Doubly deuterium-labeled patchouli alcohol from cyclization of singly labeled [2-2H 1]farnesyl diphosphate catalyzed by recombinant patchoulol synthase

Article abstract of DOI:10.1021/ja909251r

Incubations of isotopically pure [2-²H₁](E,E)- farnesyl diphosphate with recombinant patchoulol synthase (PTS) from Pogostemon cablin afforded a 65:35 mixture of monodeuterated and dideuterated patchoulols as well as numerous sesquiterpene hydrocarbons. Extensive NMR analyses ( ¹H and ¹³C NMR, ¹H homodecoupling NMR, HMQC, and ²H NMR) of the labeled patchoulol mixture and comparisons of the spectra with those of unlabeled alcohol led to the conclusion that the deuterium label was located at positions (patchoulol numbering system) C5 (both isotopomers, ca. 100%) and C12 (minor isotopomer, 30-35%), that is, an approximately 2:1 mixture of [5-²H₁]- and [5,12- ²H₂]-patchoulols. Low-resolution FIMS analyses and isotope ratio calculations further corroborated the composition of the mixture as mainly one singly deuterated and one doubly deuterated patchoulol. From a mechanistic point of view, the formation of [5,12²H₂] patchoulol is rationalized through the intermediacy of an unknown exocyclic [7,10:1,5]patchoul-4(12) ene (15-d₁), which could incorporate a deuteron at the C-12 position on the pathway to doubly labeled patchoulol. The corresponding depletion of deuterium content observed in the hydrocarbon coproducts, β-patchoulene and α-guaiene (55% d₀), identified the source of the excess label found in patchoulol-d₂. Comparison of the PTS amino acid sequence with those of other sesquiterpene synthases, and examination of an active site model, suggested that re-orientation of leucine 410 side chain in PTS might facilitate the creation of a 2-pocket active site where the observed deuteron transfers could occur. The retention of deuterium at C5 in the labeled patchoulol and its absence at C4 rule out an alternative mechanism involving two consecutive 1,2-hydride shifts and appears to confirm the previously proposed occurrence of a 1,3hydride shift across the 5-membered ring. Anew, semisystematic nomenclature is presented for the purpose of distinguishing the three different skeletal structures of the patchoulane sesquiterpenes.

Full text of DOI:10.1021/ja909251r

Published on Web 02/11/2010
Doubly Deuterium-Labeled Patchouli Alcohol from Cyclization
of Singly Labeled [2-2H1]Farnesyl Diphosphate Catalyzed by
Recombinant Patchoulol Synthase
Juan A. Faraldos,†,‡ Shuiqin Wu,§,| Joe Chappell,§ and Robert M. Coates*,†
Department of Chemistry, UniVersity of Illinois, 600 South Mathews AVenue,
Urbana, Illinois 61801, and Department of Plant and Soil Sciences, UniVersity of Kentucky,
Lexington, Kentucky 40546-00991
Received October 30, 2009; E-mail: rmcoates@illinois.edu
Abstract: Incubations of isotopically pure [2-2H1](E,E)-farnesyl diphosphate with recombinant patchoulol
synthase (PTS) from Pogostemon cablin afforded a 65:35 mixture of monodeuterated and dideuterated
patchoulols as well as numerous sesquiterpene hydrocarbons. Extensive NMR analyses (1H and 13C NMR,
1H homodecoupling NMR, HMQC, and 2H NMR) of the labeled patchoulol mixture and comparisons of the
spectra with those of unlabeled alcohol led to the conclusion that the deuterium label was located at positions
(patchoulol numbering system) C5 (both isotopomers, ca. 100%) and C12 (minor isotopomer, 30-35%),
that is, an approximately 2:1 mixture of [5-2H1]- and [5,12-2H2]-patchoulols. Low-resolution FIMS analyses
and isotope ratio calculations further corroborated the composition of the mixture as mainly one singly
deuterated and one doubly deuterated patchoulol. From a mechanistic point of view, the formation of [5,12-
2H2]patchoulol is rationalized through the intermediacy of an unknown exocyclic [7,10:1,5]patchoul-4(12)-
ene (15-d1), which could incorporate a deuteron at the C-12 position on the pathway to doubly labeled
patchoulol. The corresponding depletion of deuterium content observed in the hydrocarbon coproducts,
ꢀ-patchoulene and R-guaiene (55% d0), identified the source of the excess label found in patchoulol-d2.
Comparison of the PTS amino acid sequence with those of other sesquiterpene synthases, and examination
of an active site model, suggested that re-orientation of leucine 410 side chain in PTS might facilitate the
creation of a 2-pocket active site where the observed deuteron transfers could occur. The retention of
deuterium at C5 in the labeled patchoulol and its absence at C4 rule out an alternative mechanism involving
two consecutive 1,2-hydride shifts and appears to confirm the previously proposed occurrence of a 1,3-
hydride shift across the 5-membered ring. A new, semisystematic nomenclature is presented for the purpose
of distinguishing the three different skeletal structures of the patchoulane sesquiterpenes.
fermented leaves of P. cablin.5 Since the first total synthesis
Introduction
of patchoulol by Bu¨chi et al in 1962,6 the rare tricyclo-
The tricyclic tertiary alcohol (-)-patchoulol (patchouli alco-
hol, 2) is a crystalline solid at room temperature first isolated
by Gal in 1869.1 This volatile sesquiterpene is the main
component in patchouli oil (35-40%), an essential oil isolated
from the East Indian shrub Pogostemon cablin (Blanco) Beth.
(syn. P. patchouli Pellet). Patchouli alcohol contributes markedly
to the characteristic odor of the oil2 with a threshold of 0.93
ng/L air.3 Patchouli oil is one of the most important natural
fragrances utilized in the perfumery and cosmetic products
industries.4 In fact, well over 1200 tons of patchouli oil is
produced annually by steam-distillation of the dried and
[5.3.1.05,10]undecane carbon skeleton of 2 and the biogenetically
related sesquiterpene seychellene (11, see Figure 4, below)
continue to be interesting synthetic challenges.7
Several mechanistic proposals have been suggested for the
complex, enzyme-catalyzed cyclization of (E,E)-farnesyl diphos-
phate (FPP, 1) to patchouli alcohol (2),8–10 and experimental
work on the biosynthesis of this novel bridgehead alcohol has
been reported by Croteau et al.8 and Akhila et al.9 Native
patchoulol synthase (PTS) from P. cablin (patchouli)11 appears
to carry out this multistep reaction through the intermediacy of
five tertiary carbocations (A-E, Scheme 1) without detectable
amounts of deprotonated intermediates.8,11 On the basis of
labeling results from incubation of [12,13-14C, 1-3H]FPP and
† University of Illinois.
‡ Present address: Cardiff School of Chemistry, Cardiff University,
Cardiff, UK.
§ University of Kentucky.
(4) (a) Akhila, A.; Tewari, R. Current Research on Medicinal and
Aromatic Plants 1984, 6, 38–54. (b) Bruneton, J. Pharmacognosy,
Phytochemistry, Medicinal Plants, 2nd ed.; Lavoisier Publishing: Paris,
1999.
| Present address: Sapphire Energy, Inc., San Diego, CA.
(1) Gal, H. Justus Liebigs Ann. Chem. 1869, 150, 374–376. See also ref
3.
(2) (a) Nikiforov, A.; Jirovetz, L.; Buchbauer, G. Liebigs Ann. Chem. 1989,
489–491. (b) Nikiforov, A.; Jirovetz, L.; Buchbauer, G.; Raverdino,
V. Mikrochim. Acta 1988, 2, 193–198.
(5) Kraft, P.; Weymuth, C.; Nussbaumer, C. Eur. J. Org. Chem. 2006,
1403–1412.
(6) (a) Bu¨chi, G.; MacLeod, W. D., Jr. J. Am. Chem. Soc. 1962, 84, 3205–
3206. (b) Bu¨chi, G.; MacLeod, W. D., Jr.; Padilla, O. J. J. Am. Chem.
Soc. 1964, 86, 4438–4444.
(3) Kraft, P.; Eichenberger, W.; Frech, D. Eur. J. Org. Chem. 2005, 3233–
3245.
9
2998 J. AM. CHEM. SOC. 2010, 132, 2998–3008
10.1021/ja909251r  2010 American Chemical Society

Doubly Deuterium-Labeled Patchouli Alcohol
A R T I C L E S
Scheme 1. Mechanism Proposed for the Cyclization and
Rearrangement of FPP to Patchoulol (2)8,11
Scheme 3a
a R ) geranyl. Reagents, conditions, and yields: (a) LiCt COEt, THF,
-78 °C and (b) 10% D2SO4/D2O, -40 to 25 °C, 96% overall, 98.8%
d-incorporation; (c) i-Bu2AlH, toluene, -78 °C, 97%; (d) isomer separation;
(e) CH3SO2Cl, s-collidine, LiCl, DMF, 0 °C, 98%; (f) HOPP(NBu4)3,
CH3CN, 25 °C; ion exchange, 58%.
Scheme 2. Biosynthetic Pathways for Conversion of [2-2H1]FPP
(1-d1) to Patchouli Alcohol-d1 Isotopomersa
acterized as patchoulol synthase (PTS).13 The recombinant
cyclase catalyzed the conversion of FPP to a mixture of
patchoulol (37%) and a suite of sesquiterpene hydrocarbons
generated in a composition closely resembling that of fresh leaf
extracts. The availability of recombinant PTS and the capability
of producing substantial amounts of patchoulol prompted us to
undertake a reinvestigation of the mechanism by means of stable
isotope labeling and NMR analysis. While 1,3-hydride shifts
occur commonly in terpene synthase cyclization mechanisms,14
we are unaware of any that take place across a 5-membered
ring, ie requiring a transition state that would seem to resemble
a highly strained bicyclo[2.1.1]pentane.
PTS-catalyzed cyclization of FPP substrate labeled with
deuterium at C2 would afford patchoulol labeled at C5 if a 1,3-
hydride shift occurred (pathway a in Scheme 2), or alternatively
the label would move to C4 if two 1,2-hydride shifts had taken
place (pathway (b) in Scheme 2). The location of deuterium in
labeled patchoulol could be readily determined by high field
NMR analyses. This paper provides evidence supporting the
1,3-hydride shift pathway (a), and the remarkable finding that
some 35% of the labeled patchouli alcohol product derived from
[2-2H1]FPP by the action of recombinant PTS is actually doubly
labeled at C5 and C12.
a Pathway (a) presents the accepted route to compound 2 via a 1,
3-hydride shift (BfC).8,11 Pathway (b) is a hypothetical variant of (a)
involving two consecutive 1,2-hydride shifts to effect the B fC conversion.
[12,13-14C, 6-3H]FPP with cell-free extracts of patchouli leaves
3
and degradation analyses of the resulting [14C, H]patchoulol,
Croteau and co-workers8 concluded that the rearrangement step
BfC takes place primarily by a 1,3-hydride shift (path a,
Scheme 2) instead of an ostensibly less strained alternative
involving two consecutive 1,2-hydride shifts (path b). Sequential
1,2-hydride shifts have been proposed to occur during the in
ViVo biosynthesis of R-, ꢀ-, γ- and δ-patchoulenes from (4R)-
[4-3H, 2-14C]mevalonic acid in P. cabli.12 However, the use of
this particular labeled form of mevalonate should generate
[2,6,10-3H, 4,8,14-14C]-FPP in the plant, and therefore does not
distinguish between the different hydride shift alternatives.
Of five sesquiterpene synthase genes from P. cablin recently
cloned and functionally expressed, one (PatTps177) was char-
Results and Discussion
Synthesis of [2-2H1]farnesyl Diphosphate. The synthesis of
[2-2H1]farnesol (5) and its corresponding diphosphate (1-d1) is
outlined in Scheme 3. Addition of ethoxyacetylide to gerany-
lacetone followed by hydrolysis with D2SO4/D2O15 afforded a
mixture of ethyl esters 4 (96%) that was reduced to a 1:1 mixture
of cis and trans deuterated farnesols from which essentially pure
(E,E)-[2-2H1]farnesol (5) was isolated by chromatography. The
(7) (a) Danishefsky, S.; Dumas, D. Chem. Commun. (London) 1968, 1287–
1288. (b) Mirrington, R. N.; Schmalzl, K. J. J. Org. Chem. 1972, 37,
2871–2877; 2877-2881. (c) Naef, F.; Ohloff, G. HelV. Chim. Acta
1974, 57, 1868–1870. (d) Yamada, K.; Kyotani, Y.; Manabe, S.;
Suzuki, M. Tetrahedron 1979, 35, 293–298. (e) Bertrand, M.; Teisserie,
P.; Pelerin, G. Tetrahedron Lett. 1980, 21, 2051–2054; 2055-2056.
(f) Naef, F.; Decorzant, R.; Giersch, W.; Ohloff, G. HelV. Chim. Acta
1981, 64, 1387–1397. (g) Bertrand, M.; Teisserie, P.; Pelerin, G. NouV.
J. Chem 1983, 7, 61–65. (h) Niwa, H.; Hasegawa, T.; Ban, N.;
Yamada, K. Tetrahedron 1987, 43, 825–834. (i) Cory, R. M.; Bailey,
M. D.; Tse, D. W. C. Tetrahedron Lett. 1990, 31, 6839–6842. (j)
Magee, T. V.; Stork, G.; Fludzinski, P. Tetrahedron Lett. 1995, 36,
7607–7610. (k) Kaliappan, K. P.; Subba Rao, G. S. R. J. Chem. Soc.,
Perkin Trans. 1 1997, 1385–1389. (l) Srikrishna, A.; Satyanarayana,
G. Tetrahedron: Asymmetry 2005, 16, 3992–3997.
1
500 MHz H NMR spectrum showed the total absence (>98
2
1
atom % H) of the diagnostic H NMR triplet (J ) 7.0 Hz)
centered at 5.413 ppm and assigned to the vinyl hydrogen at
C2 of unlabeled (E,E)-farnesol. The 77 MHz 2H NMR spectrum
of [2-2H1]-farnesol (5) revealed a single olefinic resonance at δ
5.469 ppm (broad singlet) locating directly the deuterium label
and proving the absence of detectable amounts of label at other
positions. In addition, the doublet (J ) 7.0 Hz) centered at δH
(8) Croteau, R.; Munck, S. L.; Akoh, C. C.; Fisk, H. J.; Satterwhite, D. M.
Arch. Biochem. Biophys. 1987, 256, 56–68.
(13) Deguerry, F.; Pastore, L.; Wu, S.; Clark, A.; Chappell, J.; Schalk, M.
Arch. Biochem. Biophys. 2006, 454, 123–136.
(9) Akhila, A.; Sharma, P. K.; Thakur, R. S. Phytochemistry 1988, 27,
2105–2108.
(14) (a) Coates, R. M. In Progress in the Chemistry of Organic Natural
Products; Herz, W., Grisebach, H., Kirby, J. W., Eds.; 1976; Vol. 33,
pp 73-230. (b) Cane, D. E., Ed. Isoprenoids Including Carotenoids
and Steroids; Barton, D.; Nakanishi, K., Meth-Cohn, O., Eds;
Comprehensive Natural Product Chemistry; Elsevier Science: Am-
sterdam, 1999; Vol. 2, Chapters 5, 6, 8, 10, and 11. (c) Davis, E. M.;
Croteau, R. Top. Curr. Chem. 2000, 209, 53–95.
(10) (a) Hendrickson, J. B. Tetrahedron 1959, 7, 82–89. (b) Parker, W.;
Roberts, J. S.; Ramage, R. Chem. Soc. ReV. 1967, 21, 331–363. (c)
Francis, M. J. O. Planta Medica 1972, 22, 201–204.
(11) Munck, S. L.; Croteau, R. Arch. Biochem. Biophys. 1990, 282, 58–
64.
(12) Akhila, A.; Sharma, P. K.; Thakur, R. S. Phytochemistry 1987, 26,
2705–2707.
(15) Anastasia, M.; Allevi, P.; Ciuffreda, P.; Fiecchi, A.; Gariboldi, P.;
Scala, A. J. Chem. Soc., Perkin Trans. I 1985, 595–599.
9
J. AM. CHEM. SOC. VOL. 132, NO. 9, 2010 2999

A R T I C L E S
Faraldos et al.
1
Table 1. 13C (125 MHz) and H (600 MHz) NMR Chemical Shifts,
Multiplicities, J Values, and Assignments for Patchoulol (2) in
CDCl3
assignmenta
δ13Cb
multiplicitya c
δ1Ha b
,
,
C1
C2
75.64
32.64
C
CH2
1.511 dd (5.9, 3.6)
1.720 dd (12.6, 5.6)
1.373 dt (12.3, 5.9)
1.481 t (5.4)
C3
28.56
CH2
C4
C5
C6
28.07
43.66
24.29
CH
CH
CH2
1.967 dq (6.6, 3.0)
1.447 ddd (8.7, 5.4, 3.0)
1.298 dd (10.8, 3.6)
1.833 dd (11.4, 6.6)
1.195 br singlet
C7
C8
39.04
24.53
CH
CH2
1.268 dt (10.2, 3.0)
1.467 m
C9
28.81
CH2
1.043 t (13.8)
1.821 dd (11.4, 7.2)
C10
C11
C12
C13
C14
C15
37.63
40.08
18.52
20.62
26.81
24.27
C
C
CH3
CH3
CH3
CH3
0.796 d (6.6)
0.848 (s)
1.070 (s)
1.082 (s)
a Determined from HMQC 2D NMR experiments. For carbon
numbering see Scheme 1. b In agreement with ref 18. c Determined from
DEPT experiments.
4.15 ppm for the two C1 allylic protons was replaced by a sharp
singlet (4.14 ppm) in the 1H NMR spectrum of [2-2H1]farnesol.
Low resolution FI MS analyses of both unlabeled and labeled
farnesols and isotopic ratio calculations confirmed a deuterium
content of 98.8% in the labeled compound. Conversion of
alcohol 5 to the corresponding allylic chloride 6 following
Meyer’s conditions,16 and SN2 displacement with tris(tetrabu-
tylammonium) pyrophosphate in CH3CN17 afforded diphosphate
1-d1 as the trisammonium salt in 58% overall yield after ion
exchange and recrystallization from MeOH at 0 °C. The
Figure 1. 1H NMR spectra (600 MHz, CDCl3) of (A) unlabeled and (B)
[2H]patchoulol. The three arrows mark regions where the spectra in (A)
and (B) differ as a consequence of deuterium labeling.
patchoulol (Figures 1 and 2) thus obtained clearly showed: (1)
the complete absence of the characteristic signal corresponding
to the C5 methine proton (1.447 ppm) and (2) a reduced
integration for the C12 methyl doublet (0.796 ppm, 3H, ([5-
2H1]-2, 65%)).20a Careful integrations in the high field region
relative to the C13 methyl singlet (0.848 ppm, 3H, ([5-2H1]-
and [5,12-2H2]-2, 100%)) showed, in addition to the C12 methyl
doublet mentioned above (Figure 2A), a new resonance at 0.778
ppm (2H, [5,12-2H2]-2, 35%) with the multiplicity of a doublet
of triplets (J ) 6.6 and 1.8 Hz) (Figure 2B) assigned to a CH2D
group. The small upfield displacement (∆δ -0.02 ppm) is
typical for a geminal deuterium isotope shift.21 In 1H homode-
coupling NMR experiments (500 MHz), selective and complete
saturation of the downfield signal at δH 1.967 (H4, Table 1)
resulted, as expected, in the observation of the characteristic
C12 CH3 as a singlet (0.793 ppm, 1.94 H) accompanied by a
new singlet (0.775 ppm, 1.06 H) shifted 0.018 ppm upfield due
to the presence of a geminal deuterium in the CH2D group
(Figure 2C).21 Similarly, analysis of the HMQC spectrum of
[2H]patchoulol (2; see Supporting Information) showed that the
cross peak to C5 (43.66 ppm) was absent, and in addition to
the characteristic H12/C12 crosspeak at 18.52 ppm, a small new
crosspeak (18.05) shifted upfield 0.47 ppm by the 2H13C isotope
effect22 was also noticeable. These deductions were confirmed
1
deuterated FPP was characterized by satisfactory H and 31P
NMR spectra.
1
Complete assignments of the H (600 MHz) and 13C (125
MHz) NMR spectra of authentic unlabeled patchoulol (Table
1) were made by two-dimensional 1H-13C chemical shift
correlations acquired in the inverse mode (HMQC) and were
confirmed by comparisons with the existing literature data.18,19
Of particular importance are the assignments of the resonances
corresponding to the methine protons C4 and C5 as well as the
C12 methyl group. The assignments in Table 1 are consistent
with those previously reported by Barton et al.18b
Enzymatic Cyclization of (E,E)-[2-2H1]farnesyl Diphosphate. A
preparative-scale incubation of [2-2H1]farnesyl PP (1-d1) with
recombinant PTS performed according to Deguerry et al13 (pH
7.4; 5 mM MgCl2, 30 °C) afforded deuterated patchoulols (2)
in ca. 13% yield following purification by silica gel chroma-
1
tography. Initial 600 MHz H NMR spectra of the deuterated
(16) Meyers, A. I.; Collington, E. W. J. Org. Chem. 1971, 36, 3044–3045.
(17) Woodside, A. B.; Huang, Z.; Poulter, C. D. Org. Synth. 1993, 8, 616–
620, Coll.
(18) For previous 13C NMR assignments see: (a) Neszmelyi, A.; Lukacs,
G. J. Chem. Soc., Chem. Commun. 1981, 999–1001. (b) Barton,
D. H. R.; Beloeil, J. C.; Billion, A.; Boivin, J.; Lallemand, J. Y.;
Mergui, S. HelV. Chim. Acta 1987, 70, 273–280. (c) Arantes, S. F.;
Hanson, J. R.; Hitchcock, P. B. Phytochemistry 1999, 52, 635–638.
(d) Aleu, J.; Hanson, J. R.; Hernandez Galan, R.; Collado, I. G. J.
Nat. Prod. 1999, 62, 437–440. (e) Wu, S.; Schalk, M.; Clark, A.; Miles,
R. B.; Coates, R. M.; Chappell, J. Nat. Biotechnol. 2006, 24, 1441–
1447.
2
by the 77 MHz H NMR spectrum (Figure 3) of the PTS-
(20) The % value in parenthesis refers to the relative populations of each
[2H]patchoulol calculated by integration of the individual (a) 1H NMR
or (b) 2H NMR signals.
(21) Bovey, F. A. Nuclear Magnetic Resonance Spectroscopy, 2nd ed.;
Academic Press: San Diego, 1988; p 141.
1
(19) For previous H NMR assignments see ref 18b.
(22) Garson, M. J.; Staunton, J. Chem. Soc. ReV. 1979, 8, 539–561.
9
3000 J. AM. CHEM. SOC. VOL. 132, NO. 9, 2010

Doubly Deuterium-Labeled Patchouli Alcohol
A R T I C L E S
Figure 3. 2H NMR spectrum (77 MHz, CHCl3) of the isomeric mixture of
[5-2H1]- and [5,12-2H2]-patchoulols (2) generated by incubation of
[2-2H1]FPP with recombinant PTS. (Inset) Expansion of the upfield region.
ppm). Instead, a new broad resonance was observed upfield (∆δ
-0.54 ppm)22 at 43.12 ppm as a weak 1:1:1 triplet (J ) 20
Hz). Furthermore, in addition to the sharp peak at 18.52 ppm
for the C12 CH3 group, a new signal (18.05 ppm, weak 1:1:1
triplet (J ) 21.2 Hz)) shifted upfield by 0.47 ppm was observed
and assigned to the CH2D group.
The NMR evidence presented above strongly suggests that
the incubation of isotopically pure [2-2H1]farnesyl diphosphate
(1-d1) with recombinant PTS generated an inseparable mixture
of two deuterated patchoulols ([5-2H1]- and [5,12-2H2]-2) in
about 65:35 ratio. The total loss of the H5/C5 HMQC crosspeak
at δ 1.447, F2 (1H) and δ 43.66, F1 (13C), and the apparent
disappearance of the sharp C5 13C NMR signal at 43.66 ppm
unambiguously locate the deuterium label at the C5 position in
both [2H]patchoulols, thus ruling out two sequential 1,2-H shifts
(path (b), Scheme 2) as an alternative mechanism for the cationic
transformation B f C. As illustrated in Figure 2 (B and C)
1
and Figure 3, the appearance of a second H NMR resonance
for H12 centered at 0.778 ppm (2H NMR δ 0.803 ppm, Figure
3) and the presence of a new 13C NMR signal at 18.05 ppm,
both shifted upfield by 0.02 ppm and 0.47 ppm respectively,21,22
locate a second deuterium label at the C12 position. Thus the
minor PTS-generated [2H]patchoulol isotopomer (35%) corre-
sponds to the doubly deuterated form [5,12-2H]-2.
The partial incorporation (35%) of a second deuterium atom
into the [2H]patchoulol product was also verified by careful
analysis of the low resolution FI mass spectrum. The molecular
ion region of labeled 2 displayed peaks at m/z 222 (2%), 223
(100%), 224 (68%), 225 (34%), 226 (17%) and 227 (6%), and
therefore indicated, as shown in Table 2, that the mixture is
constituted mainly of d1 and d2 species in about 2:1 ratio (1.96:
1). This result is in excellent agreement with the NMR
integrations which show a composition mixture of d1 and d2
species in a 1.86:1 ratio (65:35).
Figure 2. 1H NMR spectra (600 MHz, CDCl3) of the H12 methyl doublet
of unlabeled (A) and [2H]-labeled patchoulol (B), and 500 MHz 1H
homodecoupled NMR spectrum (C) with selective and complete saturation
of the downfield signal corresponding to H4 at δH 1.967 ppm.
Analysis of the Hydrocarbon Fraction by 1H NMR
Spectroscopy and GC-MS. In addition to patchoulol (2),
recombinant PTS is known to produce at least 15 other
sesquiterpenes.13 In the present work, diagnostic 1H NMR
signals23,24 in combination with GC/MS data were used to
identify 8 sesquiterpene hydrocarbons produced by recombinant
PTS. Thus, preparative TLC on silica gel afforded two C15
generated [2H]patchoulol which displays two resonances at δ
1.44 (D at C5, 70%) and 0.80 (D at C12, 30%) ppm as broad
singlets, respectively.20b
The 13C NMR spectrum (125 MHz, CDCl3) of the [2H]patchou-
lol product was very similar to that of unlabeled patchouli
alcohol (Supporting Information). Noticeable, however, was the
strong attenuation of the signal corresponding to C5 (43.66
1
hydrocarbon fractions according to GC-MS data (Table 2). H
NMR (500 MHz) analysis of the less polar fraction (fraction
9
J. AM. CHEM. SOC. VOL. 132, NO. 9, 2010 3001

A R T I C L E S
Faraldos et al.
Table 2. Isotope Ratios of Sesquiterpene Hydrocarbons and
Patchoulol Calculated from GC/MS (FI) Data for Products from
Incubation of [2-2H1]FPP with Recombinant PTS
synthetic δ-patchoulene at 60 MHz (0.78, 0.88, 0.95, 1.00, and
5.00 ppm).24 Based on these considerations the structure of this
product was tentatively identified as δ-patchoulene (9, ca. 4%),
a minor sesquiterpene known to be present in patchouli oil,12,25
but not previously recognized as a product of recombinant
PTS.13 Similarly, analysis of fraction 2 (more polar fraction)
b
compound and no.
rel. %a GC tR
d0
d1
d2
d3
d4
Sesquiterpene hydrocarbons
ꢀ-Patchoulene (7)
Caryophyllene (8)
δ-Patchoulene (9)c
R-Guaiene (10)
Seychellene (11)
R-Patchoulene (12)
Unidentified
1.7 17.8 57.40 22.37 13.64 6.46 0.12
3.5 18.5
2.1 19.2
4.80 84.96 4.96 5.28 0.00
0.00 57.00 32.49 10.26 0.25
1
by H NMR spectroscopy revealed, in addition to R-guaiene
(26%) and seychellene (9%), and an unidentified sesquiterpene
(9%) (Table 2), the presence of diagnostic resonances for (in
order of GC elution): ꢀ-patchoulene (7, ca. 4%),23b ꢀ-caryo-
phyllene (8, ca. 7%),23c guaia-4,11-diene (13, 18%),23d and
R-bulnesene (14, 26%).23a The sesquiterpene profile observed
in the present work is in excellent agreement with that previously
reported by Deguerry et al.13
29.4 19.8 55.51 28.92 11.38 3.57 0.61
25.2 20.0
10.1 20.4
4.4 21.2
0.70 58.57 27.57 11.98 1.49
0.78 57.50 29.50 11.85 0.61
0.00 58.44 31.73 8.73 1.10
0.81 73.68 24.27 0.79 0.46
0.63 56.80 28.01 10.29 4.28
Guaia-4,11-diene (13) 10.4 21.7
R-Bulnesene (14)
13.2 22.0
Sesquiterpene alcohols
[2H]Patchoulol (2)
1.16 52.62 27.06 12.65 6.52
1.14 98.86
Starting FPP (1-d1)d
As shown in Table 2, incorporation of deuterium into the
sesquiterpene hydrocarbon fractions took place at three distinc-
tive and characteristic levels. While hydrocarbons 7 and 10
(Figure 4 and Table 2) were produced by the recombinant
cyclase mainly as a 2:1 mixtures of d0 and d1 species,
sesquiterpenes 9, 11, 12, 14, and an unidentified hydrocarbon
as well as patchoulol 2 were generated mainly as 2:1 mixtures
of d1 and d2 labeled forms. On the other hand, ꢀ-caryophyllene
(8) and guaia-4,11-diene (13) were found to be essentially
monodeuterated. It is clear that the enrichment of deuterium in
labeled patchoulol is accounted for primarily by the depletion
of deuterium in ꢀ-patchoulene (7, 57% d0) and R-guaiene (10,
56% d0). The enrichment of deuterium in seychellene (11, 59%
d1 + 28% d2) is very similar to that of patchoulol, consistent
with their formation from a common bridgehead ion (Scheme
1, intermediate E).
a Estimated by measurements of GC and 1H NMR peaks. (%)
represent the relative percentage from the total of hydrocarbons. b GC
conditions: see Experimental Part. c Tentative identification. d Deuterium
content of farnesol-d1.
It is worthwhile to comment here upon the confusing
structural relationships and nomenclature for the patchoulane
sesquiterpenes. As is evident from the structures of patchoulol
(2) and the three patchoulene isomers in Figure 4, there are in
fact three different skeletal structures for the “patchoulane”
parent, independent of stereochemistry: deoxypatchoulol
(tricyclo[5.3.1.05,10]undecane), the dihydro derivative of ꢀ- and
δ-patchoulene (7 and 9; tricyclo[6.2.1.02,6]undecane), and di-
hydro R-patchoulene (12, tricyclo[5.3.1.01,5]undecane).26 Since
all three tricycloundecane isomers have the same cyclodecane
part structure (C1-C10) and they differ only in the locations
of the one-carbon bridge (C11) and the zero carbon connections
(0,5,10 0,2,6 and 01,5), we propose the following shorthand names
to distinguish the three patchoulane carbon skeletons as follows
[C11 bridge: ring fusion]: [1,7:5,10]-, [1,7:1,5]-, and [7,10:
1,5]patchoulanes. Thus patchoulol (2), R-patchoulene (12),
ꢀ-patchoulene (7), and δ-patchoulene (9) would be unambigu-
ously named as [1,7:5,10]patchoulan-1-ol, [1,7:1,5]patchoul-
9(10)-ene, and [7,10:1,5]patchoul-1(5)-ene, and [7,10:1,5]patchou-
li-1(2)-ene, respectively.
Figure 4. Structures of the PTS-generated sesquiterpenes identified in the
hydrocarbon fractions by 1H NMR spectroscopy in order of GC elution.
The structural assignment of δ-patchoulene (9) is regarded as tentative.
1) showed the presence of three main constituents: R-guaiene
(10, 33%),23a seychellene (11, 42%),23a and R-patchoulene (12,
21%)23a (Figure 4). Traces of other minor hydrocarbons
accompanied the three major components, the more abundant
1
of which showed four H NMR resonances at δ 0.78 (s), 0.89
The conversion of [2-2H1](E,E)-FPP (1) to patchoulol (2)
catalyzed by recombinant PTS proceeds with total retention of
the deuterium label at C5 (patchoulol skeleton, or C2 farnesol
skeleton). Thus pathway (b) in Scheme 2 is conclusively ruled
out, and evidently intermediate carbocation B undergoes an
apparently strained 1f3 hydride-shift to generate the cationic
intermediate C en route to patchoulol (pathway a). This
(s), 1.00 (s), and 0.98 (d, J ) 6.5 Hz) ppm attributed to three
tertiary and one secondary methyl groups, respectively. In
addition, a broad singlet at δ 5.00 ppm assignable to an olefinic
proton was observed. These H NMR values at 500 MHz are
in excellent agreement with the data previously reported for
1
(23) R-Guaiene, seychellene, R-patchoulene and R-bulnesene, see: (a)
Rakotonirainy, O.; Gaydou, E. M.; Faure, R.; Bombarda, I. J. Essent.
Oil Res. 1997, 9, 321–327. ꢀ-Patchoulene, see: (b) Thomas, A. F.;
Ozainne, M. HelV. Chim. Acta 1979, 62, 361–368. Caryophyllene,
see: (c) Maurer, B.; Grieder, A. HelV. Chim. Acta 1977, 60, 2177–
2190. Guaia-4,11-diene, see: (d) Saritas, Y.; Bulow, N.; Fricke, C.;
Ko¨nig, W. A.; Muhle, H. Phytochemistry 1998, 48, 1019–1023.
(24) δ-Patchoulene, see: Mookherjee, B. D.; Trenkle, R. W.; Ledig, W. O.
J. Agr. Food Chem. 1974, 22, 771–773.
(25) (a) Bunrathep, S.; Lockwood, G. B.; Songsak, T.; Ruangrungsi, N.
ScienceAsia 2006, 32, 293–296. (b) Paknikar, S. K.; Veeravalli, J.
Indian J. Chem, Sect. B 1980, 19B, 432. (c) Bure, C. M.; Sellier, N. M.
J. Essent. Oil Res. 2004, 16, 17–19.
(26) γ-Patchoulene ([1,7:1,5]patchoul-10(12)-ene) is the exocyclic 10(12)
isomer of R-patchoulene (12).
9
3002 J. AM. CHEM. SOC. VOL. 132, NO. 9, 2010

Doubly Deuterium-Labeled Patchouli Alcohol
A R T I C L E S
Scheme 4. Proposed Reaction Pathway to the Deuterated Patchoulols 2-d1 and 2-d2
observation corroborates previous conclusions from biosynthetic
studies by Croteau et al.8 using the native enzyme, and is
consistent with the mechanism presented in Scheme 1 as the
biosynthetic pathway to patchouli alcohol (2). However, this
scheme does not account for the formation of doubly deuterated
patchoulol from singly labeled substrate.
and fusicoccadiene catalyzed by the respective recombinant
diterpene synthases.28
Ko¨llner et al recently proposed that the multiple products
generated by the maize sesquiterpene synthase TPS4 can be
explained by the existence of two different active site pockets.29
The alterations observed in product distributions from TPS4
mutant cyclases were consistent with this model in which
monocyclic bisabolyl products are generated by cyclization of
nerolidyl PP in Pocket I and bicyclic products are formed by
proton-induced cyclization in Pocket II. We propose that the
catalytic site of PTS similarly adjoins a hydrophobic pocket in
which one (or more) of the cyclization products is (are)
sequestered during catalysis. Thus, diffusion of a neutral
intermediate and products may occur between the catalytic site
and the proximal storage pocket.
Mechanism for the Formation of Patchoulol-d2. We propose
that the second deuterium incorporated into the C12 methyl
group of patchoulol occurs through a proton elimination
followed by deuteron addition with the unknown sesquiterpene
[7,10:1,5]patchoul-4(12)-ene-5-d1 (15-d1)27 as an intermediate
(Scheme 4). Addition of D+ to the 4(12) exocyclic double bond
would regenerate intermediate B-d2 in doubly deuterated form
which could proceed through the remaining steps to [5, 12-
2H2]patchoulol (2-d2). The low level of d3 species present in
[2H]patchoulol according to GC/MS analysis (Table 2, 12-13%)
can be explained by a small amount of reversal of the 15-d1 f
B-d2 step by kinetically favored H+ elimination giving some
15-d2 that could incorporate a second deuterium into the C12
methyl group. The elimination reactions forming the double
bonds of ꢀ-patchoulene (7) and R-guaiene (10) provide plausible
sources of the deuteron in view of their isotope depletions. Since
a direct, intermolecular deuteron transfer between intermediates
15-d1 and C-d1 seems inconceivable, recombinant PTS presum-
ably has the capacity to retain the D+ label in the catalytic site
long enough for re-entry of intermediate 15-d1. Mechanistic
schemes involving intramolecular transfers of protons have been
proposed previously to rationalize observed migrations of
deuterium label in the biosynthesis of taxadiene, abietadiene,
Native PTS isolated from P. cabin by Munck and Croteau
was determined by SDS-PAGE analysis to be a dimer composed
of two apparently identical subunits of Mr ≈ 40 000.11 Thus
another way to explain the double deuteration phenomenon
would invoke channeling of proposed intermediate 15-d1 formed
in one monomer active site to the other monomer active site
where the D+ donor was sequestered. This alternative scenario
is similar to the two-pocket shown in Scheme 5, except that in
this case the two active sites in dimeric PTS serve similar
functions. The channeling of intermediates between two catalytic
sites on the same enzyme has been demonstrated in the cases
of the fungal cyclases, ent-kaurene synthase30 and geosmin
synthase.31 However, this homodimer explanation seems to be
excluded by the finding that recombinant PTS actually exists
as a 64-kDa monomer (see Experimental Section), in agreement
with previous results.13
(27) Literature searches with SciFinder turned up reports of three other
[7,10:1,5]patchoulene isomers besides ꢀ-and δ-patchoulenes (7 and
9)andarelatedtertiaryalcohol:[7,10:1,5]patchoul-3(4)-ene(patrinene),27a,b
[7,10:1,5]patchoul-4(5)-ene,27c,d,e [7,10:1,5]patchoul-5(6)-ene,27d and
the tricyclic alcohol [7,10:1,5]patchoul-5-ol,27d,f evidently an erroneous
structure proposed for patchouli alcohol. However, the reliability of
the structure assignments for these naturally occurring patchoulenes
is open to question. No references to [7,10:1,5]patchoul-4(12)-ene (15)
were found. However, [7,10:1,5]patchoula-1(5), 4(12)-diene was an
intermediate in Bu¨chi’s total synthesis of patchoulol.6 SciFinder
references: (a) Peking Medical College, Beijing Yixueyuan Xuebao
1976, 1, 17-22, (b) Liu, M.; Chen, Y.; Bu, Y.; Fang, Y. Shenyang
Yaoxueyuan Xuebao 1993, 10, 301–304. (c) Polovinka, M. P.;
Korchagina, D. V.; Shcherbukhin, V. V.; Gatilov, Yu. v.; Rybalova,
T. V.; Zefirov, N. S.; Barkhash, V. A. Zh. Org. Khim. 1995, 31, 214–
219. (d) Treibs, W. Liebig’s Ann. Chem. 1949, 564, 141–151. (e)
Polovinka, M. P.; Korchagina, D. V.; Shcherbukhin, V. V.; Gatilov,
Yu. v.; Rybalova, T. V.; Zefirov, N. S.; Barkhash, V. A. Tetrahedron
Lett. 1995, 36, 8093–8096. (f) Esafov, V. I.; Novikov, N. I. Zh.
Obshch. Khim. 1949, 19, 1344–1350.
Scheme 5
9
J. AM. CHEM. SOC. VOL. 132, NO. 9, 2010 3003

A R T I C L E S
Faraldos et al.
Scheme 6
The pictorial model illustrated in Scheme 5 offers a credible
explanation for the intermolecular deuterium transfer observed.
Proposed intermediate 15-d1 formed in the catalytic site can be
sequestered in the storage pocket while another [2-2H1]FPP
substrate is bound in the catalytic site where it is converted to
ꢀ-patchoulene (7-d0) or R-guaiene (10-d0) by elimination of D+.
Diffusive exchange of ꢀ-patchoulene (or R-guaiene) with 4(12)-
patchoulene 15-d1 would set the stage for generating the
dideuterated products, provided that the exchange equilibrium
of the deuteron with the external medium is slower than active
site-storage pocket exchange and D+ transfer to C12.
Inspection of the X-ray crystal structure of FPP synthase
containing the inert substrate analogue dimethylallyl thioloPP
and the normal cosubstrate isopentenyl PP in the catalytic site
lead to the suggestion that the OPP ·Mg32+ leaving group accepts
the 2R proton eliminated in the electrophilic coupling reaction.32
Similarly it seems plausible that the OPP ·Mg32+ leaving group
generated by the catalytic action of PTS could serve the same
function. In a least-motion or minimal motion mechanism, the
multidentate leaving group would be positioned close to C5 and
C12 (corresponding to C2 and C15 of FPP). Hence the
DOPP·Mg32+ generated in the D+ elimination forming ꢀ-patchou-
lene (or R-guaiene) would be in a suitable position for deuteron
transfer to C12 of 4(12)-patchoulene (15-d1), if the D/H
exchange with the surrounding aqueous medium is relatively
slow.
The two C-C bonds formed in the germacrenyl f patchou-
lenyl cyclization A-1 f B-1, presumably proceed by a stepwise
syn addition of the carbocation carbon (C11) and C2 across
the 6,7 double bond (Scheme 6, FPP numbering). This
exothermic process must involve considerable adjustment of
bond angles as the two trans double bonds are converted to
C-C single bonds in the rigid tricyclic ion B-1. A least-motion
or minimal-motion process would form the central 6-membered
ring initially in a boat conformation B-1 with orbital alignment
ill-suited for the C1 f C4 hydride transfer. Conformational
inversion to a more stable chair form B-2 would improve the
orientation of the carbocation sp2 orbital for the hydride transfer,
although further distortions of the 5-membered ring would seem
necessary to attain the configuration of a bicyclo[3.1.1]hexane.
Since the proton removed from C12 to form the exocyclic
double bond of 15 must be added again in a later step, the
apparent necessity of passing through this neutral intermediate
is puzzling. One possibility might be a proton elimination at
2+
C12 enforced by the nearby OPP ·Mg3 counterion/base (not
shown) in boat intermediate B-1 to form the unstable conforma-
tion of 15-1 illustrated in Scheme 6. Perhaps the 5-membered
ring of ion B-2 is generated in a puckered conformation in the
protonation of 15-2 so as to facilitate the 1f4 hydride transfer
step. Alternatively the overall B-1 f C-1 transformation might
take place by a mechanism other than the 1f4 hydride shift.
This issue could be resolved by deuterium labeling at C1 (C6
of FPP). Also obscure are the reasons why pathway b via the
more stereoelectronically favorable sequence of 1, 2-hydride
shifts (5f4 and 1f5) does not occur.
The obvious question arises whether the patchoulol product
is actually formed by two different pathways operating in
parallel as implied by Scheme 4. In other words, is [7,10:
1,5]patchoul-4(12)-ene (15) an intermediate on a minor pathway
to patchoulol, competing with the direct sequence B f C f 2
or is 15 in fact an obligatory intermediate? An alternative
scenario is that all of the patchoulol is formed through the neutral
4(12)-patchoulene intermediate 15 and that the d1 + d2 content
of the patchoulol product formed from [2-2H1]FPP reflects a
The PTS-catalyzed formation of the bicyclic sesquiterpenes
R-guaiene (10), guaia-4,11-diene (13) and R-bulnesene (14)
(Figure 4) was previously proposed to occur by a proton-induced
cyclization of the endocyclic double bonds of germacrene A, a
minor hydrocarbon also detected in the product mixture.13
However, with [2-2H1]FPP as the substrate, this route with
protonation at C3 would generate 10-d0, 13-d1, and 14-d1 or by
protonation at C10 would form 10-d1, 13-d0, and 14-d1. In fact,
GC/MS analyses in the present work established that the three
guaienes are actually formed as mixtures of d0 + d1 (2:1), d1 +
d2 (3:1) and d1 + d2 (2:1) species, respectively, (see Table 2).
This discrepancy in deuterium content can be explained by an
alternative route to the bicyclic hydrocarbons independent of
germacrene A. The deuterium label distribution present in these
olefins can be explained by a later divergence from the pathway
to patchoulol (2) after the bridged tricyclic ions C and D. As
shown in Scheme 7, fragmentation of the C11-C10 or C11-C1
bridging bond of carbocation C or D followed by a PTS-
mediated deprotonation at C12 would generate R-bulnesene (14)
as the required mixture of d1 + d2 species. Reprotonation of
the neutral olefin 14, followed by a 1,2-deuteride shift would
give rise to a new carbocation that could lose either a deuteron
or a proton to generate R-guaiene (10) or guaia-4,11-diene (13),
2+
kinetic isotope effect on H/D exchange of DOPP ·Mg3 with
the surrounding aqueous medium. If a deuterium from
DOPP ·Mg32+ can be incorporated into patchoulol by D+ transfer
to intermediate 15, it seems highly likely that the analogous
H+ transfer would also be occurring. Thus, in this scenario the
dominant pathway from FPP to patchoulol passes through 4(12)-
patchoulene as an intermediate.
(28) (a) Williams, D. C.; Carroll, B.; Jin, Q.; Rithner, C.; Lenger, S. R.;
Floss, H. G.; Coates, R. M.; Williams, R. M.; Croteau, R. Chem. Biol.
2000, 7, 969–977. (b) Coates, R. M.; Ravn, M. M.; Jetter, R.; Croteau,
R. Chem. Commun. 1998, 21–22. (c) Toyomasu, T.; Tsukahara, M.;
Kenmoku, H.; Anada, M.; Nitta, H.; Ohkanda, J.; Mitsuhashi, W.;
Sassa, T.; Kato, N. Org. Lett. 2009, 11, 3044–3047.
(29) Ko¨llner, T. G.; O’Maille, P. E.; Gatto, N.; Boland, W.; Gershenzon,
J.; Degenhardt, J. Arch. Biochem. Biophys. 2006, 448, 83–92.
(30) Fall, R. R.; West, C. A. J. Biol. Chem. 1971, 246, 6913–6928.
(31) Jiang, J.; Cane, D. E. J. Am. Chem. Soc. 2008, 130, 428–429.
(32) Hosfield, D. J.; Zhang, Y.; Dougan, D. R.; Broun, A.; Tari, L. W.;
Swanson, R. V.; Finn, J. J. Biol. Chem. 2004, 279, 8526–8529.
9
3004 J. AM. CHEM. SOC. VOL. 132, NO. 9, 2010

Doubly Deuterium-Labeled Patchouli Alcohol
A R T I C L E S
Scheme 7. Proposed Biosynthetic Pathways to R-Guaiene (10), Guai-4,11-diene (13), and R-Bulnesene (14) via Bridge Fragmentation of
Carbocations C and D
respectively (Scheme 7, lower right paths). Alternatively, the
formation of olefin 10 could take place through a protonation/
dedeuteration sequence involving again a tightly bound R-bulne-
sene intermediate (14) (Scheme 7, upper right path).
enzymes.37 The RXR motif (outlined in blue) is juxtaposed to
the first DDXXD motif to assist in initiating ionization and
sequestering of the diphosphate substituent outside the active
site, and releasing the reactive carbocation intermediate into the
active site pocket.14c The amino acid side chains of all the
residues in the first DDXXD and RXR motifs are directly
superimposed upon one another in the PTS and TPS4 structural
models and the solved EAS structure.
The amino acid sequences for patchoulol synthase (PTS;
AAS86323) and the maize TPS4 (Zea mays) (TPS4; AY518310)
enzyme were aligned with those for γ-humulene synthase (gHS;
GenBank accession: AAC0528) and δ-selinene synthase (dSS;
AAC05727) from Abies grandis, and tobacco (Nicotana to-
bacum) epiaristolochene synthase (EAS; AAA19216) using the
publically available ClustalW (v1.4) multiple sequence align-
ment program in order to facilitate identification and discussion
of conserved residues and domains (Figure 5). The PTS
3-dimensional structure was also modeled similarly to that
described for the maize TPS4 enzyme by Ko¨llner et al,29 except
that the Modeler 6.2 software program35 was used to thread
the PTS amino acid sequence upon the structural coordinates
for the N. tabacum EAS enzyme33 (GenBank pdb file: 5EAT)
and to energy minimize the final structural predictions. The EAS
structure solved with the FPP analog farnesyl hydroxyphos-
phonate was used to position substrate docked into the various
models. The PTS structural model was overlaid with that for
the maize TPS4 enzyme29 and the solved EAS structure33 using
the PYMOL software suite (www.pymol.org) (Figure 6).
In Figure 5, identical residues aligning in all 5 enzymes are
back-highlighted in yellow, while residues of similar charge or
polarity conserved among the enzymes are back-highlighted in
green. Regions and residues of the enzymes suspected of playing
particular roles in catalysis are outlined in colors as follows:
The first DDXXD motif (I; outlined in red) positions a divalent
cation near the entrance to the active site cavity and serves to
fold the hydrocarbon portion of the FPP molecule into the active
site pocket while anchoring the diphosphate substituent near
the mouth of the pocket.33,36 A second DDXXD motif (II;
outlined in red), conserved only in γ-humulene and δ-selinene
synthases, has been suggested to provide a second FPP binding
site and possible second active site facilitating the formation of
large numbers of diverse reaction products emanating from these
The residues outlined in magenta have been shown by Ko¨llner
et al29 to be associated with the early steps in the reaction
mechanism of the TPS4 enzyme, especially those steps involved
in the initial ionization, isomerization and cyclization of the FPP
substrate; the same steps correlated with activities putatively
associated with pocket I of the TPS4 enzyme. The residues
outlined in cyan have been associated with the second round
of cyclization reactions occurring in the TPS4 enzyme,29 and
are suspected of either directly participating in the reaction
mechanisms or possibly serving as a gatekeeper for intermediate
movement between reaction pockets I and II. Although leucine
413 of TPS4 is conserved in sequence alignments with leucine
410 and 409 of PTS and EAS, respectively, its spatial orientation
in the molecular model of PTS is directed away from the active
site pocket in contrast to the inward and overlapping orientation
found in the TPS4 and EAS structures (Figure 6). In contrast,
threonine 301 of TPS4 occupies a similar spatial orientation as
isoleucine 296 in EAS, but leucine 297 of PTS appears to extend
a terminal methyl substituent across the active pocket sufficiently
to distort the orientation of leucine 410 on the G2 helix. This
may in turn distort the orientation of the entire G2 helix and
the closely aligned J helix, two helices contributing directly to
active site topology, and thus possibly providing for a larger
reactive pocket or a second active site pocket.
Conclusion
This work has established that recombinant patchouli alcohol
synthase catalyzes the conversion of the singly labeled substrate
[2-2H1]farnesyl diphosphate to doubly deuterated patchoulol
([5,12-2H2]-2) to the extent of ca. 35%, in addition to 65% of
the singly labeled product bearing deuterium at C5 ([5-2H1]-2.
The unexpected incorporation of label into the C12 methyl group
implicates the unknown sesquiterpene [7,10:1,5]patchoul-4(12)-
ene-d1 (15-d1) as a neutral, enzyme-bound intermediate. The
deuterium eliminated in producing the hydrocarbon coproducts,
(33) Starks, C.; Back, K.; Chappell, J.; Noel, J. P. Science 1997, 277, 1815–
1820.
(34) (a) Mathis, J. R.; Back, K.; Starks, C.; Noel, J.; Poulter, C. D.;
Chappell, J. Biochem. 1997, 36, 8340–8348. (b) Greenhagen, B. T.;
O’Maille, P. E.; Noel, J. P.; Chappell, J. Proc. Natl. Acad. Sci. U.S.A.
2006, 103, 9826–9831.
(35) Fiser, A. S.; Sali, A. Methods Enzymol. 2003, 374, 461–491.
(36) Christianson, D. W. Chem. ReV. 2006, 106, 3412–3442.
(37) Little, D. B.; Croteau, R. Arch. Biochem. Biophys. 2002, 402, 120–
135.
9
J. AM. CHEM. SOC. VOL. 132, NO. 9, 2010 3005

A R T I C L E S
Faraldos et al.
Figure 5. Amino acid sequence alignment of patchoulol synthase (PTS) to two multireaction product sesquiterpene synthases, γ-humulene (gHS) and
δ-selinene synthases (dSS) from A. grandis, a maize synthase (TPS4) purported to have an active site divided into 2 pockets,29 and epi-aristolochene synthase
(EAS) from tobacco, with a single active site pocket as evident from crystallographic33 and biochemical studies.34
9
3006 J. AM. CHEM. SOC. VOL. 132, NO. 9, 2010

Doubly Deuterium-Labeled Patchouli Alcohol
A R T I C L E S
Figure 7. Gel chromatography profile of recombinant PTS enzyme
separated on a calibrated Sephadex G-75 column. A BSA molecular size
standard of 66 kDa eluted in fraction 10 (not shown). PTS protein and
enzyme activity eluted in fractions 10-12 with an apparent molecular size
of 64-68 kDa, consistent with the monomeric size of 66 kDa calculated
from the deduced amino acid sequence of the gene.
Figure 6. Close-up view of putative active site residues influencing catalytic
activity in patchoulol synthase (magenta highlighting), epi-aristolochene
synthase (green highlighting) and TPS4 from maize (blue highlighting).
This rendition attempts to accentuate the possible reorientation of leucine
410 in PTS such that it could facilitate creation of a second active site
pocket.
CDCl3) δ 142.7, 135.6, 131.4, 124.3, 123.4, 41.1, 39.6, 39.3, 26.6,
26.0, 25.7, 17.7, 16.1, 16.0. Noticeable is the absence of the olefinic
resonance for C2 (dCD).
(2E,6E)-[2-2H1]Farnesyl Diphosphate, Trisammonium Salt
(1-d1). Reaction of chloride 6 (178.1 mg, 0.74 mmol), and
HOPP(NBu4)3 (1.5 g, 1.7 mmol) in dry CH3CN (15 mL) in the
presence of 4 Å molecular sieve powder (1.5 g) for 24 h at room
temp according to a literature procedure,17 dilution with CH3CN,
filtration, and concentration under reduced pressure (40 °C), cation
exchange, and lyophilization gave the crude NH4 salt as a white
solid. Extractions with methanol followed by concentration, and
recrystallization from MeOH at 0 °C provided 187.1 mg (58%) of
essentially pure diphosphate 1-d1: 1H NMR (500 MHz, CD3OD) δ
5.09 (broad t, J ) 7.5 Hz, 1H), 5.97 (app t, J ) 6.0 Hz, 1H), 4.41
(d, J ) 5.0 Hz, 2H), 2.02-1.84 (m, 8H), 1.73 (s, 3H), 1.58 (s,
3H), 1.55 (s, 3H), 1.49 (s, 3H), 1.48 (s, 3H); 31P NMR (202 MHz,
CD3OD) δ -8.78 (broad d, J ) 100.6 Hz).
Protein-Size Determination. Recombinant PTS harboring a
hexa-histidine amino terminal purification tag (see Supporting
Information) was purified from 100 mL of Escherichia coli cells
overexpressing the corresponding PTS gene as described previously
(i.e., using Ni2+ affinity column).13 The isolated enzyme was
concentrated to 100 µL, dialyzed, and fractionated using a Sephadex
G-75 column (0.9 × 28.5 cm, 40-120 µm particle size, Pfizer,
New York, NY). The column was equilibrated (100 mM Tris, 50
mM NaCl, pH 8.0) and calibrated using the following standards
(500-600 µg): bovine serum albumin (BSA, 66.0 kDa), ovalbumin
(43.0 kDa), ꢀ-lactoglobulin (35.0 kDa), or lysozyme (14.3 kDa).
The molecular size of the recombinant PTS enzyme was then
determined by chromatography on the calibrated Sephadex G-75
column (Figure 7) and shown to be ca. 66 kDa (monomer).
Patchoulol synthase activity eluted from the column with a similar
pattern, but slightly later than the bovine serum albumin (66.0 kDa)
standard, which eluted in fraction 10. This observation is consistent
with previous sizing results obtained from SDS-PAGE experiments
(Figure 7).13 The protein concentration of each fraction (0.43 mL)
was determined with a protein assay reagent (Bio-Rad, Hercules,
CA), and the PTS enzyme activity of each fraction measured in an
enzyme assay with [1-3H] FPP as previously reported.13 The enzyme
showed a maximum specific activity of ∼1 pmol/min/mg enzyme,
corresponding to the fractions 10-12.
ꢀ-patchoulene and R-guaiene (7 and 10), predominantly in
unlabeled form (55% d0) is considered the source of the excess
label found in 2-d2. A “two-pocket” catalytic site capable of
sequestering 15-d1 and a D+ source (eg DOPP·Mg32+) in close
proximity is proposed to rationalize the double labeling. It seems
reasonable to predict that [7,10:1,5]patchoul-4(12)-ene (15) is
a naturally occurring sesquiterpene awaiting discovery.
The complete retention of deuterium at C5 in both labeled
forms of patchoulol excludes a mechanism proceeding by two
consecutive 1,2-hydride shifts (C5 to C4 and C1 to C5), and is
consistent with the occurrence of the previously proposed 1,3-
hydride shift (from C1 to C4)8 through what would appear to
be a stereoelectronically unfavorable transition state resembling
a strained bicyclo[2.1.1]hexane.
Experimental Section
[2-2H1](2Z,6E)- and [2-2H1](2E,6E)-Farnesols (5). A 1:1 mix-
ture of cis,trans and trans,trans esters 4 (560 mg, 2.11 mmol) in
dry toluene (5.0 mL) was reduced with i-Bu2AlH (1 M in hexane,
4.60 mL, 4.60 mmol) under N2 for 3 h at -78 °C. Hydrolysis,
extractions with 20% EtOAc/hexane, and purification by silica gel
flash chromatography using 5% EtOAc/hexane as eluent gave pure
trans,trans [2-2H1]farnesol 5 (168 mg, 34%): 1H NMR (500 MHz,
CDCl3) δ 5.10 (m, 2H), 4.14 (s, 2H), 1.95-2.14 (m, 8H), 1.68 (s,
6H), 1.59 (s, 6H), 1.25 (br s, 1H, OH); 13C NMR (125 MHz, CDCl3)
δ 139.7, 135.3, 131.3, 124.3, 123.7, 59.3, 39.7, 39.5, 26.7, 26.3,
25.7, 17.7, 16.2, 16.0. Noticeable was the apparent absence of the
2
olefinic resonance of C2 (dCD-); H NMR (77 MHz, CHCl3) δ
5.47 (broad s); MS (EI) m/z (rel. int.) 224 (M+ + 1, 12), 223 (M+,
62), 206 (M+ + 1 - H2O, 11), 205 (M+ - H2O, 53), 192 (100),
190 (44), 180 (51); HRMS (EI) m/z calcd for C15H25D1O (M+)
223.2046. Found: 223.2043.
(2E,6E)-[2-2H1]Farnesyl Chloride (6). Reaction of
[2-2H1]farnesol 5 (168 mg, 0.75 mmol), with LiCl (500 mg, 11.8
mmol), s-collidine (500 µL, 547 mg, 4.5 mmol), methanesulfonyl
chloride (250 µL, 170 mg, 1.5 mmol) in DMF (10.0 mL) for 3 h
at 0 °C and extractive workup gave 196 mg of a light-yellow oil.
Purification by rapid flash chromatography on silica gel afforded
178.1 mg (98%) of pure deuterated chloride 6: 1H NMR (400 MHz,
CDCl3) δ 5.09 (t, J ) 6.8 Hz, 2H), 4.10 (s, 2H), 1.93-2.16 (m,
8H), 1.73 (s, 3H), 1.68 (s, 3H), 1.60 (s, 6H); 13C NMR (100 MHz,
Incubation of (E,E)- [2-2H1]Farnesyl Diphosphate with
Recombinant Patchouli Alcohol Synthase. A typical large scale
incubation experiment (500 mL) was set up as follows: [2-2H1]FPP
(10 mg/mL) in 2.5 mL of buffer solution (250 mM Tris-HCl, pH
7.4; 50 mM MgCl2) was diluted with 50 mL of buffer solution,
purified PTS synthase (ca. 1 µg/µL, 20 mL), and distilled water
9
J. AM. CHEM. SOC. VOL. 132, NO. 9, 2010 3007

A R T I C L E S
Faraldos et al.
(427 mL). The reaction mixture was incubated in a 30 °C water
bath overnight. Two such incubations were combined (1 L) and
extracted with 15% EtOAc-hexane (2 × 1 L). The volume of the
organic extracts was reduced to 20 mL by evaporation under
reduced pressure. The crude extracts were shipped to the University
of Illinois.
Isolation of Sesquiterpenes. The organic extract (20 mL) was
evaporated using a stream of N2, and the residue (14.1 mg) was
loaded into a preparative TLC (silica) plate. The plate was first
developedwithHPLC-gradepentaneandthenwith5%EtOAc-hexane
to afford, in order of elution, a m/z 204 sesquiterpene hydrocarbon
fraction (3.2 mg, 13.6%) and three m/z 222 sesquiterpene alcohol
fractions (Alcohol Fractions 1 (3.2 mg, 12.5%), 2 (0.8 mg, 3%),
and 3 (1.2 mg, 4.7%) according to GC-MS analysis. Initial 1H NMR
(500 MHz, CDCl3) experiments revealed that Alcohol Fractions 1,
2, and 3 were composed of pure (>95%) patchoulols (2), [2-2H1]-
nerolidol, and (E,E)-[2-2H1]-farnesol (5), respectively.
δ-Patchoulene (9). Partial 1H NMR (500 MHz, CDCl3) data from
fraction 1: δ 5.00 (broad s, 1H), 1.00 (s, 3H), 0.98 (d, J ) 6.5 Hz,
3H), 0.89 (s, 3H), 0.78 (s, 3H). LR(FI)MS m/z (rel. int.) 208 (4),
207 (28), 206 (73), 205 (100).
r-Guaiene (10). Partial 1H NMR (500 MHz, CDCl3) data from
fraction 1: δ 4.68 (broad s, 1H), 4.62 (broad s, 1H), 1.72 (s, 3H),
1.00 (d, J ) 6.9 Hz, 3H), 0.93 (d, J ) 6.8 Hz, 3H). LR(FI)MS m/z
(rel. int.) 208 (2), 207 (10), 206 (30), 205 (67), 204 (100). The 1H
NMR data are in agreement with those reported in ref 23a at 200
MHz.
Seychellene (11). Partial 1H NMR (500 MHz, CDCl3) data from
fraction 1: δ 4.79 (d, J ) 1.5 Hz, 1H), 4.58 (d, J ) 1.5 Hz, 1H),
0.96 (s, 3H), 0.83 (s, 3H), 0.74 (d, J ) 6.9 Hz. 3H). LR(FI)MS
m/z (rel. int.) 208 (6.5), 207 (29), 206 (64), 205 (100), 204 (1). 1H
NMR data are in agreement with those reported in ref 23a at 200
MHz.
1
r-Patchoulene (12). Partial H NMR (500 MHz, CDCl3) data
Patchoulol (2, as 2:1 Mixture of [5-2H1]- and
[5,12-2H2]-Isotopomers). white solid; 1H NMR (600 MHz, CDCl3)
See Table 1 and Figure 1; 1H NMR (500 MHz, C6D6) δ 1.97-1.90
(m, 1H), 1.88 (broad d, J ) 12.5 Hz, 1H), 1.85 (m, 1H), 1.59 (dd,
J ) 12.5, 4.2 Hz, 1H), 1.40-1.11 (m, 7H), 1.10 (s, 3H), 1.07 (broad
s, 1H), 1.00 (s, 3H), 0.97 (d, J ) 13.5 Hz, 1H), 0.80 (s, 3H), 0.75
(dd, J ) 6.5, 0.8 Hz, 3H, 12-CH3), 0.73 (d, J ) 6.5 Hz, 2H, 12-
CH2D); 13C NMR (125 MHz, CDCl3) δ 75.62 (C1), 43.12 (1:1:1 t,
J ) 20 Hz, C5), 40.10 (C11), 39.08 (C7), 37.55 (C10), 32.66 (C2),
32.57 (C2, minor), 28.80 (C9), 28.57 (C3), 27.99 (C4), 27.91 (C4,
minor), 26.83 (C14), 24.45 (C8), 24.30 (C6), 24.28 (C15), 20.60
from fraction 1: δ 5.07 (broad s, 1H), 1.62 (s, 3H), 0.95 (s, 3H),
0.90 (d, J ) 6.9 Hz, 3H), 0.88 (s, 3H). LR(FI)MS m/z (rel. int.)
1
208 (4.7), 207 (27), 206 (63), 205 (100), 204 (1). The H NMR
data are in agreement with those reported in ref 23a at 200 MHz.
Guaia-4,11-diene (13, Aciphyllene). Partial 1H NMR (500 MHz,
CDCl3) data from fraction 2: δ 4.67 (broad s, 1H), 4.59 (broad s,
1H), 1.75 (broad s, 3H), 1.57 (s, 3H), 0.75 (d, J ) 7.1 Hz, 3H).
LR(FI)MS m/z (rel. int.) 208 (1), 207 (8), 206 (49), 205 (100), 204
(1). The 1H NMR data are in agreement with those reported in ref
23d at 400 MHz.
r-Bulnesene (14). Partial 1H NMR (500 MHz, CDCl3) data from
fraction 2: δ 4.65 (s, 1H), 4.63 (s, 1H), 1.70 (s, 3H), 1.66 (s, 3H),
0.90 (d, J ) 7.0 Hz, 3H). LR(FI)MS m/z (rel. int.) 208 (11), 207
(27), 206 (66), 205 (100), 204 (1). The 1H NMR data are in
agreement with those reported in ref 23a at 200 MHz.
Unidentified. LR (FI)MS m/z (rel. int.) 208 (5), 207 (25), 206
(71), 205 (100).
2
(C13), 18.52 (C12), 18.05 (1:1:1 t, J ) 21.2 Hz, C12 minor). H
NMR (77 MHz, CHCl3) δ 1.442 (broad s, D5), 0.803 (12-CH2D);
LR(FI)MS m/z (rel. int.) 228 (1.4), 227 (6), 226 (17), 225 (34),
224 (68), 223 (100) 222 (2).
Further purification of the m/z 204 hydrocarbon fraction by
preparative TLC on silica gel using hexane as developing solvent
gave two fractions: fraction 1 (ca. 1.5 mg) was composed of an
approximate 10:8:5:1 mixture of sesquiterpenes 11, 10, 12 and 9,
Acknowledgment. We thank the National Institutes of Health
for financial support of this work through grants GM13956 (R.M.C.)
and GM54029 (J.C.) and the National Science Foundation for
providing support for the VOICE NMR facility at the University
of Illinois at Urbana-Champaign. We are grateful to Steve Mullen
of the School of Chemical Sciences Mass Spectrometry Laboratory
at the UIUC for the GC-MS analyses and to Brandon Miles for an
initial preparation of the [2-2H1]FPP substrate.
1
respectively, according to H NMR analysis (500 MHz, CDCl3).
In the same manner, 1H NMR analysis (500 MHz, CDCl3) of
fraction 2 (ca. 1.5 mg) revealed that this fraction was composed of
(in addition to R-guaiene (10, 26%), seychellene (11, 9%) and an
unidentified hydrocarbon (9%)) R-bulnesene (14, 26%), guaia-4,11-
diene (13, 18%), caryophyllene (8, 7%) and ꢀ-patchoulene (7, ca.
4%).
ꢀ-Patchoulene (7). Partial 1H NMR (500 MHz, CDCl3) data from
fraction 2: δ 0.93 (d, J ) 6.7 Hz, 3H), 0.89 (s, 6H), 0.88 (s, 3H).
LR(FI)MS m/z (rel. int.) 208 (2), 207 (15), 206 (30), 205 (55), 204
(100). The 1H NMR data are in agreement with those reported for
synthetic 7 at 360 MHz.23b
Supporting Information Available: General experimental and
instrumentation, detailed preparative procedures and character-
ization data, expression and purification of recombinant patchou-
ꢀ-Caryophyllene (8). Partial 1H NMR (500 MHz, CDCl3) data
from fraction 2: δ 5.31 (broad s, 1H), 4.94 (broad s, 1H), 4.82
(broad s, 1H), 1.61 (s, 3H), 0.99 (s, 3H), 0.97 (s, 3H). LR(FI)MS
m/z (rel. int.) 208 (8), 207 (22), 206 (6), 225 (100), 204 (5.6). The
1H NMR data are in agreement with those reported in ref 23c at 90
MHz.
1
2
li alcohol synthase, and reproductions of H, H, 13C, and 31P
NMR spectra. This material is available free of charge via the
Internet at http://pubs.acs.org.
JA909251R
9
3008 J. AM. CHEM. SOC. VOL. 132, NO. 9, 2010