a 1.68 m 1.67 m (6.9) (9.9,2.7,1.0) (17.6,4.7,0.8) (17.6,4.7,1.0) b 1.48 br m b 0.82 dtd b 1.87 br dt b 1.83 br dt (13.6,3.4,1.3) (17.5,2.3) (17.6,2.2) 4 5.13 t sext 5.49 ddd 5.37 ddd 5.35 ddd 2H, 1.27 m a 1.16 m a 1.25 m 2H, 1.99 br 2H, 1.46 m a 1.74 m (6.9,1.2) (9.9,5.6,2.1) (9.9,4.7,2.5) (9.9,4.7,2.6) b 1.26 m b 1.10 m m b 2.20 m 5– a 1.84 dd 5.72 ddd 5.67 ddd 2H, 1.27 m a 1.18 m a 1.30 m 5.35 br m a 1.91 m 5.32 br dt (16.2,5.6) (9.9,2.7,1.4) (9.9,2.7,1.4) b 1.26 m b 1.52 m b 2.09 dtd (4.7,2.4) b 1.89 br d (12.1,3.7,1.2) (16.3) 6 2H, 2.04 dd – –––––––– (8.0,7.3) 7 2H, 2.13 q a 1.23 dd a 1.20 dd a 1.21 dd 2H, 2.15 brc 2H, 1.37 d a 2.18 dm a 2.14 m 5.26 br a 2.21 m (7.3) (13.4,9.5) (12.3,2.8) (12.6,2.8) (9.1) (14.9) b 2.32 dd b 2.30 m b 1.70 dd b 1.62 t (11.7) b 1.62 t (11.9) b 1.87 d (14.7,8.8) (13.2,9.9) (14.9) 8 5.16 t pent 1.55 dddd 1.79 dtd 2.02 dtd – 2.64 m – 1.58 qd 2.55 ddt 1.59 dddd (6.9,1.4) (9.9,9.7,7.8,6.0) (11.5,7.8,2.8) (11.6,7.9,2.9) (8.5,5.5) (8.2,6.4,1.9) (14.2,9.2,8.0,6.3) 9–––––––––– 10 3H, 1.52 br 3H,1.16 s 3H, 1.21 s 3H, 1.16 sE5.46 s E 4.87 d 3H, 1.37 3H, 1.09 s 3H, 1.19 s 3H, 0.86 s s Z 4.68 s (2.3) br m Z 4.79 d (2.7) 11 3H, 1.64 br 3H, 0.97 s 3H, 1.15 s 3H, 1.08 sE4.92 br t 3.45 dd 3.95 d 3H, 0.92 s 3H, 0.77 s 3H, 0.97 s m (2.0) (10.5,7.0) (12.0) Z 5.53 br td 3.55 dd 3.98 d (2.4,0.6) (10.5,5.1) (12.0) 12 3H, 1.72 d 3.25 dd 3.33 dd 4.11 dd 3H, 0.89 s 3H, 0.92 s 3H, 1.05 s 3.29 dd 3.36 dd 3.30 dd (0.8) (10.3,7.8) (10.3,7.8) (11.0,8.0) (10.1,8.7) (10.0,8.1) (10.2,8.4) 3.52 dd 3.57 dd 4.22 dd 3.52 dd 3.48 dd 3.52 dd (10.3,6.0) (10.3,7.8) (11.0,8.0) (10.2,5.6) (10.0,6.4) (10.3,6.1) 13 – 3H, 1.17 d (0.7) 3H, 1.19 s 3H, 1.15 s 3H, 0.83 d 3H, 0.85 d 3H, 0.86 d 3H, 1.04 s 3H, 1.12 s 3H, 0.98 d (7.3) (7.2) (7.0) (6.8) 14 3H, 1.49 s 3H, 1.12 s 3H, 1.03 s 3H, 0.99 s 3H, 0.85 d 3H, 0.86 d 3H, 0.79 d 3H, 0.95 s 3H, 0.93 s 3H, 0.91 d (7.2) (7.3) (7.4) (7.4) 15 3H, 1.53 s 3H, 1.15 s 3H, 1.11 s 3H, 1.06 s 3H, 0.91 s 3H, 0.83 s 3H, 0.80 s 3H, 0.91 s 3H, 0.91 s 3H, 0.93 s OH – 0.50 br s 0.55 br s – – 0.48 br s 0.75 br s 0.53 br s 0.68 br s 0.59 br s – – – 3H, 1.72 s ––––– – CH3–CO
a Solvent reference, d from TMS: 1H 7.16 ppm. Diastereotopic protons designated by a (down) and b (up) (known stereochemistry), or a and b (unknown stereochemistry). Coupling constants (Hz) are in parentheses. b Assignments by analogy with the corresponding acetate. c At 70 °C these were resolved into d 2.13 dt (J = 15.9, 2.4 Hz) and 2.17 br d (J = 15.9 Hz). 1548 L.G. Cool et al. / Phytochemistry 71 (2010) 1545–1557
Table 3 1 a H NMR spectroscopic data of Ligusticum grayi sesquiterpenoids 10–18 in C6D6, d from TMS (ppm).
Position 10 11 12 13 14 15 16 17 18b 18-OAc 1 – – 2.07 br s – 1.91 s ––––– 2 1.84 qd ––– –– –––– (7.1,5.4) 3 1.64 m 1.97 sext – 5.42 br m a 1.12 ddd a 1.52 td a 1.32 m a 1.23 m a 1.89 td a 1.86 td (7.5) (12.5,6.0,1.5) (13.8,4.9) b 1.17 m b 1.13 m (12.5,7.3) (12.6,7.4) b 1.39 td b 1.09 ddd b 1.34 ddd b 1.31 ddd (12.5,6.1) (13.8,5.0,2.0) (13.4,10.0,3.4) (13.4,10.1,3.4) 4 a 1.51 ddt a 1.78 m a 1.92 m a 2.03 m a 1.99 br m a 1.36 m a 1.32 m a 1.36 m a 1.49 m a 1.46 m (12.8,4.7,2.5) b 1.47 m b 2.07 m b 1.87 m b 1.82 br m b 1.63 tt b 1.51 m b 1.52 m b 1.61 m b 1.58 m b 1.43 tt (13.3,4.7) (13.3,4.7) 5 a 2.19 tdt a 1.33 m a 1.50 m a 1.08 m – a 1.43 td a 1.25 m a 1.54 m a 1.41 ddd a 1.34 m (14.1,4.7,2.0) b 1.52 ddd b 1.40 m b 1.85 m (12.8,4.2) b 1.41 m b 1.26 m (12.4,9.3,2.1) b 2.16 ddd b 2.11 m (11.2,4.9,2.0) b1.24 m b 2.21 ddd (12.0,10.3,8.7) (12.3,11.2,9.0) 6– – – – – – – – – – 7 5.22 br t (2.5) a 1.48 m a 1.25 dd a 1.25 t (12.5) a 1.83 br m a 1.81 br ddd a 1.79 br a1.78 br d 5.28 br m 5.21 br m b 1.73 m (13.0,5.0) b 1.36 dd (12.5,7.2) b 2.45 ddm (13.0,10.8,1.0) d (16.9) (15.6) b 1.53 m (16.3,7.6) b 1.18 dd a 2.72 br b 2.43 br (13.0,2.6) d (16.9) dm (15.6) 8 2.08 m 1.77 m 1.84 dtd 1.95 dq (12.7,7.0) 1.53 dtd 2.63 m – – 2.28 br m 2.50 br m (11.5,7.6,5.0) (11.6,7.8,5.2) 9– – – – – – – – – – 10 3H, 1.01 s 3H, 1.32 s 3H, 1.26 s 3H, 1.15 s 3H, 1.19 sZ5.02 d (2.8) Z 4.75 s 3H, 1.48 br 3H, 1.14 s 3H, 1.10 s E 4.91 d (2.4) E 5.52 s s 11 3H, 1.02 s 3H, 1.13 s 3H, 1.04 s 3H, 0.96 s 3H, 0.67 s 3.76 dd Z 5.51 br 3.99 d 3H, 0.95 s 3H, 0.91 s (10.3,5.6) m (12.1) 3.42 dd E 4.83 br 4.04 d (10.2,8.6) m (12.1) 12 3.37 dd 3.39 dd 3.36 dd 3.26 dd (10.3,6.7) 3.51 dd 3H, 0.95 s 3H, 0.79 s 3H, 0.89 s 3.48 dd 4.16 ddd (10.4,7.1) (10.3,7.5) (9.9,8.1) 3.47 dd (10.3,7.0) (10.3,5.2) (10.1,6.2) (10.8,6.2,0.9) 3.52 dd 3.66 dd 3.58 dd 3.25 dd 3.36 dd 4.08 ddd (10.4,5.9) (10.3,6.0) (9.7,7.9) (10.3,8.0) (9.9,8.3) (10.7,8.5,1.1) 13 3H, 0.94 d 3H, 1.08 s 3H, 1.00 s 3H, 0.86 s 3H, 1.58 br m 3H, 1.11 s 3H, 0.96 s 3H, 0.80 s 3H, 1.83 dd 3H, 1.80 m (7.5) (2.7,1.5) 14 3H, 0.83 d 3H, 1.06 d 3H,1.55 s 3H, 1.61 br m 3H, 1.09 s 3H, 1.17 s 3H, 0.85 s 3H, 0.88 s 3H, 0.97 s 3H, 0.93 s (7.1) (7.0) 15 3H, 0.95 s 3H, 1.66 s 3H,1.77 s 3H, 0.85 s 3H, 0.91 s 3H, 1.01 s 3H, 0.94 s 3H, 0.87 s 3H, 1.13 s 3H, 1.08 s –OH 0.80 br s 0.68 br s 0.59 br s 0.48 br s 0.44 br s 0.95 br s – 0.64 br s 0.52 br s – – – – – – – – – – 3H, 1.73 d CH3–CO (1.2)
a Solvent reference, d from TMS: 1H 7.16 ppm. Diastereotopic protons designated by a = down and b = up (known stereochemistry), or a and b (unknown stereochemistry). Coupling constants (Hz) are in parentheses. b Assignments by analogy with the corresponding acetate.
Alcohol 2, with apparent Mr = 222, was the major sesquiterpenoid to H-5a, Me-11 and Me-15; H-3 to Me-15; H-5b to Me-13; and in all L. grayi specimens examined. The 1H NMR spectrum (1D Me-11 to H-8 and Me-10. A long-range 13C–1H HETCOR experiment NMR spectroscopic data in Tables 2 and 4) pointed to a sesquiter- (correlations in Table 5) allowed unambiguous connection of the pene alcohol with six one-carbon substituents: five methyls three fragments (ruling out 2A0) to give the gross structure of 2. (three-proton singlets between d 0.97 and 1.17) and two dd signals The relative stereochemistry was determined by a 1H–1H NOESY at d 3.25 and 3.52 attributable to the methylene protons of a experiment (Table 5), which gave cross-peaks from H-1 to H-5b
-CH2OH group. There were also two vinyl protons (d 5.21 and and from H-5a to Me-13, confirming the trans ring fusion, and from 5.49). Integration of all signals gave 26 protons consistent with H-1 to H-8, establishing the a orientation of the –CH2OH group. The 13 the expected molecular formula C15H26O. The C and DEPT NMR trivial name a-prethapsenol was given to 2, ‘prethapsane’ having spectra showed 15 carbons bearing a total of 25 hydrogens: five been chosen for this carbon skeleton because methyl migration
CH3, three CH2, four CH, and three quaternary carbons; the broad from C-9 to C-1 leads to the thapsane skeleton A. 1H signal at d 0.50 was identified as a hydroxyl proton. Presence of two downfield methine groups (13C: d 123.6 and 141.8) and lack of vinyl quaternary carbons confirmed the presence of one disubsti- tuted double bond; thus the compound is bicyclic. Assignment of the protons and carbons (Table 2) was by 13C–1H HETCOR. The five methyl groups were assigned as two geminal pairs and a bridge- head methyl. This was consistent with the proton singlet at d 1.36, which was interpreted as a bridgehead methine flanked by three quaternary carbons, i.e. fragment 2A. (The other possibility, spiro structure 2A0 with a tert-butyl substituent, was considered The C-4, C-5 double-bond isomer of 2, b-prethapsenol 3, could not biosynthetically unlikely.) Analysis of the 1H–1H COSY spectrum be effectively separated by either HPLC or GC from several other indicated two coupled spin systems, giving fragments 2B sesquiterpene alcohols, so the mixture was esterified to the (H-3 ? H-4 ? H-5a,b) and 2C (H-7a,b ? H-8 ? H-12a,b ? OH), acetates, which were separable by normal-phase HPLC. A portion with weak signals due to the following long-range couplings: H-1 of the purified b-prethapsenyl acetate was hydrolyzed back to the Table 4 13 a C NMR spectroscopic data of Ligusticum grayi sesquiterpenoids 1–18 in C6D6, d from TMS (ppm).
Position 123b 3-OAc 456789101112131415161718b 18-OAc 1 189.3 64.3 62.9 62.7 51.2 51.4 54.5 50.7 57.5 47.6 52.2 150.8 61.6 52.5 62.7 55.3 52.4 55.5 67.5 67.3 2 136.7 38.5 34.5 34.4 38.5 39.6 38.5 37.1 42.0 35.1 37.5 127.2 127.3 137.7 33.8 36.5 36.4 37.3 44.2 44.1 [9] [6] 3 24.7 141.8 46.9 46.8 31.8 31.7 33.2 37.3 41.6 34.5 35.1 35.7 127.0 122.2 42.7 38.2 37.7 38.7 43.5 43.4 [5] [2] [8] 4 122.6 123.6 122.6 122.8 26.5 25.8 25.3 22.6 25.4 33.5 33.7 28.9 31.9 22.7 29.4 19.5 17.3 19.0 19.9 19.8
[6] [5] 1545–1557 (2010) 71 Phytochemistry / al. et Cool L.G. 5 135.2 45.7 138.9 138.6 34.2 36.6 38.3 116.3 26.6 115.7 22.4 38.8 40.5 32.9 122.6 31.3 33.8 35.2 34.5 34.4 [70] [28] [7] 6 40.2 39.3 44.3 44.3 41.2 42.5 41.9 146.8 148.3 145.7 151.2 41.2 41.4 42.5 133.7 43.5 42.7 44.0 148.0 148.6 7 27.1 43.1 43.1 43.2 44.8 41.3 48.1 34.4 121.4 34.7 120.7 43.2 42.2 42.8 32.5 42.7 44.6 46.5 128.0 127.2 [21] [13] [4] 8 124.8 52.1 52.1 48.3 147.5 43.6 133.4 52.8 56.4 50.5 59.5 52.4 53.4 50.7 51.4 42.4 149.6 134.8 57.3 53.7 9 131.1 44.1 43.3 43.4 157.6 161.9 143.2 47.5 48.2 45.9 46.0 45.7 42.4 44.2 44.5 157.9 155.6 139.5 48.8 48.9 [9] 10 17.7 32.5 37.2 36.8 102.8 104.4 10.0 32.5 24.3 31.5 30.3 30.6 35.3 33.6 29.6 103.6 104.8 13.0 25.1 24.8 [8] [3] 11 25.8 19.3 20.6 20.6 105.3 67.3 59.7 22.1 23.7 21.1 21.9 23.2 19.2 23.5 17.9 66.8 104.3 59.5 20.3 20.3 [11] 12 16.2 64.3 65.3 67.0 23.5 23.4 22.6 65.8 63.1 66.6 65.2 64.8 65.3 64.7 63.7 27.8 30.4 30.5 62.3 64.2 [15] [12] [3] 13 153.0 26.6 25.5 25.4 18.5 19.9 20.3 26.6 24.0 16.8 14.5 30.1 22.7 23.3 19.0 30.8 27.2 27.3 19.4 19.3 [20] [15] [4] [3] 14 22.9 34.1 32.5 32.4 10.6 9.6 9.0 27.4 25.6 15.5 15.6 22.5 18.7 21.8 31.2 25.1 29.2 31.1 28.0 28.0 [13] [9] [5] [9] 15 18.7 24.1 25.5 25.5 21.5 23.5 19.8 21.5 17.0 18.1 20.4 17.4 16.0 20.1 22.2 21.7 17.3 15.0 25.8 25.8 [41] [20] [4] [11] – – – 20.6 –––––––––––––––20.6 H3–CO C – – – 170.1 –––––––––––––––170.2 CH3–CO
a Solvent reference, d from TMS: 128.04 ppm. For broadened signals, w ½ is in square brackets [Hz]. b Assignments by 135° DEPT and analogy with the corresponding acetate. 1549 1550 L.G. Cool et al. / Phytochemistry 71 (2010) 1545–1557 alcohol for 1D NMR spectroscopic analysis, which had data (Tables The 13C spectrum of 4 at 25 °C was marked by extreme broadening 2 and 4) indicating its structural similarity to 2: five non-vinylic of a number of peaks, in one case (C-5) amounting to 70 Hz at half- methyl singlets, a -CH2OH group (two dd signals at d 3.33 and height (see Table 4). This was attributed to exchange between two 3.57), a disubstituted double bond (13C DEPT: methines at d 122.6 conformations on the time scale of the NMR experiment. Broaden- and 138.9, corresponding to 1H signals at d 5.37 and 5.72); and a ing of 13C signals due to conformational exchange has also been re- bridgehead methine (1H: singlet at d 1.48). 2D NMR experiments ported for anisotomene diterpenoids (van Klink et al., 1999, 2000). were done on the acetate, and the 1H–13C HSQC data permitted In variable temperature experiments at 45, 60 and 70 °C, the broad- 1 13 13 assignment of all H and C resonances. The HMBC data (Table 5) ened C peaks of 4 became progressively narrower, with w½ of C-5 verified the prethapsane skeleton and showed that Me-13 was cor- and -15 dropping to about 11 and 9 Hz, respectively, at 70 °C. Low related to vinyl C-5; NOESY cross-peaks (Table 5) confirmed the rel- temperature experiments in acetone-d6 at 60 and 80 °C also ative stereochemistry shown. The isomeric relationship of 2 and 3 gave increasingly sharp peaks, but in these spectra 30 signals ap- was verified by hydrogenating them. Each yielded a single product, peared, 15 for each of two conformers (ratio of major to minor con- and these had identical mass spectra, GC retention times, and opti- former ca. 3:1). Details of these experiments and analysis of cal rotations (see Section 4.3.4). conformational exchange in 4 will be presented elsewhere. The 1H NMR spectrum of hydrocarbon 4, which had MS data giv- The 1H–1H NOESY experiments on 4 at 27 and 70 °C were ing Mr = 204, showed two methyl doublets, two non-vinylic methyl unhelpful in determining the relative stereochemistry, but in ace- singlets, and two unsaturated exo-methylenes (data in Table 2). The tone-d6 at 80 °C the conformational exchange was slow enough presence of two double bonds required a bicyclic structure. The that informative NOE peaks appeared in the spectrum. Correlations 1H–1H COSY spectrum showed that the methyl doublets were on given in Table 5 are for the major conformer at 80 °C and fully adjacent carbon atoms, since they correlated with different methine agree with interatomic distances calculated for the lowest-energy protons that were in turn coupled with each other. Further correla- conformation (MM3 force-field). Cross-peaks from Me-12 and tions from this system to protons on C-4 and -5 gave fragment 4A. -15 to each other and to H-7a established the cis ring fusion, while The 1H–1H COSY spectrum also showed H-7 was only correlated to correlations from H-2 to H-10Z and from axial H-3 to H-5 were H-11 (long-range coupling), while H-11 was weakly coupled to H- only consistent with a orientation of Me-13 and -14. 10, suggesting fragment 4B. For the remaining four carbon atoms The 1H–1H COSY spectrum of b-isothapsenol 5 delineated a cou- (two methyl singlets and two non-vinylic quaternaries), there were pled spin system as depicted in 4A, another from H-7 ? H-8 ? two possible substructures, 4C and 4C0. An HMBC experiment H-12 ? OH, and coupled doublets (J = 2.2 Hz) at d 4.79 and 4.87 showed H-15 (but not H-12) to be no more than three bonds re- attributable to the exo-methylene group on C-9. H-8 (d 2.64, m) moved from C-9, which ruled out substructure 4C0. This experiment was clearly allylic. The 13C signals were significantly broadened permitted unambiguous connection of fragments 4A–C, thus delin- for the atoms corresponding to those of 4 (Table 4). The HMBC cor- eating the gross structure of 4. Its skeleton is particularly interesting relations in Table 5 also confirmed the isothapsane skeleton, and in that all six one-carbon substituents are on different ring carbon the relative stereochemistry at the five stereogenic atoms was atoms. The trivial name isothapsane was assigned to this remark- determined by comparing the NOESY spectrum with the MM3 able sesquiterpene skeleton, 4 being called isothapsadiene. minimum-energy conformations of the candidate diastereomers. A cross-peak from Me-15 to H-11 proved the cis relationship of these substituents, and correlations from H-2 and -3 to H-10Z were only consistent with cis ring fusion and the depicted stereochemis- try at C-2 and C-3. The 1D NMR spectroscopic data (Tables 2 and 4)ofa-isothapse- nol 6 suggested its structure, which was confirmed by 2D 1H–1H COSY, 13C–1H HETCOR and long-range HETCOR and HMBC experi- ments. As in 4 and 5, the three-proton doublets at d 0.79 and 0.86 were part of the coupled chain 4A, while H-7 and H-11 only
Table 5 Selected long-range NMR correlations of Ligusticum grayi sesquiterpenoids 1–18.a
Compound 1H–13C HMBC or 13C–1H HETCOR (H Nos. – C Nos.)b 1H–1H NOESY (H No. – H Nos.)c 1 1-2,3,13; 10,11-8,9; 12-4,5,6;4-2,3,6,12; 14,15-2,13 3-12; 1-15; 7-10 2 10,11-1,8,9; 13-1,5,6,7; 14,15-1,2,3;3-1,2,5,14 1-5b,7b,8; 5a-13; 7a-12 3-OAc 10,11-1,8,9; 13-1,5,6,7/9; 14,15-1,2,3 12-11,13; 1-3b,8,10; 14-3b; 5-7b 4 10-1,8,9; 11-7,8,9; 12-1,5,6,7; 13-2,3,4; 14-1,2,3; 15-1,2,6,9 2-10Z; 3-5; 7a-12,15; 7b-11E; 12-15 5 10-1,8,9; 11-7,8,9; 12-1,5,6,7; 13-2,3,4; 14-1,2,3; 15-1,2,6,9 10Z-2,3; 15-11 6 10-1,8,9; 11-7,8,9; 12-1,5,6,7; 13-2,3,4; 14-1,2,3; 15-1,2,6,9 7a-12,15; 12-14 7 10,11-1,8,9; 12-7,8,9; 13,14-1,2,3; 15-1,2,6,9 8-10; 12-11,15 8 7-1,5,6,8,9,12; 10,11-1,8,9; 12-7,8,9; 13,14-1,2,3; 15-1,2,6,9 8-10,13; 11-12; 13-3b; 14,15-3a 9 10,11-1,8,9; 13-2,3,4; 14-1,2,3; 15-1,2,6,9 10-2,8; 12-7a,11,15; 2-4b; 13-4a 10 7-1,5,6,8,9; 10,11-1,8,9; 12-7,8,9; 13-2,3,4; 14-1,2,3; 15-1,2,6,9 10-2,8; 12-11,15; 2-4b; 13-4a 11 10,11-1,8,9; 12-7,8,9; 13-1,5,6,7; 14-2,3,4; 15-1,2,3 8-10; 12-11,13 12 10,11-1,8,9;7a-1,8,12,13; 5a-1,3,4,6?,13; 14-4; 15-1 1-10; 12-11,13 13 10,11-1,8,9; 13-1,5,6/7; 14-1,2,3; 15-1,2,6/7,9 7a-11,12,15; 10-5b,8,14 14 10,11-1,8,9; 12-7,8,9; 13-4,5,6 1-8 15 4-2,3,5,6; 10-1,8,9; 11-7,8,9; 12-1,5,6,7; 13,14-1,2,3; 15-1,2,6,9 15-7b,8; 11-7a,12 16 4-2,3,5,6; 10-1,8,9; 11-7,8,9; 12-1,5,6,7; 13,14-1,2,3; 15-1,2,6,9 3a-14,15; 12-7a; 13-5b,7b;5b-7b 17 10-1,8,9; 11-7,8,9; 12-1,5,6,7; 13,14-1,2,3; 15-1,2,6,9 12-5a,7a; 13-3b,4b,7b 18-OAc 10,11-1,8,9; 12-7,8,9; 13-1,6,7; 14,15-1,2,3 5b-10,14; 5a-8; 13-15
a Compounds 3 and 18: 2D data on acetates only. b Forward slash ‘/’ means ‘‘and/or”. For all geminal methyl pairs, CH3–C–CH3 correlations were seen but are not tabulated. c For compound 4, NOESY correlations are for the major conformer in acetone-d6 at 80 °C. L.G. Cool et al. / Phytochemistry 71 (2010) 1545–1557 1551 showed long-range coupling to Me-10. Again, a number of 13C sig- C-7) and verified the locations of all methyl substituents. The nals were broadened due to conformational exchange. The tetra- NOESY spectrum (Table 5) showed a cis relationship between substituted double bond (13C singlets at d 133.4 and 143.2) bore a-oriented Me-11, -15, and H-12, and between b-oriented Me-10, a methyl (3H, d 1.37, br m) and a -CH2OH group (doublets at d H-8 and H-2. (H-2 also correlated with H-4b, indicating that both 3.95 and 3.98), thus locating the double bond in the cyclopentene are axial.) The stereochemistry at C-3 was established by presence ring. The 1H–1H NOESY spectrum (Table 5) gave the relative stereo- of cross-peaks from Me-13 to H-4a and H-5, which requires a (i.e., chemistry at C-1, -2 and -6. No diagnostic signal from Me-10 to H-3 axial) orientation for this methyl group. In the case of the 3S diaste- confirming the expected 3R stereochemistry could be discerned reomer, the MM3 minimum-energy conformation has Me-13 in an due to the similarity of chemical shifts of these protons. Neverthe- equatorial orientation, which is inconsistent with the observed less, examination of the MM3 minimum-energy conformation of spectrum. the 1S,2R,3S,6S diastereomer showed that (a) the observed NOESY The 1H NMR spectrum of 11 (Table 3) had one methyl doublet interaction from H-7a to Me-15 would not occur, and (b) a clear and four methyl singlets, one of them vinylic (broad peak at d
NOESY cross-peak from H-3 to Me-15 would be expected that 1.66). The -CH2OH group typical of this series of compounds was was in fact absent in the spectrum of compound 6. also evident (two dd, d 3.39 and 3.66; -OH br s at d 0.68), and there Alcohol 7, b-ligustigrenol, is the sesquiterpenoid analogue of were no vinyl protons. The 13C and DEPT spectra confirmed there anisotomenol (7 with R = isopentenyl). The 1H NMR spectrum (Ta- was one tetrasubstituted double bond (two s, d 127.2 and 150.8). ble 2) showed five methyl singlets (none of them vinylic) and a The 1H–1H COSY spectrum indicated coupled chains H-3 ?
-CH2OH group. Presence of a single vinyl proton (d 5.35) and a H-4 ? H-5 and H-7 ? H-8 ? H-12 ? -OH, with H-3 being appar- vinylic quaternary carbon (13C: d 146.8) indicated a trisubstituted ently allylic (d 1.97). The carbon skeleton suggested by these data bridgehead double bond. The 1H–1H COSY spectrum indicated H- was proven by HMBC correlations in Table 5. The NOESY correla-
3 ? H-4 ? H-5 and H-7 ? H-8 ? H-12 coupled systems, while tions established a cis relationship between the -CH2OH group the 13C–1H HETCOR and long-range HETCOR experiments (Table 5) and Me-13, but no cross-peaks confirming the expected a orienta- confirmed the presence of two geminal methyl pairs and a bridge- tion of Me-14 were discernible. However, the minimum-energy head methyl arranged as shown and clarified all the features of the conformation of the C-3 epimer would place Me-14 in a pseudo- gross carbon skeleton, which is given the trivial name ligustigrane. axial orientation leading to a cross-peak with H-5b; however, no As in anisotomenol (van Klink et al., 2000), conformational ex- such peak appeared in the spectrum. The skeleton of 11 would change was evident in the 13C spectrum, where atoms correspond- be expected to arise during biosynthesis of the isothapsane com- ing to those in the diterpenoid were similarly broadened (Table 4). pounds 4–6, which suggested the trivial name ‘‘preisothapsane”, MM3 minimum-energy calculations yielded two conformers of 7 11 being called a-preisothapsenol. whose steric energy differed by 1.5 kcal mol 1. The NOESY spec- Two other preisothapsenols might be anticipated, the D2 and trum had strong exchange peaks for H-3a and b, which switch be- D2(15) isomers, but only the former, b-preisothapsenol 12, was tween axial and equatorial orientations in the two conformers. Any found. Its 1H NMR spectrum had five methyl singlets, two of them exchange peaks for the methyls with broadened 13C signals (Me- vinylic (broadened signals at d 1.55 and 1.77). A tetrasubstituted 11, -13 and -15) were not visible due to proximity to the diagonal. double bond was indicated by 13C singlets at d 127.0 and 127.3; NOE cross-peaks in the spectrum (Table 5) proved the cis relation- the possibility that the vinyl methyls might be geminal (i.e. as a ship between H-12, Me-15, H-7a and Me-11. A cross-peak be- propylidene substituent) rather than vicinal was considered bio- tween b-oriented Me-10 and H-8 was also observed. synthetically unlikely, and the HMBC spectrum established the The 1H NMR spectra of a-ligustigrenol 8, the C-6, C-7 double- attachments of all quaternary carbons to protonated ones (Table 5). bond isomer of 7, differed from the latter in that H-8 was now The NOESY correlations showed H-11, -12 and -13 are a-oriented, apparently allylic (d 2.55 vs. 1.58 in 7); the 1H–1H COSY spectrum while H-1 and -10 are b-oriented. confirmed that it was coupled to vinyl H-7 as well as H-12a,b. The Yet another substitution pattern is found in allothapsenol 13. HSQC and HMBC correlations led to complete assignment of all sig- The 1D data (Table 3) showed a trisubstituted double bond (vinyl nals, while the NOESY data (Table 5) showed H-8, -10, -13 and -3b methine at d 5.42; vinyl methyl at d 1.61) and four high-field were on the opposite faces of the molecule from H-11, -12, -14, -15 methyl singlets. The location of the double bond at C-2, C-3 was and -3a, confirming the cis relationship between Me-15 and H-12. shown by HMBC correlations from H-14 to C-1, -2 and -3, while Another major compound was b-isoligustigrenol 9. The 1H NMR the C-1 methyl substituent protons H-15 showed HMBC correla- spectrum showed two methyl doublets, three high-field methyl tions to C-1, -2, -6 and -9. The relative stereochemistry at C-1, -6 1 1 singlets, a -CH2OH group, and a vinyl proton. The H– H COSY data and -8 was evident from the NOESY correlations (Table 5). showed that the methyl doublets represent vicinal substituents as An unexpected finding was isoprethapsenol 14. Its 1D 1H spec- in the coupled spin system 4A, but in this case H-5 is vinylic (d trum showed similarities to the other hydroindane alcohols in L. 5.32). Another chain H-7 ? H-8 ? H-12 ? -OH was observed, with grayi: five methyl singlets (one being vinylic), and the usual allylic H-7a also showing long-range coupling to H-5. The relative -CH2OH group. There were no vinyl protons, and a broad singlet stereochemistry at C-2, C-8 and C-15 was evident from the NOESY (1H, d 1.91) was interpreted as an allylic bridgehead methine. data (Table 5), which showed cross-peaks from b-oriented Me-10 Two allylic methylenes (at C-4 and C-7 by HSQC) were also indi- to H-2 and H-8, and from a-oriented Me-11 and -15 to H-12. cated. The 13C spectroscopic data confirmed one tetrasubstituted Me-15 also correlated with proton H-7a, which in turn was corre- double bond, so the compound is bicyclic. Analysis of the 1H–1H lated with H-12a,b. The relative stereochemistry at C-3 was de- COSY spectrum delineated the coupled systems H-3 ? H-4 and duced by the combination of a cross-peak from H-2 to H-4 (d H-7 ? H-8 ? H-12. The HMBC spectrum showed that the four 2.21) that established the b orientation of the latter, and the corre- high-field methyls occur as geminal pairs on C-2 and C-9 (all lation between Me-13 and H-4a (but not H-4b), which is only con- showed cross-peaks to C-1), while vinyl Me-13 is attached at C-5 sistent with an a orientation of Me-13. (correlations with C-4, -5 and -6). These NMR spectroscopic data a-Isoligustigrenol 10 is the C-6, C-7 double-bond isomer of 9. unambiguously defined the gross carbon skeleton, and a NOESY The 1H–1H COSY spectrum delineated the coupled spin system cross-peak established the relative stereochemistry at C-1 and -8. 4A, H-5a and -5b now being clearly allylic (d 2.19 and 2.11, respec- The structure of 14 suggests it is the rearrangement product of a tively). The HMBC data (Table 5) confirmed the bridgehead double prethapsane precursor in which the angular methyl at C-6 mi- bond is in the cyclopentene ring (cross-peaks from H-12a,b to vinyl grates to C-5 with deprotonation at the latter position. 1552 L.G. Cool et al. / Phytochemistry 71 (2010) 1545–1557
The 1D and 2D NMR spectra of alcohol 15, b-epithapsenol, indi- skeleton, compound 18 being called oshalagrenol. No correspond- cated a thapsane skeleton: an unsaturated exo-methylene group at ing D6(13) exo-methylene isomer was found in the oil. C-9 (HMBC correlations from H-10 to C-1, -8 and -9), geminal Production of complex mixtures of sesquiterpenoids by a single methyls at C-2 (HMBC cross-peaks from H-13 and -14 to C-1, -2 enzyme catalyst has been demonstrated in a number of plants (re- and -3), angular methyls at C-1 and C-6, and a -CH2OH group at viewed by Degenhardt et al., 2009), so the biosynthesis of sesqui- C-8. Although the gross structure is the same as the known terpenoids 2–18 in L. grayi may be catalyzed by a small number 9(10)-thapsen-11-ol (15 with Me-15 a-oriented; ‘‘b-thapsenol” of synthases. Because plant-to-plant variation in the relative 1 13 hereinafter), its H and C NMR spectroscopic data (in CDCl3; amounts of terpenoids can indicate possible biosynthetic relation- see Section 4.3.16) and optical rotation ( 3.4°) are different (Smitt ships (Jones et al., 2006; Zavarin, 1970), GC-FID analyses of a num- et al., 1990). Analysis of the NOESY data (Table 5) demonstrated ber of L. grayi specimens from three populations were undertaken. that the methyl at C-1 of 15 was b-oriented, i.e. the ring fusion is Variability within a population was generally low, but the data trans. pointed to two individuals with markedly atypical proportions of The 1D NMR spectroscopic data (Tables 3 and 4) of the minor particular sesquiterpenoids. One plant from population 2 had neg- hydrocarbon 16 (GC–MS: Mr = 204) suggested that, like 4, there ligible amounts of b-epithapsanol 15, thapsadiene 16 and oshala- are two vicinal unsaturated exo-methylenes (1H–1H COSY: long- grenol 18 (the other sesquiterpenoids being at typical levels), range couplings from H-10 to H-11), but in 16 all four methyl which suggested the possibility of a separate pathway for these groups appear as singlets. The HMBC spectrum proved the geminal compounds. Against the implied tight linkage in the biosynthesis attachment of Me-13 and -14 at C-2; location of Me-12 at C-6 and of 15, 16 and 18 is the fact that all plants analyzed from popula- Me-15 at C-1; the two- or three-bond proximity of exo-methylene tions 1 and 3 had much less 15 than those in population 2 (average protons H-10E,Z to C-1, -8 and -9; and proximity of H-11E,Z to C-7, of 0.4% vs. 1.9%, respectively) but had comparable amounts of 16 -8 and -9. Given the presence of substantial amounts of b-epi- and 18. Another anomalous plant (from population 3) had much thapsenol 15 in L. grayi, it was anticipated that this hydrocarbon reduced levels of hydrocarbons 4 and 16 (0.2% and 0.05%, respec- might be the dehydration product of 15. However, this proved tively, vs. 3.8% and 0.9% average for the other plants in this popu- not to be the case. The NOESY spectrum (Table 5) had cross-peaks lation). The decrease in amount of 4 was accompanied by a that were only compatible with cis ring fusion, i.e. 16 is commensurate increase in the amount of a-isothapsenol 6 (5.2% thapsadiene. vs. 1.5% average for the other plants). This suggests that dehydra- a-Thapsenol 17, a minor component, was identified as the tion of 6 to 4, and presumably of 17 to 16, is catalyzed by a separate endocyclic double-bond isomer of known b-thapsenol. The 1H enzyme (a comparison of the amounts of 16 and 17 in individual NMR spectroscopic data had signals close to those of substituents plants could not be made due to GC co-elution of 11 and 17). on the cyclopentene ring of 6, i.e. an allylic –CH2OH group (dou- Though the number and activities of enzymes responsible for blets at d 3.99 and 4.04) and a vinyl methyl (d 1.48) on a tetrasub- biosynthesis of 2–18 are not known, plausible routes leading to stituted double bond (13C: singlets at d 134.8 and 139.5). The the L. grayi sesquiterpenoids can be proposed. The scheme shown 1H–1H COSY spectrum showed long-range correlations from vinyl in Fig. 1 includes the putative cyclization sequence (shaded por- Me-10 to H-7a and -7b, and (by W-coupling) from Me-12 to tion) proposed by van Klink et al. (2000) to account for biosynthe- H-7b and from Me-13 to H-3a. In contrast to 6, all four high-field sis of the thapsane skeleton. For simplicity, it is arbitrarily assumed methyl resonances of 17 were singlets, and their expected loca- that the diphosphate group of cosubstrate DMAPP has been hydro- tions at C-1, C-2 (geminal pair) and C-6 were confirmed by analysis lyzed before condensation with GPP rather than at a later step.1 of the HMBC spectrum (Table 5). Some potentially diagnostic cor- Deprotonation of the prethapsanol-3-carbocation C1 (path 1a) relations in the NOESY spectrum could not be observed due to sim- gives the dominant L. grayi sesquiterpenoid 2.2 The alternate path- ilarity of the chemical shifts for Me-12, -14 and -15, but, based on ways in Fig. 1 lead to bicyclic cations C2–C13 that are the likely the MM3 minimum-energy conformation, the observed correlation intermediates in the formation of 3–18. Cations C2, C6 and C12 arise from Me-13 to H-7b was incompatible with a trans ring fusion. from C1 by 1,2 methyl or 1,3 hydride shifts (paths 1d, 1c and 1b, The acetate of 18 was purified from an acetylated mixture that respectively). Since cation C6 lacks a-protons, a 1,2 methyl shift included several other L. grayi alcohols, and the 2D NMR spectro- (paths 6a–d) or Wagner–Meerwein rearrangement (6e, leading to scopic data were acquired on this derivative. The chemical shifts C11) must occur before deprotonation to neutral products is possi- given in this discussion are those of the regenerated alcohol (Ta- ble. It is striking that compounds resulting from four of the five pos- bles 3 and 4). Compound 18 shared some features with 13: bicyclic sible suprafacial methyl migrations are produced from C6 in sesquiterpenoid; one vinyl methine (1H d 5.28; 13C d 128.0); six significant amounts. In the case of C2, H-1, -3 and -15 are available one-carbon substituents that included a vinyl methyl (d 1.83), four for termination of the reaction sequence by deprotonation, and pre- high-field methyl singlets (d 0.95–1.14), and a -CH2OH group (two isothapsenols corresponding to the first two of these (11 and 12) dd at d 3.36 and 3.48; –OH br s at d 0.52); and (by 13C DEPT) one were in fact found. However, further rearrangement of C2 by path vinylic and three non-vinylic quaternary carbons. 1H–1H COSY elu- 2c (1,2 hydride shift, giving C3) also occurs, with subsequent methyl cidated the two coupled systems: H-3a,b ? H-4a,b ? H-5a,b shifts (path 3b or 3c) leading to the isothapsane (C4) and isoligustig- (non-allylic methylenes); and (vinyl) H-7 ? (allylic) H-8 ? rane (C5) skeletons. Deprotonation of C3 (path 3a) may contribute H-12a,b. Taken together, these data could not be reconciled with additional a-preisothapsenol 11 to any produced by path 2b. a hydroindane structure, and an HMBC experiment showed that all five methyl substituents, including vinyl Me-13, were attached 1 to ring atoms that were one bond removed from quaternary C-1 (d The presence of hydrocarbons 4 and 16 in L. grayi might actually argue against 67.5). This is only possible with the spiro[4.4]nonane structure this assumption. If the diphosphate moiety of DMAPP were to remain intact throughout the reaction cascades, it would make possible a more favourable process shown. The absolute stereochemistry at C-8 is assumed to be the for generation of 4 and 16 than the dehydration of primary alcohols 6 and 17 shown same as the other sesquiterpene alcohols in L. grayi, while the rel- in Fig. 1. This alternative would involve ionization of putative allylic diphosphates 6- ative stereochemistry of the spiro atom was determined by com- PP and 17-PP with subsequent deprotonation of the allylic cations to give the dienes. paring the NOESY data to the minimum-energy conformations of For the remaining sesquiterpenoids, the last biosynthetic step would be enzymatic hydrolysis of the diphosphate group to give the alcohols. the two candidate diastereomers. The definitive cross-peak was 2 In this discussion, atom numbering of a carbocation corresponds to that of its from H-8 to H-5a, which is incompatible with S stereochemistry product structure(s), as depicted for 2–18. Atom numbers are omitted from Fig. 1 for at the spiro atom. We chose the name oshalagrane for this carbon clarity. L.G. Cool et al. / Phytochemistry 71 (2010) 1545–1557 1553
Fig. 1. Proposed biosynthesis of Ligusticum grayi sesquiterpenoids 1–18. Steps in the shaded area are according to van Klink et al. (2000).
It will be noted that the relative stereochemistry of most prod- would lend support to this possibility, but no such compound ucts deriving from C1 can be explained by reaction cascades was isolated. However, a trace unknown (LiGrR222e, Table 1) involving suprafacial hydride and methyl shifts, with termini at had a distinctive mass spectrum essentially identical to that of 2. C-1 and/or C-2, where the migrating substituents have an anti- In all likelihood, this compound is indeed the cis-fused stereoiso- periplanar geometry. (This may be true at C-5 as well, i.e. path mer of 2. This scenario, i.e. biosynthesis of 15 by a separate enzyme 1b+12b, though nothing about the stereochemistry of the hydride via 1-epi-C1, could account for the fact that its proportion was dra- and methyl shifts can be deduced from the structure of 14.) Thus matically lower in populations 1 and 3 than in population 2. path 1d+2c accounts for the a orientation of the methyl groups at C-2 and -3, and, by subsequent steps 3b or 3c, inversion of con- figuration at C-1. Likewise, paths 1c+6a,b,c lead to inversion of C-1 3. Conclusion configuration in C7, C8 and C9. An apparent exception to this pattern is trans-fused 15 (path 1c+6d), biogenesis of which would The roots of Sierran L. grayi plants produce large amounts of require a syn-periplanar relationship in the C-1 ? C-3 hydride and irregular sesquiterpenoids. In addition to acyclic (+)-sesquilav- C-9 ? C-1 methyl shifts, as depicted. This is quite possible: both andulol and sesquilavandulal, there are 17 bicyclic sesquiterpene anti- and syn-periplanar tandem hydride and methyl migrations alcohols representing the following carbon skeletons: thapsane A, have been encountered in sesquiterpene biosynthesis, as, for epithapsane C, prethapsane D, isothapsane E, ligustigrane F, isolig- example, in aristolochene and 5-epi-aristolochene biosynthesis ustigrane G, preisothapsane H, allothapsane I, isoprethapsane J, (Cane, 1990). Another possibility is that 15 derives, not from C1, and oshalagrane K. These skeletons are presumed to arise from a but rather from a cis-fused prethapsanol-3-carbocation (i.e., 1- prethapsane carbocation (skeleton D) by either methyl migrations epi-C1, not shown) in which sequential anti-periplanar 1,3 hydride (A, C and E–J) or Wagner–Meerwein rearrangement (K). The struc- and 1,2 methyl shifts give the appropriate stereochemistry at C-1. tural relationships among the products suggest that relatively few Occurrence of a cis-fused prethapsenol (e.g. 1-epi-2) in the oil synthases may be involved in their biogenesis. 1554 L.G. Cool et al. / Phytochemistry 71 (2010) 1545–1557
Preparative GC was on a 6 mm i.d. 2.5 m aluminum column packed with 10% SE-30 stationary phase, using He carrier gas at 50 ml min 1, with thermal conductivity detection. Injector and detector temperatures were 220 °C. Exact masses were determined by high resolution positive-ion ESI-QTOF MS on an ABI Qstar Pulsar mass spectrometer, with inter- nal mass standards caffeine and Substance P loaded with the sam- ple in NanoES spray capillaries (Proxeon Biosystems). Melting points were determined by differential scanning calorimetry (DSC 2910, TA Instruments) as the onset temperature of the endo- therm curve, using a heating rate of 10 °C min 1. NMR spectra of 2, 6 and 7, except 1H–1H NOESY, were acquired at 25 °C on a Bruker ARX-400 spectrometer using a 5 mm BB XYZ probe. All other 1D and 2D NMR spectra were acquired on a Bruker Avance 500 spectrometer at 27 °C using a 5 mm BBO probe and TOPSPIN 1.3 pl 8 software. Samples for analysis on the latter sys- tem were shipped neat in sealed glass vials after addition of ca. 0.05 mg of Ionox 330 antioxidant. Upon receipt, they were recon-
stituted in 0.6 ml of benzene-d6 and refrigerated until needed. Variable temperature experiments on 4 were run in benzene-d6 (45, 60 and 70 °C) or acetone-d6 ( 80 °C). All IUPAC names were generated using the ACD/Name module of ACD/ChemSketch v. 12.0 software, Advanced Chemistry Devel- 4. Experimental opment, Inc., Toronto, Canada. Global minimum-energy conforma- tions were identified with the Monte Carlo-Minimization routine 4.1. General experimental procedures of the TINKER molecular mechanics package (Ponder, 2009), using MM3 force-field parameters. The NEWTON minimization routine Survey analyses of individual L. grayi specimens were done by (Ponder and Richards, 1987) was used with MM3 parameters for thinly slicing, freezing (liq. N2) and grinding ca. 1 g of fresh roots determination of energy differences between pairs of low-energy from plants from populations 2 and 3 and thawed frozen roots of conformations. plants from population 1 (see Table 6 for collection locations), extracting overnight with n-pentane with addition of Na2SO4 (ca. 4.2. Plant material 0.5 g), and analyzing the extracts by GC-FID (polydimethylsiloxane WCOT column) under these conditions: injector 220 °C, FID detec- The locations of plant populations sampled are given in Table 6. tor 250 °C, column temperature 35 °C (0.7 min), 6 °C min 1 pro- Plants from population 2 were used for compound isolation, and a gram to 240 °C. The same GC conditions were used for GC–MS voucher from this collection has been deposited in the University and GC-FID analyses of the essential oil, the LC fractions and the of California Herbarium (Accession No. UC 1948336). Small purified compounds. Known compounds were identified by GC– amounts of five of these samples were extracted and analyzed as MS comparison with published data (Adams, 2007). described in Section 4.1, after which the entire collection was Four normal-phase LC columns were employed for oil separa- ground and extracted for production of the oil used for compound tions. Low-pressure column I consisted of nominal 60 l silica gel isolation (see Section 4.3). Survey analyses of individual plants (50 g) in a 25 600 mm i.d. glass column with a PTFE bed support. from populations 1 (8 plants) and 3 (21 plants) were done as de- A head pressure of ca. 110 kPa of N2 was used to maintain an ade- scribed in Section 4.1. quate flow rate (ca. 5 ml min 1), with automatic collection of elu- ate fractions at 3 min intervals. Preparative HPLC column II was 21.4 250 mm (stainless steel) packed with 8l silica gel packing. 4.3. Compound isolation and identification Solvent flow rate was 16.0 ml min 1. Semi-preparative HPLC col- umn III was 10 250 mm (stainless steel) packed with 5l spheri- The washed roots from population 2 (360 g total) were sliced, cal silica gel packing. Semi-preparative HPLC column IV was frozen in liq. N2, ground (mortar and pestle), covered with hexane 10 250 mm (stainless steel) with 5l diol bonded-phase silica with addition of Na2SO4 (200 g), and left overnight. The solution packing. Semi-preparative HPLC column V was 10 250 mm was removed by suction (fritted glass funnel) and the filter cake (stainless steel) packed with 5l silica gel impregnated with 15% displacement-washed with a small amount of hexane. Overnight w/w AgNO3. For all HPLC analyses, peak detection was by refractive extraction of the cake was then repeated with fresh hexane. index, with manual switching of fraction collection vessels based The combined extracts were dried (Na2SO4, 10 g), and the solvent on detector output. removed by rotary evaporation. The oily extract (21.7 g) was Organic solvents were Fisher Scientific’s Optima grade, with hydrodistilled for two days from 150 ml satd. NaCl and 3 g NaH-
‘‘hexane” or ‘‘n-hexane” in this section referring to Fisher’s ca. CO3. The condensate receiver contained a layer of hexane on 75:25 mixture of unbranched and branched hexanes. H2O, the condensed H2O phase being continuously recycled to
Table 6 Sampled locations of Ligusticum grayi in California.
Population No. Latitude Longitude Elevation (m) Date Number of plants 1 39.67347°N 120.62949°W 2035 July 15, 2007 8 2 38.45459°N 119.75941°W 2456 July 21, 2007 20 3 40.47160°N 121.41523°W 1927 September 15, 2009 21 L.G. Cool et al. / Phytochemistry 71 (2010) 1545–1557 1555 the distillation flask. After distillation was complete, the organic 4.3.1. Sesquilavandulal, (4E)-5,9-dimethyl-2-(1- layer was dried (Na2SO4) and the hexane removed by rotary evap- methylethylidene)deca-4,8-dienal (1) oration. Yield of oil was 13.4 g (3.7%, fr. wt basis). Oil; for 1H and 13C NMR spectroscopic data, see Tables 2 and 4. + The root oil was separated in 2.7 g portions by LC on column I HRESIMS: [M+H] m/z found 221.1776, calcd. for C15H25O using step gradients of hexane and EtOAc-hexane mixtures (ratios 221.1906; GC–MS 70 eV, m/z (rel. int.): 220 [M]+ (8), 205 (13), are v/v) as follows: hexane (25 ml), 2.5:97.5 (25 ml), 5:95 (75 ml), 202 (10), 187 (16), 177 (21), 162 (6), 151 (13), 138 (100), 137 7.5:92.5 (100 ml), 1:9 (100 ml), 1:4 (75 ml), 1:9 (100 ml), hexane (33), 133 (29), 123 (66), 122 (24), 121 (21), 109 (50), 107 (25), (100 ml). After GC analysis of each collection tube, similar tubes 105 (29), 95 (24), 93 (26), 91 (28), 81 (59), 69 (65), 67 (39), 55 were combined to give five fractions: 1, 2, 3, 4a, 4b. (26), 53 (20), 43 (31), 41 (94). LC fraction 1 consisted solely of hydrocarbons. Part of this frac- tion was separated by repeated prep. GC at 180 °C to give pure vir- 4.3.2. ( )-a-Prethapsenol, [(2S,3aS,7aS)-1,1,3a,7,7-pentamethyl- idene and sub-fractions 1–1 (95:5 4 and 16) and 1–2 (mixture of 4 2,3,3a,4,7,7a-hexahydro-1H-inden-2-yl]methanol (2) 21 and 16 with b-barbatene and b-acoradiene). HPLC of fraction 1–1 Colorless crystals, mp (uncorr.) 49.6 °C; ½a D 47 (c 2.1; n-hex- on column V using toluene–hexane (15:85, v/v) gave pure 4; HPLC ane); for 1H and 13C NMR spectroscopic data, see Tables 2 and 4; + of fraction 1–2 under the same conditions gave 97% pure 16 and HRESIMS: [M+H] m/z found 223.2067, calcd. for C15H27O analytical samples of b-barbatene and b-acoradiene. 223.2062; GC–MS 70 eV, m/z (rel. int.): 222 [M]+ (3), 207 (13), LC fraction 2 consisted of small amounts of several unknowns 189 (8), 168 (12), 167 (100), 149 (21), 139 (28), 135 (19), 123 with Mr = 220, including 1, along with non-terpenoids myristicin, (15), 122 (16), 121 (68), 119 (18), 107 (51), 93 (39), 91 (28), 79 dill apiole, elemicin and dihydroagarofuran. It was subjected to (17), 69 (17), 67 (16), 55 (18), 41 (26). HPLC on column III using EtOAc–hexane (4.5:95.5, v/v), giving 94% pure 1. Fraction 3 consisted of >90% sesquilavandulol. It was 4.3.3. (+)-b-prethapsenol, [(2S,3aS,7aR)-3,3,4,4,7a-pentamethyl- purified to 97% by HPLC on column IV using EtOAc–hexane 2,3,3a,4,5,7a-hexahydro-1H-inden-2-yl]methanol (3) 21 (8:92, v/v). The measured specific optical rotation was +7 (c 3.9, Colorless crystals, mp (uncorr.) 99.3 °C. ½a D +61 (c 2.1; n-hex- MeOH). ane); for 1H and 13C NMR spectroscopic data, see Tables 2 and 4; + LC fraction 4a was separated on HPLC column III using EtOAc– HRESIMS: [M+H] m/z found 223.2022, calcd. for C15H27O hexane (8:92, v/v) giving the following sub-fractions rich in certain 223.2062; GC–MS 70 eV, m/z (rel. int.): 222 [M]+ (2), 207 (48), compounds (parentheses): 4a-1 (sesquilavandulol); 4a-2 (5, 8, 10); 189 (18), 181 (25), 167 (56), 166 (28), 161 (12), 151 (12), 149 4a-3 (3, 11, 12, 13, 18); 4a-4 (6, 17); 4a-5 (96% pure 6). Fraction 4a- (17), 136 (34), 135 (50), 133 (22), 123 (33), 121 (98), 107 (81), 2 was separated on diol HPLC column IV with EtOAc–hexane (5:95, 105 (39), 97 (100), 93 (76), 91 (49), 83 (15), 81 (26), 79 (33), 77 v/v), giving analytical samples of 5, 8 and 10. Fraction 4a-3 was (27), 69 (34), 55 (40), 41 (50). acetylated with Ac2O and a catalytic amt. of DMAP in pyridine at 85 °C for 45 min. After quenching in ice H2O, extraction with 4.3.4. Hydrogenation of 2 and 3 to (+)-prethapsanol, [(2S,3aS,7aR)- hexane (3 2 ml), drying with Na2SO4 and solvent removal, the 1,1,3a,7,7-pentamethyloctahydro-1H-inden-2-yl]methanol resulting mixture was separated on HPLC column III using Starting material (15 mg) was dissolved in MeOH (0.5 ml), a
EtOAc–hexane (1:99, v/v) giving samples of 3-OAc, 11-OAc, 12- small amount of 5% PdO2 on activated carbon added, and H2 bub- OAc, 13-OAc and 18-OAc. Each of these was subsequently hydro- bled in at 1 atm for 2 h under exclusion of air. 2 and 3 each gave a lyzed in 0.5 ml 1 N KOH in MeOH with ca.20mgH2Oat single product; these were identical by GC–MS. Colorless crystals, 21 90–100 °C for 1 h. After diluting in cold H2O, extraction with hex- mp (uncorr.) 72.8 °C; ½a D +8 (c 2.0, MeOH); GC–MS 70 eV, m/z + ane (3 2 ml), washing the organic phase with H2O, drying with (rel. int.): 224 [M] (4), 209 (6), 191 (7), 163 (3), 139 (19), 138 Na2SO4, and solvent removal, 3, 13 and 18 were obtained with (29), 125 (9), 124 (19), 123 (67), 109 (35), 97 (12), 96 (17), 95 acceptable purity; 11 and 12 (each ca. 90% pure) were further puri- (56), 85 (14), 83 (22), 82 (100), 81 (29), 79 (11), 71 (11), 69 (49), fied on diol column IV with EtOAc–hexane (6:94, v/v) to give 98% 68 (13), 67 (21), 57 (11), 55 (24), 43 (11), 41 (27). pure analytical samples. Fraction 4a–4 was acetylated as above and separated on column III with EtOAc–hexane (0.9:99.1, v/v), 4.3.5. (+)-Isothapsadiene, (3aS,6R,7R,7aS)-3a,6,7,7a-tetramethyl-1,2- giving 97% pure 17-OAc; this was hydrolyzed as above to give an dimethylideneoctahydro-1H-indene (4) 21 1 13 analytical sample of 17. Oil; ½a D +34 (c 1.2; n-hexane); for H and C NMR spectro- Part of LC fraction 4b was separated on column III with EtOAc– scopic data, see Tables 2 and 4; HRESIMS: [M+H]+ m/z found hexane (1:9, v/v) giving sub-fractions 4b-S1 through -S9. Fraction 205.1952, calcd. for C15H25 205.1956; GC–MS 70 eV, m/z (rel. 4b-S5 contained 9% 14; it was retained for further purification. int.): 204 [M]+ (18), 189 (11), 175 (2), 161 (2), 147 (11), 135 (17), Fraction 4b-S6 and -S7 were rich in 9; these were combined and 134 (12), 133 (12), 123 (37), 122 (39), 121 (100), 120 (17), 119 re-chromatographed on the same column using EtOAcAhexane (25), 105 (44), 91 (17), 79 (9), 77 (10), 55 (8), 41 (11). (8:92, v/v), giving an analytical sample of 96% pure 9. Fraction 4b-S9 was a mixture of 2 and 7; it was separated by prep. GC at 4.3.6. (+)-b-Isothapsenol, [(2S,3aS,6R,7R,7aS)-3a,6,7,7a-tetramethyl- 225 °C giving analytical samples of 2 and 7. 1-methylideneoctahydro-1H-inden-2-yl]methanol (5) 21 1 13 Another portion of LC fraction 4b was separated on prep. HPLC Oil; ½a D +73 (c 0.8; n-hexane); for H and C NMR spectro- column II using EtOAc–hexane (15:85, v/v), giving sub-fractions scopic data, see Tables 2 and 4; HRESIMS: [M+H]+ m/z found
4b-P1 to -P7. Fraction 4b-P5 had 14% of 14; it was combined with 223.2059, calcd. for C15H27O 223. 2062; GC–MS 70 eV, m/z (rel. fraction 4b-S5, acetylated, and subjected to HPLC on column III int.): 222 [M]+ (9), 207 (27), 191 (14), 189 (12), 167 (16), 152 using EtOAc–hexane (1:99, v/v), giving a fraction with 68% of 14- (10), 140 (64), 139 (100), 135 (27), 133 123 (44), 121 (93), 109 OAc. This was hydrolyzed as above and separated on column III (78), 107 (40), 105 (44), 95 (27), 93 (28), 91 (43), 79 (22), 69 using EtOAc–hexane (9:91, v/v), giving an analytical sample of (16), 55 (26), 41 (27). 98% pure 14. A third portion of LC fraction 4b was separated on diol column 4.3.7. (+)-a-Isothapsenol, [(3aS,4R,5R,7aS)-3,3a,4,5,7a-pentamethyl- IV with EtOAc–hexane (9:91, v/v), giving several sub-fractions, one 3a,4,5,6,7,7a-hexahydro-1H-inden-2-yl]methanol (6) 21 of which was rich in 15. It was separated on column III with Colorless crystals, mp (uncorr.) 84.6 °C; ½a D +36 (c 1.3; n-hex- EtOAc–hexane (9:91, v/v) giving 94% pure 15. ane); for 1H and 13C NMR spectroscopic data, see Tables 2 and 4; 1556 L.G. Cool et al. / Phytochemistry 71 (2010) 1545–1557
+ HRESIMS: [M+H] m/z found 223.2085, calcd. for C15H27O 4.3.14. (+)-Allothapsenol, [(2S,3aS)-3,3,4,4,7-pentamethyl- 223.2062; GC–MS 70 eV, m/z (rel. int.): 222 [M]+ (23), 207 (3), 2,3,3a,4,5,6-hexahydro-1H-inden-2-yl]methanol (13) 21 191 (8), 165 (4), 152 (23), 151 (16), 139 (90), 138 (30), 137 (27), Colorless crystals, mp 70 °C (decomp.); ½a D +36 (c 0.4; n-hex- 133 (23), 123 (50), 121 (100), 119 (28), 109 (43), 107 (19), 105 ane); for 1H and 13C NMR spectroscopic data, see Tables 3 and 4; + (27), 95 (13), 93 (15), 91 (24), 79 (14), 69 (10), 67 (10), 55 (21), HRESIMS: [M+H] m/z found 223.2066, calcd. for C15H27O 43 (12), 41 (23). 223.2062; GC–MS 70 eV, m/z (rel. int.): 222 [M]+ (7), 207 (8), 124 (8), 123 (82), 122 (81), 121 (100), 120 (23), 119 (9), 108 (12), 4.3.8. ( )-b-Ligustigrenol, [(2S,3aR)-3,3,3a,4,4-pentamethyl- 107 (50), 105 (18), 95 (8), 93 (11), 91 (21), 81 (10), 79 (12), 77 2,3,3a,4,5,6-hexahydro-1H-inden-2-yl]methanol (7) (10), 69 (10), 67 (8), 55 (9), 41 (20). 21 1 13 Oil; ½a D 11 (c 1.6; n-hexane); for H and C NMR spectro- + scopic data, see Tables 2 and 4; HRESIMS: [M+H] m/z found 4.3.15. (+)-Isoprethapsenol, [(3aS,5S)-3,3,4,4,7-pentamethyl- 223.2048, calcd. for C15H27O 223.2062; GC–MS 70 eV, m/z (rel. 2,3,3a,4,5,6-hexahydro-1H-inden-5-yl]methanol (14) int.): 222 [M]+ (3), 207 (38), 189 (14), 166 (49), 151 (100), 135 21 1 13 Oil; ½a D +12 (c 0.4; n-hexane); for H and C NMR spectro- (30), 133 (27), 121 (29), 119 (16), 107 (28), 105 (23), 95 (14), 93 scopic data, see Tables 3 and 4; HRESIMS: [M+H]+ m/z found (16), 91 (22), 79 (14), 69 (12), 55 (12), 41 (17). 223.2035, calcd. for C15H27O 223.2062; GC–MS 70 eV, m/z (rel. int.): 222 [M]+ (31), 208 (10), 207 (56), 189 (14), 166 (68), 151 4.3.9. (+)-a-Ligustigrenol, [(2S,7aR)-1,1,7,7,7a-pentamethyl- (45), 137 (18), 136 (17), 135 (100), 133 (25), 121 (51), 119 (20), 2,4,5,6,7,7a-hexahydro-1H-inden-2-yl]methanol (8) 107 (30), 105 (30), 95 (21), 93 (27), 91 (28), 81 (15), 79 (17), 77 21 1 13 Oil; ½a D +146 (c 0.7; n-hexane); for H and C NMR spectro- (16), 69 (15), 55 (15), 43 (10), 41 (25). scopic data, see Tables 2 and 4; HRESIMS: [M+H]+ m/z found 223.2009, calcd. for C H O 223.2062; GC–MS 70 eV, m/z (rel. 15 27 4.3.16. ( )-b-Epithapsenol, [(2S,3aS,7aR)-3a,7,7,7a-tetramethyl-1- int.): 222 [M]+ (9), 207 (49), 191 (100), 189 (23), 151 (22), 135 methylideneoctahydro-1H-inden-2-yl]methanol (15) (46), 133 (18), 123 (19), 121 (30), 119 (24), 109 (23), 107 (25), 21 Colorless crystals, mp (uncorr.) 68.1 °C; ½a 3.7 (c 2.1; n-hex- 105 (20), 95 (11), 93 (18), 91 (25), 79 (13), 77 (12), 69 (30), 55 D ane); for 1H and 13C NMR spectroscopic data, see Tables 3 and 4; (14), 41 (19). + HRESIMS: [M+H] m/z found 223.2054, calcd. for C15H27O 223.2062; GC–MS 70 eV, m/z (rel. int.): 222 [M]+ (7), 207 (25), 4.3.10. (+)-b-Isoligustigrenol, [(2S,3aR,4R,5R)-3,3,3a,4,5-pentamethyl- 191 (27), 189 (24), 137 (34), 135 (34), 133 (35), 123 (84), 122 2,3,3a,4,5,6-hexahydro-1H-inden-2-yl]methanol (9) 21 1 13 (26), 121 (65), 120 (24), 119 (50), 109 (50), 108 (19), 107 (100), Oil; ½a D +22 (c 0.7; n-hexane); for H and C NMR spectro- + 105 (48), 95 (57), 93 (45), 91 (55), 81 (47), 79 (32), 77 (27), 69 scopic data, see Tables 2 and 4; HRESIMS: [M+H] m/z found 1 (48), 67 (24), 55 (42), 43 (27), 41 (50); H NMR (400.13 MHz, CDCl3, 223.2060, calcd. for C H O 223.2062; GC–MS 70 eV, m/z (rel. 15 27 solvent ref. 7.26, d from TMS): 0.98 (3H, s), 1.09 (3H, s), 1.11 (3H, s), int.): 222 [M]+ (7), 207 (100), 189 (28), 165 (5), 161 (7), 147 (8), 1.16 (3H, s), 1.23 (1H, dd, J = 13.0, 2.6 Hz), 1.15–1.35 (3H, m), 1.40– 133 (18), 121 (37), 107 (29), 105 (24), 95 (14), 93 (18), 91 (21), 1.75 (4H, m), 1.96 (1H, dd, J = 12.1, 11.7 Hz), 2.78 (1H, m), 3.62 (1H, 83 (11), 81 (11), 79 (13), 77 (12), 69 (14), 55 (14), 41 (19). dd, J = 10.5, 8.7 Hz), 3.94 (1H, dd, J = 10.5, 5.3 Hz), 4.86 (1H, d, J = 2.4 Hz), 5.00 (1H, d, J = 2.8 Hz). 4.3.11. (+)-a-Isoligustigrenol, [(2S,6R,7R,7aR)-1,1,6,7,7a-pentamethyl- 2,4,5,6,7,7a-hexahydro-1H-inden-2-yl]methanol (10) 21 4.3.17. (+)-Thapsadiene, (3aS,7aR)-3a,7,7,7a-tetramethyl-1,2- Colorless crystals, mp (uncorr.) 76.7 °C; ½a D +24 (c 1.1; n-hex- 1 13 dimethylideneoctahydro-1H-indene (16) ane); for H and C NMR spectroscopic data, see Tables 3 and 4; 21 1 13 + Colorless crystals; ½a D +120 (c 0.46; n-hexane); for H and C HRESIMS: [M+H] m/z found 223.2002, calcd. for C15H27O NMR spectroscopic data, see Tables 3 and 4; HRESIMS: [M+H]+ m/z 223.2062; GC–MS 70 eV, m/z (rel. int.): 222 [M]+ (8), 207 (45), found 205.1949, calcd. for C H 205.1956; GC–MS 70 eV, m/z (rel. 191 (100), 189 (23), 161 (4), 147 (8), 135 (19), 121 (48), 119 15 25 int.): 204 [M]+ (27), 189 (37), 161 (11), 148 (11), 147 (10), 135 (11), (24), 109 (16), 107 (18), 105 (15), 95 (11), 93 (11), 91 (17), 83 133 (18), 123 (12), 122 (37), 121 (100), 120 (24), 119 (25), 107 (14), (11), 79 (8), 77 (8), 69 (12), 55 (10), 41 (11). 106 (8), 105 (61), 93 (8), 91 (19), 79 (10), 77 (12), 69 (7), 55 (9), 41 (15). 4.3.12. (+)-a-Preisothapsenol, [(2S,3aS,6R)-1,1,3a,6,7-pentamethyl- 2,3,3a,4,5,6-hexahydro-1H-inden-2-yl]methanol (11) 21 1 13 4.3.18. ( )-a-Thapsenol, [(3aS,7aS)-3,3a,4,4,7a-pentamethyl- Oil; ½a D +76 (c 0.5; n-hexane); for H and C NMR spectro- scopic data, see Tables 3 and 4; HRESIMS: [M+H]+ m/z found 3a,4,5,6,7,7a-hexahydro-1H-inden-2-yl]methanol (17) Oil; 21 16 (c 1.2; n-hexane); for 1H and 13C NMR spectro- 223.2065, calcd. for C15H27O 223.2062; GC–MS 70 eV, m/z (rel. ½a D + + int.): 222 [M] (39), 207 (59), 191 (14), 189 (31), 161 (18), 147 scopic data, see Tables 3 and 4; HRESIMS: major [M+H H2O] m/ + (16), 139 (29), 137 (14), 136 (32), 135 (68), 133 (26), 123 (22), z found 205.1943, calcd. for C15H25 205.1956; minor [M+H] m/z 122 (25), 121 (100), 120 (13), 119 (36), 109 (27), 107 (35), 105 found 223.2044, calcd. for C15H27O 223.2062; GC–MS 70 eV, m/z + (27), 95 (19), 93 (18), 91 (26), 81 (11), 79 (15), 77 (14), 69 (15), (rel. int.): 222 [M] (43), 191 (23), 152 (12), 151 (32), 139 (40), 55 (18), 41 (21). 138 (46), 137 (36), 135 (16), 133 (34), 123 (66), 122 (26), 121 (100), 120 (16), 119 (29), 109 (64), 108 (10), 107 (23), 105 (40), 4.3.13. (+)-b-Preisothapsenol, [(2S,3aS,7aR)-1,1,3a,6,7-pentamethyl- 95 (14), 93 (22), 91 (36), 81 (11), 79 (19), 77 (21), 69 (19), 67 2,3,3a,4,5,7a-hexahydro-1H-inden-2-yl]methanol (12) (15), 55 (26), 53 (11), 43 (19), 41 (38). 21 1 13 Oil; ½a D +31 (c 0.9; n-hexane); for H and C NMR spectro- scopic data, see Tables 3 and 4; HRESIMS: [M+H]+ m/z found 4.3.19. (+)-Oshalagrenol, [(2S,5R)-1,1,4,6,6-
223.2054, calcd. for C15H27O 223.2062; GC–MS 70 eV, m/z (rel. pentamethylspiro[4.4]non-3-en-2-yl]methanol (18) + 21 int.): 222 [M] (33), 207 (47), 191 (24), 189 (24), 161 (10), 153 Colorless crystals, mp (uncorr.) 46.9 °C; ½a D +189 (c 0.3; n-hex- (7), 147 (10), 139 (24), 136 (21), 135 (49), 133 (16), 123 (27), ane); for 1H and 13C NMR spectroscopic data, see Tables 3 and 4; + 122 (33), 121 (100), 120 (13), 119 (25), 109 (24), 108 (11), 107 HRESIMS: [M+H] m/z found 223.2030, calcd. for C15H27O (48), 105 (23), 95 (16), 93 (19), 91 (24), 81 (13), 79 (14), 77 (12), 223.2062; GC–MS 70 eV, m/z (rel. int.): 222 [M]+ (15), 192 (15), 69 (12), 67 (10), 55 (13), 41 (19). 191 (100), 135 (31), 133 (8), 123 (11), 121 (40), 119 (11), 109 L.G. Cool et al. / Phytochemistry 71 (2010) 1545–1557 1557
(18), 107 (15), 105 (14), 93 (12), 91 (16), 79 (9), 77 (9), 69 (15), 55 Pascual Theresa, J. de, Morán, J.R., Fernández, A., Grande, M., 1986a. Hemiacetalic (10), 43 (8), 41 (14). thapsane derivatives from Thapsia villosa var. minor. Phytochemistry 25, 703– 709. Pascual Theresa, J. de, de Morán, J.R., Fernández, A., Grande, M., 1986b. Nonacetalic Acknowledgements thapsane sesquiterpenoids from Thapsia villosa var. minor. Phytochemistry 25, 1171–1174. Piva, O., 1995. Enantio- and diastereoselective protonation of photodienols: total We are grateful to Christopher Hobbs for collection and identifi- synthesis of (R)-( )-lavandulol. J. Org. Chem. 60, 7879–7883. cation of plant samples, Bor-Sen Chiou for help with DSC measure- Ponder, J.W., 2009. TINKER: Software Tools for Molecular Design, Version 5.0, ments, Delilah Wood for assistance in several NMR experiments, Copyright 1990–2009. Washington University, Saint Louis, MO.