On the early steps of cineol biosynthesis in Eucalyptus globulus

Article abstract of DOI:10.1002/1522-2675(20000906)83:9<2504::AID-HLCA2504>3.0.CO;2-Z

Samples of [4-²H₁]-1-deoxyxylutose (17a) and [2-¹³C, 4-²H₁]-1-deoxyxylulose (17b), have been prepared by modification of known procedures and fed in aqueous solution to twiglets of Eucalyptus globulus

Full text of DOI:10.1002/1522-2675(20000906)83:9<2504::AID-HLCA2504>3.0.CO;2-Z

2504
Helvetica Chimica Acta ± Vol. 83 (2000)
On the Early Steps of Cineol Biosynthesis in Eucalyptus globulus
by Christoph Rieder, Bernhard Jaun, and Duilio Arigoni*
Laboratorium für Organische Chemie der Eidgenössischen Technischen Hochschule Zürich,
Universitätstrasse 15, ETH-Zentrum, CH-8092 Zürich
Dedicated, on the occasion of his 75th birthday, to Albert Eschenmoser, from whom I learned, among many
other things, that one plus one gives much more than two, if the ones communicate.
Samples of [4-2H1]-1-deoxyxylulose (17a) and [2-13C, 4-2H1]-1-deoxyxylulose (17b), have been prepared by
modification of known procedures and fed in aqueous solution to twiglets of Eucalyptus globulus. The probes of
cineol (6) isolated from these experiments were analyzed by GC/MS, 2H- and 13C-NMR techniques. In the
experiments with 17b, the formation of five isotopomers of 6 could be detected. Their structure and relative
abundance demonstrate that the 13C-label is incorporated to the same extent into the two C5-units of 6, and that
2
the H label is retained to an extent of 57% in the starter dimethylallyl-diphosphate unit (DMAPP; 12), but
completely or almost completely lost in the unit derived from isopentenyl diphosphate (IPP; 11), in the
elongation step which leads to geranyl diphosphate (GPP; 1). These results confirm that the recently discovered
mevalonate-independent pathway to IPP and DMAPP is operative in the biosynthesis of cineol, and indicate,
together with previous finding, that, within this pathway, formation of IPP and DMAPP occurs in independent
rather than in sequential steps. In addition, the demonstration of different metabolic origins for the olefinic H-
atoms of GPP (1), the aliphatic C10-precursor of 6, paves the way for a realistic interpretation of the strikingly
consistent but hitherto unexplained anomalies detected in the natural-abundance 2H-NMR spectra of (‡)- and
(À)-a-pinene and of (‡)-limonene.
Introduction. ± Extensive studies carried out by different groups in the last 25 years
have established that the biosynthesis of a large variety of cyclic monoterpenes
proceeds from the universal aliphatic precursor geranyldiphosphate (GPP; 1),
according to the mechanistic route outlined in Scheme 11). Anionotropic rearrange-
ment of 1, followed by rotation about the newly created single bond, leads to the boat-
like conformation of (3R)-linaloyl diphosphate (2). Subsequent ionization of 2 with
concomital participation of the distal double bond generates the (4S)-a-terpinyl cation
3, which can dismiss its charge by proton extrusion to provide (‡)-limonene (4),
capture a molecule of H2O to give (‡)-a-terpineol (5), or suffer an additional
cyclization to generate 7, the immediate precursor of (‡)-a-pinene (9), and (À)-b-
pinene (8). An equivalent set of reactions proceeding through the (3S)-enantiomer of
2 leads to the mirror images of all chiral cationic intermediates and substances
indicated in Scheme 1.
Some years ago, three groups reported independently on natural-abundance
2H-NMR as a tool for investigating the biosynthesis of cyclic monoterpenes [2 ± 4]. In
the paper by Martin et al. [2], it was pointed out specifically that the 2H content at the
position labelled with HA in the formula of (‡)-a-pinene (9) was markedly depleted
1
)
For a recent detailed review, see [1].

Helvetica Chimica Acta ± Vol. 83 (2000)
2505
Scheme 1
(by a factor of 4) with respect to the value corresponding to HB in the same compound,
and that a similar depletion was also evident in the 2H-NMR spectrum of (À)-a-pinene.
A critical examination of the data published by Pascal et al. on (À)-a-pinene [3] has
2
confirmed this anomalous H distribution [5]. In addition, inspection of the natural-
abundance 2H-NMR of (‡)-limonene (4) has revealed once again a significant
depletion (by a factor of 2.5) in the 2H content of the HA-atom of 4 [5]. No explanation
was offered in any of the three reports for this consistent anomalies. In subsequent work
triggered by these striking findings, it could be shown that, with a set of cyclases from
Salvia officinalis, the HA-atom of the aliphatic precursor 1 is completely retained during
formation of a variety of cyclic monoterpenes including inter alia 4, 8, and their
enantiomers, as well as the achiral cineol (6), and it was concluded that the anomalies
previously detected had to be connected with a pre-existing differential labelling of the
two olefinic H-atoms HA and HB in the aliphatic precursor 1 [5].
The notion of an aberrant alteration in the natural 2H content at the C(2)-atom of 1
rests, of course, on the preconceived and ill-supported conviction that 1 is assembled as
illustrated in Scheme 2 from the two building blocks isopentenyl diphosphate (IPP; 11)
and dimethylallyl diphosphate (DMAPP; 12), generated sequentially from the
ꢀuniversalꢁ terpene precursor mevalonic acid (10). Since in this scheme the same H-
atom of IPP (11) is lost in the formation of DMAPP (12) and in the elongation step
which leads to GPP (1) [6], the two olefinic H-atoms of the latter share a common

2506
Helvetica Chimica Acta ± Vol. 83 (2000)
Scheme 2
source, the dotted H at C(4) of the mevalonic-acid precursor 10, and are, accordingly,
2
expected to display identical values for their levels of natural H abundance.
For a long time, experimental evidence in support of a mevalonate route for the
biosynthesis of monoterpenes in higher plants has been at best scanty and altogether
less than convincing [7]. A dramatic turn of events has been brought about in recent
years by the discovery of a new mevalonate-independent metabolic pathway to IPP
(11) and DMAPP (12), which by now is known to be widespread in eubacteria (both
Gram-positive and Gram-negative), green and red algae, as well as higher plants [8].
An updated version covering the latest developments [9] in our still incomplete
knowledge of the new pathway is presented in Scheme 3. The specific demonstration
that the new pathway plays a major, if not exclusive, role in the formation of the two C5-
units required for the biosynthesis of monoterpenes in higher plants [10] provided a
welcome platform for tackling anew the problem of the suspected difference in the
Scheme 3

Helvetica Chimica Acta ± Vol. 83 (2000)
2507
metabolic origin of the olefinic HA- and HB-atoms in the GPP intermediate 1. For
practical reasons, we decided to concentrate our efforts on the formation of cineol (6),
the preponderant monoterpene in leaves of Eucalyptus globulus.
In a set of incisive studies carried out with a specific synthetase from Salvia
officinalis, Croteau et al. succeeded in showing that the structure of 6 is elaborated from
1 via the intermediacy of 2 and 5 (cf. Scheme 1), and that the ether O-atom as well as
the proton that triggers off the last cyclization step are derived from molecules of H2O
[11]. To date, no evidence is available on the cryptic stereochemical course of the
sequence 3 ! 5 ! 6.
Results and Discussion. ± The feasibility of the selected approach was first tested in
exploratory experiments with a specimen of 1-deoxyxylulose labelled with deuterium
at C(4) (cf. 17a in Scheme 4), and the preliminary results were then refined in a second
set of experiments involving a probe of the doubly labelled [2-13C,4-2H1]-1-deoxyxy-
lulose (17b).
Scheme 4
The synthesis of 17a was achieved according to known procedures [12][13],
involving a Wittig reaction of the deuterated aldehyde 13 with 14 to provide the (E)-
enone 15, followed by asymmetric Sharpless dihydroxylation of the latter to 16 and
hydrogenolytic removal of the protecting Bn group. The required deuterated aldehyde
13 was conveniently prepared by Swern oxidation of the [1-2H2]alcohol, which is easily

2508
Helvetica Chimica Acta ± Vol. 83 (2000)
accessible from commercially available 2-(benzyloxy)acetyl chloride in two steps
involving methanolysis and subsequent reduction of the resulting methyl ester with
LiAl(2H4). The 2H content of all the intermediates and of the final compound 17a was
1
found by H-NMR analysis to be larger than 98%.
The same deuterated aldehyde 13 could be used for the synthesis of the doubly
labelled precursor 17b (ꢀ ˆ 13C). Condensation of 13 with the Wittig reagent prepared
with Ph3P from ethyl 2-bromo[1-13C]acetate (18) gave a 3 :1mixture of the ( E)-ester 20
and of its (Z)-counterpart. After chromatographic separation of the two isomers, the
pure (E)-compound was converted by a Tebbe olefination and subsequent hydrolysis of
the resulting enol ether with AcOH in toluene to the (E)-enone 22, and the synthesis of
the desired material 17b was then completed as described before. Once again, 1H-NMR
monitoring of the doubly labelled compounds ensured that the deuterium content was
>98%.
With these tools in hand, we could now proceed with the critical biological
experiments. For this purpose, twiglets of E. globulus were allowed to absorb an
aqueous solution of the labelled substrates over a period of 5 days under normal
illumination conditions. The volatile fraction from the leaves was isolated by distillation
and found by GC/MS analysis to consist of essentially pure cineol (6).
The mass spectrum of the cineol specimen from the feeding experiment with 17a
‡
indicated a slight increase of ca. 1% in the intensity of the [M ‡ 1] ion. The 2H-NMR
spectrum of this probe displayed a single enriched signal at 1.38 ppm, corresponding to
2
the H-atom at C(4) of cineol (for the numbering of cineol, see Fig. 2). Using the
background signals of molecules with natural abundance (more specifically, the signal
at 1.20 ppm, corresponding to the two enantiotopic Me groups) as a calibration, a 61-
fold intensity increase was estimated for the enriched position, corresponding to a
2
specific H incorporation of ca. 0.9%.
2
To clarify the reason for the lack of H incorporation in the second isopentane unit
of cineol, a second feeding experiment was carried out with the doubly labelled
substrate 17b, in which the 13C-label was expected to act as an internal standard for
following the fate of the 2H-atom. The mass spectrum of the resulting cineol suggested
the presence of a small amount of molecules (<1%) with up to three additional mass
units, and once again the 2H-NMR spectrum of the probe displayed a single
recognizably enriched signal at 1.38 ppm with a relative intensity corresponding to a
specific incorporation of ca. 0.45%.
The 13C-NMR spectrum of the same material proved to be rewardingly rich in
information. A comparison of the relative intensities of its lines with those of a sample
with natural 13C-abundance isolated from the same source (see Table 1) revealed an
identical (or nearly identical) increase of ca. 57% in the 13C-content of both C(1) and
C(8). This corresponds to a specific incorporation of 0.61% at both sites. An expanded
version of the signals corresponding to these two C-atoms is presented in Fig. 1 and
uncovers the presence of different satellites. From the relevant data collected in
Table 2, the presence and relative abundance of the five isotopomers of cineol (6)
depicted in Fig. 2 can be inferred. Capital letters are used in Table 2 and Fig. 2 to
correlate each isotopomer with the signals for which it is responsible; note that A and B
refer to molecules generated de novo from the proffered precursor, whereas A0 and B0
correspond to molecules with normal background abundance. The relative intensity of

Helvetica Chimica Acta ± Vol. 83 (2000)
2509
Awas estimated by correcting the value of C with the ratio of the values of E and D, and
the value of B was approximated by the sum of the values for A and E. The positive
value of the d shift caused in isotopomer D by the presence of a 2H-atom at C(4) on the
13C-signal of C(1) may, at first sight, seem unusual but is in fact well-precedented by a
similar value of 13 ppb detected in a bicyclo[2.2.2]octane derivative [15].
Table 1. Relative Intensities of the 13C-NMR Signals in Samples of Cineol from E. globulus (A: reference sample
with natural abundance; B: sample from the experiment with 17b)
Positiona)
d Valuesb)
(CD2Cl2)
Relative intensitiesc)
A
B
C(8)
C(1)
C(4)
C(2/6)
C(9/10)
C(7)
73.76
69.92
33.35
31.91
29.09
27.71
23.22
23.5
23.5
100
217
223
114
217
37
37
100
231
238
116
222
C(3/5)
a) Numbering as in Fig. 2. b) For signal assignment, see [14]. c) Normalized for the intensity of the C(4) signal.
Table 2. Selected Data for the 13C-NMR Signals (CD2Cl2) of C(1) and C(8) in the Cineol Sample from the
Feeding Experiment with 17b
Position
Isotopomera)
d Value
[ppm]
Relative
intensities
J(C,C)
[Hz]
Isotopic shift
[ppb]
C(8)
A0
A
C
E
D
73.766
73.766
73.763
73.716
73.712
128
5.5
19.5
8
1.74
b-(CÀC) ˆ À3.38
30
1.74
1.76
b-(DÀC) ˆ À51.2
d-(DÀC) ˆ ‡11.6
C(1)
D
B0
B
69.934
69.923
69.923
69.92123
30
124
13
C
.76
1
b-(CÀC) ˆ À2.6
a) Cf. Fig. 2.
Among the de novo synthesized labelled material, isotopomers C and D doubly
labelled with 13C have a preponderance of 67%; this shows that the starting material,
which was essentially isotopically pure, was diluted only to an extent of ca. 20% by
endogenous material, and, accordingly, the relatively low value of the specific
incorporation is caused mainly by a large pool of cineol already present at the
beginning of the feeding experiment.
A comparison of the relative abundancies of species A, C, D, and E reveals that
2
57% of all the molecules labelled with 13C at C(8) have retained a H companion at
C(4); in sharp contrast, all (or nearly so; cf. below) of the 2H originally associated in the
precursor with the 13C-label eventually detected at C(1) of cineol has been lost during
formation of the monoterpene. This is tantamount to the statement that the 2H label of
the precursor 17b is largely retained in the C5-unit of cineol derived from DMAPP (12)

2510
Helvetica Chimica Acta ± Vol. 83 (2000)
Fig. 1. Expansions of the 13C-NMR spectrum of the cineol sample from the feeding experiment with 17b showing
the signals of C(8) (left) and C(1) (right). Conditions: 125 MHz, {1H}-broad-band-decoupled, inverse-gated;
processed with a gaussian filter function.
Fig. 2. Cineol numbering and structure of the isotopomers responsible for the 13C-NMR signals displayed in Fig. 1
(ꢀ ˆ 13C)

Helvetica Chimica Acta ± Vol. 83 (2000)
2511
and completely or almost completely lost in the C5-unit of cineol derived from IPP (11)
in the elongation process. Since it is firmly established that HA of GGP (1) is fully
retained during cineol biosynthesis [5], it follows that, as suspected, the HA- and HB-
atoms of the GGP (1) precursor of cineol (6) have different metabolic origins. This, in
turn, provides a long-sought new starting point for a realistic interpretation of the
unexplained spectroscopic results discussed in detail in the Introduction.
The findings reported here match closely the ones previously reported for the
mevalonate-independent formation of the octaprenyl side chain of ubiquinone in
Escherichia coli [13]. In the meantime, work carried out in the Poulter group [16] has
demonstrated that in E. coli, as in all eukariotes, the same H-atom is eliminated from
C(2) of IPP (11) in the isomerization and elongation steps depicted in Scheme 5. From
the bulk of the evidence, one is forced to conclude that in the new pathway IPP (11)
cannot be a committed precursor of DMAPP (12), and that the two compounds must
be formed independently from a late and yet unidentified intermediate X. Results
recently secured in Rohmerꢁs group for the biosynthesis of bacteriohopanes in
Zymomonas mobilis suggest that the H-atom of IPP (11) retained in the main
elongation step of Scheme 5 has been introduced from NADPH in one of the late and
still unknown steps of the pathway [17]. The operation of a similar step during
monoterpene biosynthesis in higher plants would be eminently well-suited for efficient
isotopic discrimination (both in the formation of NADPH and in the subsequent
hydride transfer) and might well represent the immediate cause for the marked
depletion in natural 2H consistently observed at the HA-position of monocyclic terpenes
[2 ± 4].
Scheme 5

2512
Helvetica Chimica Acta ± Vol. 83 (2000)
2
A final comment is required for the apparent loss of 43% of the H label of 17b
detected in this work during formation of the DMAPP unit involved in cineol
biosynthesis. Lack of information concerning the nature of the last two reductive steps
leading to DMAPP does not allow to rule out a true partial loss associated with the
mechanistic details of such steps. Alternatively, one can assume that the 2H label of 17b
is fully retained in DMAPP molecules generated without the intermediacy of IPP and
that 2H-free DMAPP molecules stem from the subsequent isomerization of independ-
ently formed IPP labelled as indicated in Scheme 5. One might, of course, object that by
the same token action of the isomerase on the originally formed DMAPP should
generate another form of IPP, which, in contrast to the original one is expected to retain
its 2H in the elongation step; it is, however, important to note that in the early stages of
the (incomplete) isomerization process the relative contribution of such a species to the
pool of labelled IPP molecules will be critically dependent on the relative size of the
pools of IPP and DMAPP; in other words, if the size of the original IPP pool is much
larger than the original size of the DMAPP pool, the relative contribution of IPP
molecules with a different location of the label emerging from the isomerization
process will be mitigated accordingly. In this context, it may be relevant to point out
that a closer look at the traces of Fig. 1 uncovers the presence of what looks like a set of
two low-intensity doublets centered at ca. 69.865 and 73.764 ppm and displaying an
apparent J value of ca. 1.75 Hz. While the significance of these signals remains
questionable, it is interesting to note that, both in terms of their spacing and of their
estimated shifts from the main peaks, they would be compatible with the presence of
minor amounts (less than 10% of C) of yet another isotopomer of cineol, generated
from IPP derived from the isomerization of DMAPP as indicated in Scheme 5 and
2
labelled with 13C at both C(1) and C(8) and with H at C(6).
We are indebted to Prof. P. Chen for allowing us to use his GC/MS facilities, to Mrs. P. Enz and R. Stalder of
the Zürich botanical garden for providing the biological material, and Prof. H.-J. Borschberg for technical
assistance in the preparation of the manuscript. Financial support by Novartis AG, Basel, is gratefully
acknowledged.
Experimental Part
General. GC/MS Measurements were carried out with a Fisons MD 800 instrument equiped with a DB-
5MS capillary columm. 2H-NMR: Bruker AMX-400 (61.42 MHz); in CH2Cl2 with a CFCl3 lock. 13C-NMR:
Bruker AMX-400 (100 MHz) and Bruker DRX-500 (125 MHz), in CD2Cl2.
Feeding Experiments and Isolation of Cineol (6). a) The stems of two 20-cm long twiglets of E. globulus,
carrying ca. 20 leaves each, were cut under running water with a sharp razor blade at an angle of ca. 308 and
placed in a narrow 50-ml beaker containing 20 ml of H2O. The labelled precursor 17a (400 mg) was dissolved in a
mixture of 8 ml of H2O and 4 ml of EtOH; approximately half of this soln. was administered to the twiglets at
the beginning of the experiment, and the rest was added in aliquots at regular intervals together with enough
H2O to make up for the amount lost through evaporation. To facilitate absorption, the immersed tips of the
twiglets were cut anew twice a day under running water. The experiment was stopped on the fifth day after
150 ml of H2O had been absorbed, and the detached leaves powered under liq. N2. The powder (6.30 g)
was suspended in 50 ml of H2O, and the volatile components were distilled at r.t./1Torr and trapped in a flask
cooled with liquid N2. Following addition of H2O to the residue, the distillation was repeated three times, and
the collected distillates were extracted with CH2Cl2 (4 Â 100 ml). The org. phase was dried (MgSO4) and the
solvent removed at 408/300 Torr to give 92 mg of an oil, shown by GC/MS analysis to consist essentially of pure
cineol.
b) From a similar experiment with 390 mg of 17b and a total absorption of 235 ml of H2O, a sample of
427 mg of cineol was isolated.

Helvetica Chimica Acta ± Vol. 83 (2000)
2513
REFERENCES
[1] M. L. Wise, R. Croteau in ꢀComprehensive Natural Products Chemistryꢁ, Ed. D. E. Cane, Elsevier, Oxford,
UK, 1999, Vol. 2, p. 97.
[2] G. J. Martin, P. Janvier, S. Akoka, F. Mabon, J. Jurczak, Tetrahedron Lett. 1986, 27, 2855.
[3] R. A. Pascal, Jr., M. W. Baum, C. K. Wagner, R. L. Rodgers, D. S. Huang, J. Am. Chem. Soc. 1986, 108,
6477.
[4] M. F. Leopold, W. W. Epstein, D. M. Grant, J. Am. Chem. Soc. 1988, 110, 616.
[5] D. Arigoni, D. E. Cane, J. H. Shim, R. Croteau, K. Wagschal, Phytochemistry 1993, 32, 623.
[6] C. D. Poulter, H. L. Rilling in ꢀBiosynthesis of Isoprenoid Compoundsꢁ, Ed. J. W. Porter, S. L. Spurgeon,
John Wiley & Sons, N. Y., 1981, Vol. 1, p. 161.
[7] D. V. Banthorpe, B. V. Charleswood, M. J. D. Francis, Chem. Rev. 1972, 72, 115.
[8] a) M. Rohmer, Nat. Prod. Rep. 1999, 16, 565; b) M. Rohmer in ꢀComprehensive Natural Products
Chemistryꢁ, Ed. D. E. Cane, Elsevier, Oxford, UK, 1999, Vol. 2, p. 45; c) D. Arigoni, M. K. Schwarz, in
ꢀComprehensive Natural Products Chemistryꢁ, Ed. D. E. Cane, Elsevier, Oxford, UK, 1999, Vol. 2, p. 367;
d) W. Eisenreich, M. Schwarz, D. Cartayrade, D. Arigoni, M. H. Zenk, A. Bacher, Chem. Biol. 1998, 5,
R 221.
[9] a) S. Herz, J. Wungsintaweekul, C. A. Schur, S. Hecht, H. Lüttgen, S. Sagner, M. Fellermeier, W.
Eisenreich, M. H. Zenk, A. Bacher, F. Rohdich, Proc. Natl. Acad. Sci. U.S.A. 2000, 97, 2486; b) M. Tagagi,
T. Kuzuyama, K. Kaneda, H. Watanabe, T. Dairi, H. Seto, Tetrahedron Lett. 2000, 41, 3395.
[10] a) W. Eisenreich, S. Sagner, M. H. Zenk, A. Bacher, Tetrahedron Lett. 1997, 38, 3889; S. Sagner, C. Latzel,
W. Eisenreich, A. Bacher, M. H. Zenk, Chem. Commun. 1998, 221; c) J. Piel, J. Donath, K. Bandemer, W.
Boland, Angew. Chem., Int. Ed. 1998, 37, 2478; d) D. J. Fowler, J. T. C. Hamilton, A. J. Humphrey, D. O.
Hagan, Tetrahedron Lett. 1999, 40, 3803.
[11] R. Croteau, R. Alonso, F. Koepp, M. J. Johnson, Arch. Biochem. Biophys. 1994, 309, 184.
[12] J. L. Giner, Tetrahedron Lett. 1998, 39, 2479.
[13] J. L. Giner, B. Jaun, D. Arigoni, Chem. Commun. 1998, 1857.
[14] a) R. M. Carman, A. C. Garner, K. D. Klika, Austr. J. Chem. 1994, 47, 1509; b) Y. Asakawa, M. Matsuda, M.
Tori, T. Hashimoto, Phytochemistry 1988, 27, 3861.
[15] W. Adcock, A. N. Abeywickrema, V. S. Iyer, G. B. Cock, Magn. Res. Chem. 1986, 24, 213.
[16] a) A. E. Leyes, J. A. Bacher, F. M. Hahn, C. D. Poulter, Chem. Commun. 1999, 717; b) A. E. Leyes, J. A.
Bacher, C. D. Poulter, Org. Lett. 1999, 1, 1071.
[17] L. Charon, C. Pale-Grosdemange, M. Rohmer, Tetrahedron Lett. 1999, 40, 7231.
Received August 4, 2000