Stereochemistry of formation of the β-ring of lycopene: Biosynthesis of (1R,1′R)-β,β-[16,16,16,16′,16′,16′- 2H6]carotene from [16,16,16,16′,16′,16′-2H6] lycopene in flavobacterium R 1560

Article abstract of DOI:10.1002/1522-2675(20000809)83:8<2036::AID-HLCA2036>3.0.CO;2-5

Incubation of deuteriated precursors in cultures of Flavobacterium produced specifically deuteriated carotenoids. Analysis of these led to several conclusions: i) Lycopene is a direct precursor of β,β-carotene. ii) Its terminal Me groups retain their inte

Full text of DOI:10.1002/1522-2675(20000809)83:8<2036::AID-HLCA2036>3.0.CO;2-5

2036
Helvetica Chimica Acta ± Vol. 83 (2000)
Stereochemistry of Formation of the b-Ring of Lycopene: Biosynthesis
of (1R,1'R)-b,b-[16,16,16,16',16',16'-2H6]Carotene from
[16,16,16,16',16',16'-2H6]Lycopene in Flavobacterium R 1560
by Sasank Sekhar Mohantya)1), Peter Uebelhartb), and Conrad Hans Eugster*b)
a) Novartis Agro AG, CH-4002 Basel,
b) Organisch-chemisches Institut der Universität Zürich, Winterthurerstrasse 190, CH-8057 Zürich
(e-mail Conrad.Eugster@access.unizh)
Dedicated to Prof. Albert Eschenmoser on the occasion of his 75th birthday
Incubation of deuteriated precursors in cultures of Flavobacterium produced specifically deuteriated
carotenoids. Analysis of these led to several conclusions: i) Lycopene is a direct precursor of b,b-carotene. ii) Its
terminal Me groups retain their integrity during cyclization: there is a stereospecific folding of the 1,5-diene. The
Me(16,16') groups of lycopene become the Me(16,16') of b,b-carotene. Consequently, the folding must follow
the C2(E,E) mode. iii) Incorporation of deuterium was sufficiently extensive to permit CD measurements on the
isolated b,b-carotene, allowing its centers of chirality to be assigned as (1S,1'S). iv) The same chirality resulted
from incorporation of [2H3]mevalonate into zeaxanthin. The syntheses of specifically deuteriated [2H3]GPP,
[2H3]FPP, and [2H3]GG are described.
1. Introduction. ± Only those carotenoids with cyclic end-groups, such as b,b-
carotene, are capable of performing this groupꢀs vital functions as supplementary light
absorbers and deactivators of electronically excited states of chlorophyll in higher
plants. As well as this, only those carotenoids with b-rings also function as precursors of
retinoids such as, e.g., vitamin A.
Using different experimental approaches, many groups have been able to show
recently that lycopene (1) is the direct precursor in the biological cyclization giving rise
to b,b-carotene2). This work offers further confirmation of that finding, and, in
addition, offers innovations in methodology.
It also represents the final conclusions of a long-running project in our group, with
important preliminary work done by Märki et al. [3] and Hofer et al. [3d][4], and
confirms findings by Britton et al. [5] that establish the stereochemistry of the ring
closure to zeaxanthin (2; Scheme 1). Our results, obtained with novel methodology, are
also significant in that Brittonꢀs findings had conflicted with some reported by BuꢀLock
and co-workers [6] (cf. Sect. 5).
2. Key Reactions. ± The controlled degradation of abietic acid (3a) and its
deuteriated form 3b to precursors of (1R,1'R)-b,b-[16,16,16,16',16',16'-2H6]carotene (4)
1
)
From the Ph.D. thesis of S.S.M. [1]. For external reasons, publication in this journal has not been possible
until now. Some fragmentary findings were presented in brief in Chapt. 9 of [2]. This work presents the
complete results of the work in Zürich and sets it in its overall context.
2
)
Authoritative review by Britton in [2].

Helvetica Chimica Acta ± Vol. 83 (2000)
2037
Scheme 1
was described in [3]. In this work, labeling has also been used to show that
the Me groups at C(1) arising from this, indistinguishable by 1H- and
13C-NMR spectroscopy, produce in the conformational equilibrium an isotope effect
sufficiently large as to make it possible to draw reliable conclusions about the chirality
of substitution at C(1) and/or C(1') from low-temperature CD measurements.

2038
Helvetica Chimica Acta ± Vol. 83 (2000)
Subsequently, we prepared the specifically deuteriated (E)-[2H6]geraniol 5c and
lycopene 63).
It is unnecessary to stress that, as reference substances for carotenes and
carotenoids of biological origin, compound 4 and other isomers not mentioned here
[3] were from the outset characterized sufficiently thoroughly that definitive
comparisons could be carried out even with the smallest quantities.
Using compound 6, we carried out numerous incorporation experiments; these,
however, did not provide any usable results. Among them, we should mention the
wicking technique in yellow roses4); injections into ripening rose hips, fruiting bodies of
Inocybe species, and fruits of Taxus baccata; incubations with acetone powders from
tomatoes [8]; cell-free enzyme preparations from spinach chloroplasts [8] and those
from ꢁfrench beansꢀ (Phaseolus vulgaris according to [9], Phycomyces vulgaris
according to [10]), etc. However, in each case5), we were unsuccessful in identifying
any incorporation (detectable by HPLC, MS, FTIR, or CD) of 4 or its enantiomer. On
Brittonꢀs advice, we finally succeeded with an incorporation of 6 in cultures of
Flavobacterium mutante R 15606) (Sect. 3.1 and 3.2).
3. Results.
± 3.1. Biosynthesis of (1S,1'S,3R,3'R)-[17,17',18,18',19,19',20,20'-
2H24]Zeaxanthin. ± For our labeled precursor, we used (Æ)-[3',3',3'-2H6]mevalonolac-
tone (14) or [2H3]mevalonic acid ([2H3]MVA), following the procedure of Chakraborty
et al. [11], with the modifications necessary for labeling (Scheme 2).
The bacterial cell mass (10000 g) obtained from centrifugation of the cultured
aforementioned Flavobacterium was incubated in 0.1m Tris ´ HCl buffer, pH 7.0, with
lysozyme, MnSO4, MgSO4, adenosine triphosphate (ATP), and the potassium salt of 14
for 24 h at room temperature, under light and with shaking. Subsequently, the
carotenoids formed were extracted, saponified, and concentrated by phase separation.
The crude zeaxanthin obtained from the hypophase was purified by column
chromatography (CC) on ZnCO3/Celite.
The FT-IR spectrum showed the presence of deuterium in the molecule. As the
signals of Me(16,16',17,17') coincide in the 1H-NMR spectrum, the mixture was acetylated
to allow discrimination of the terminal Me groups [12]. From this spectrum, it was
possible to establish that the intensities of signals of Me(17,17',18,18',19,19',20,20') were
significantly reduced in comparison with those of Me(16,16'). From this reduction, we
inferred ca. 10% incorporation of deuterium into 15.
The 2H-NMR spectrum showed three signals in a ratio of 2 :1:1 (Fig. 1). The signal
at 1.07 ppm corresponds to CD3(17,17'), that at 1.72 ppm to CD3(18,18'), and the most
intense, at 1.96 ppm, to CD3(19,19',20,20'). A CD spectrum of 15b, recorded at room
temperature (Fig. 2), agrees completely with that of unlabeled (3R,3'R)-zeaxanthin
[13]. This finding confirms that the ring closure leaves intact the identity of the terminal
3
)
The Me groups at C(7), or C(1) and C(1') of 5 and 6 may be distinguished and assigned by NMR
spectroscopy; in 5a and 5c, they can also be distinguished by GC [4].
4
5
6
)
)
)
For a review of carotenoids in rose flowers, see [7].
Incorporation experiments were carried out by S. S. M. and by Dipl. Biochem. A. Eisenring.
Flavobacterium R 1560 is a gram-negative organism, obtained by F. Hoffmann LaRoche microbiologists via
mutation of the wild type. Its main carotenoid is zeaxanthin.

Helvetica Chimica Acta ± Vol. 83 (2000)
2039
Scheme 2
a) HOCH2CH2SH, BF3 ´ OEt2. b) LiAlH4, THF. c) 3,4-Dihydro-2H-pyran, Pyridinium toluene-4-sulfonate. d)
HgCl2, CaCO3, H2O, MeCN. e) CD3MgI. f) TsOH, H2O. g) Pyridinium chlorochromate, CH2Cl2.
Me groups. Our findings complement the experiments carried out on the same
organism with [2-13C]MVA by Britton et al. [5]; (see 2 in Scheme 1 and Sect. 5).
We assume that the incorporation of 14 (or its K salt) into 15 proceeds through the
intermediates shown in brackets (Scheme 2). If (all-E)-lycopene serves as precursor
and the acid-catalyzed cyclization proceeds by the concerted re(2),si(2),si(6) electron
migration, then this mode of cyclization corresponds to the C2(E,E) scheme systemized
by us [14].
3.2. Direct Incorporation of (all-E)-[16,16,16,16',16',16'-2H6]Lycopene into (all-
E,1R,1'R)-b,b-[16,16,16,16',16',16'-D6]Carotene (17; Scheme 3). ± The main carotenoid
of Flavobacterium R 1560 is (all-E,3R,3'R)-zeaxanthin. b,b-Carotene is also formed in

2040
Helvetica Chimica Acta ± Vol. 83 (2000)
Fig. 1. 2H-NMR Spectrum of biosynthetic di-O-acetylzeaxanthin (15b)
Fig. 2. Qualitative CD spectrum of di-O-acetylzeaxanthin (15b) at room temperature in Et2O/isopentane/EtOH
5 :5 :2 (EPA)

Helvetica Chimica Acta ± Vol. 83 (2000)
2041
Scheme 3
small amounts, however [15]. There is wide agreement among other authors that this is
the precursor of the monohydroxylated product kryptoxanthin, and this in turn of
zeaxanthin.
To achieve a sufficiently high degree of labeling in the precursor, it is necessary to
reduce endogenous production of b,b-carotene. To this end, we added mevinolin (16)7)
to the culture solution, together with the labeled lycopene (6). In addition, to facilitate
incorporation of the hydrophobic lycopene, we performed these experiments using a
cell-free system: a culture of Flavobacterium was homogenized in a ꢁFrench pressꢀ in
three stages (at 10000 psi (ˆ689 bar) in each case, with some deoxyribonuclease-I
(DNAase I) also added at each stage). In this way, the greater portion of cells was
lysed. The homogenate was separated from the remaining intact cells by centrifugation
at 6000 Â G. The clear supernatant obtained contained ca. 165 mg/ml protein.
3.3. Incubations. Each of 400 individual samples containing 3 ml of supernatant,
with 20 ml of 6 in 0.1 ml of EtOH, 10 mg of 16, and 1 ml of 0.1m Tris ´ HCl buffer,
pH 7.0, added, was shaken at 308 under light. In total, 8 mg of 6 were consumed in this
way. After 24 h, incubation was terminated by addition of a small quantity of MeOH.
The carotenoids formed were extracted with Et2O, saponified, and separated into
hypophasic and epiphasic fractions; carbohydrates were then separated by CC on
aluminium oxide and silica gel. Finally, 5 mg of b,b-carotene were obtained, still
containing (Z)-isomers. Separation of these was achieved by semipreparative HPLC.
2
The HPLC profile and the H-NMR spectrum of 17 are shown in Figs. 3 and 4,
respectively; the qualitative CD spectrum of 17 is shown in Fig. 5 along with the
analogous spectrum of 4 for comparison; in Fig. 6 is demonstrated how the labeled
molecule 17 can easily be detected by FTIR even in a mixture containing unlabeled b,b-
carotene in a ratio of 1:99. The data show that we were able to obtain a largely pure
mixture of 17 and unlabeled b,b-carotene.
7
)
Mevinolin (16), isolated and structurally interpreted by Alberts et al. [16], is a potent inhibitor of HMG
coenzyme reductase (HMG ˆ 3-(hydroxymethyl)glutaric acid), which, for its part, is a precursor of
mevalonic acid (MVA) or its lactone 14.

2042
Helvetica Chimica Acta ± Vol. 83 (2000)
Fig. 3. HPLC of biosynthetic (1S,1'S)-[16,16,16,16',16',16'-2H6]-b,b-carotene (17) (on Spherisorb S5-NH2, with
hexane/N-ethyldiisopropylamine 99.9 :0.1; flow 0.5 ml; detection 445 nm)
Fig. 4. 2H-NMR Spectrum of biosynthetic (1S,1'S)-[16,16,16,16',16',16'-2H6]-b,b-carotene (17) in CHCl3

Helvetica Chimica Acta ± Vol. 83 (2000)
2043
Fig. 5. Quantitative CD spectrum of semisynthetic (1R,1'R)-[16,16,16,16',16',16'-2H6]-b,b-carotene (4) at À1808
[3] (left) and of biosynthetic (1S,1'S)-[16,16,16,16',16',16'-2H6]-b,b-carotene (17) (right). Qualitative in EPA at
À1808.
Fig. 6. IR Difference spectrum between unlabeled b,b-carotene and biosynthetic 17 (99 :1 in CHCl3)
The CD curve shows that the labeled b,b-carotene is the enantiomer of 4.
Comparing the CD spectrum with the corresponding VIS spectrum, we concluded that
the rate of incorporation was 5 ± 10%. Hence, despite the mevinolin (16), an
appreciably high degree of dilution by endogenous b,b-carotene had taken place. That
dilution by enantiomeric 4 might also have occurred, however, we were unable to rule
out. However, we believe that we can discount this possibility on the grounds of the
isolation of 15 in the first experiment.
3.4. Synthesis of Specifically Deuterated Terpenoids. To determine the absolute
configuration of a biosynthetically obtained cyclized carotene by CD spectroscopy, a
relatively high degree of incorporation is essential. Since, over a long period, we had
been unable to achieve such a level, the following deuterated terpenoids and their
diphosphoric acid esters were prepared as possible precursors.

2044
Helvetica Chimica Acta ± Vol. 83 (2000)
[2H3]Dimethylallyl Alcohol ([2H3]DMAA; 3-methyl[4,4,4-2H3]but-2-en-1-ol; 21)
and its pyrophosphate (DMAPP, 23) were prepared as outlined in Scheme 4. According
to the procedure of Corey et al. [17], propynol (18) was deprotonated, alkylated with
CD3I, and stereoselectively converted to 20 with LiAlH4/NaOMe and I2. Treatment
with Me2CuLi produced the C5 alcohol 21. [2H3]DMAPP 23 was then prepared from
this by a procedure of Poulter and co-workers [18], via the bromide 22. The structure of
1
23 was confirmed by its H- and 32P-NMR spectra. The signal of the (E)-Me group,
prominent in the unlabeled reference compound, was completely absent in 23. In
comparison with the free alcohol, HÀC(1) is shifted by ca. 0.3 ppm. The chemical shifts
of P(a) and P(b) matched literature values.
Scheme 4
a) LiNH2, NH3, CD3I. b) LiAlH4, NaOMe, THF, I2. c) Me2CuLi. d) Tris(tetrabutylammonium)hydrogen
diphosphate. e) PBr3, Et2O. f) 1. MeC(O)CH2C(O)OMe, NaH, THF; 2. BuLi, Hexane, ClPO(OEt)2; 3.
Me2CuLi. g) DIBAH, Hexane. h) CCl4, Ph3P. i) N-Chlorosuccinimide, MeSMe.
[2H3]Geraniol ([2H3]GA, 25) and [2H3]GPP (27). For the starting material, we used
the alcohol 21 and its bromide 22 (Scheme 4). Extension by five C-atoms was
accomplished in a one-pot procedure by the method of Weiler and co-workers [19]. The
dianion of ethyl acetoacetate was regioselectively alkylated in the g-position with 22,
the intermediate enolate then converted to the enol phosphate with chlorodiethyl-
phosphoric acid, and this was finally substituted with a Me group from Me2CuLi. The
C(2)ˆC(3) bond retains its (E)-configuration in this reaction sequence.
We isolated the monoterpene ester 24 in 73% yield. After reduction to geraniol 25
and its conversion to the chloride 26 and subsequent pyrophosphoric acid esterification
according to [18], we obtained [2H3]GPP (27) in good yield. Its spectroscopic data
agreed with those of a reference compound prepared by a different route [4].

Helvetica Chimica Acta ± Vol. 83 (2000)
2045
[2H3]Farnesol (29, [2H3]FA) and [2H3]FPP (31) were prepared in two different ways
(Schemes 4 and 5). First, geraniol 25 was converted to the chloride 26 with Ph3P/CCl4
[20], and this was then extended by the already mentioned Weiler procedure [19]. In
contrast to the preparation of 24, it was necessary in this instance to isolate the
intermediate enol phosphate, as the Me2CuLi treatment proceeded only incompletely
otherwise. Reduction of 28 to [2H3]farnesol (29) and pyrophosphoric acid esterification
of the chloride 30 to give 31 were carried out as described for [2H3]geraniol (25).
For the second route, the chloride 22 was converted to the sulfone 32 with
NaSO2Ph. We converted the anion of this to the more strongly acidic 33 with methyl
chloroformate. This was extended by ten C-atoms with the palladium complex of
methyl geranate according to the procedure of Trost et al. [21].
Scheme 5
a) NaSO2Ph. b) BuLi, ClCOMe. c) p-Pd complex of methyl geranate, Ph3P, NaH, THF. d) 4-Aminothiophenol,
Cs2CO3. e) DIBAH, Hexane. f) Li, EtNH2. g) N-Chlorosuccinimide, MeSMe. h) Tris(tetrabutylammonium)
hydrogen diphosphate.
As expected, only the Me signal at 1.56 ppm appeared in the NMR spectrum of
sulfone 34; hence, the reaction was stereospecific. After decarboxylation of 34 with LiI/
NaCN [21], we were able to isolate only 44% of 35. We obtained double that yield,
however, by using the method of Keinan and Eren [22]. After reduction of ester 35 to 36
and subsequent desulfonation according to [22], we once more obtained 29, which had
spectra identical to those obtained from the sample prepared by the first route. Apart
from the lack of the terminal (E)-Me group, its identity with non-deuterated farnesol
was also confirmed. Conversion of 36 to chloride 30 and diphosphorylation [18] finally
supplied 31 once more.
[2H3]Geranylgeraniol ([2H3]GGA; Scheme 6). The extension technique employed
in the second route to [2H3]farnesol 29 (Scheme 5) was used, but with 33 as starting
material. Spectroscopically, the isolated [2H3]GGA was identical to non-deuterated
geranylgeraniol [21], apart from the lack of any signal from the terminal (E)-Me group.

2046
Helvetica Chimica Acta ± Vol. 83 (2000)
2
In the H-NMR spectrum, the signal of the CD3 group lies at 1.65 ppm. The planned
diphosphorylation encountered difficulties and so was not pursued further.
Scheme 6
a) 1. PdCl2, CuCl2, methyl farnesate, 2. Ph3P, NaH, THF. b) 4-Aminothiophenol, Cs2CO3. c) DIBAH, Hexane.
d) Li, EtNH2.
4. Discussion. ± The theory, set out quite some years ago, that lycopene is the most
important precursor of carotenes with cyclic end groups (summarized by Goodwin in
[23]) has now been conclusively proven by our experiments.
As regards the problem of the stereochemisty of folding of the 1,5-diene (in
lycopene) and the chirality about C(1) in the b-end group, two groups have already
provided an answer, but with mutually contradictory results. To summarize briefly:
BuꢀLock et al. [6] isolated trisporic acid C from a Blakslea trispora culture, and the
structure 42 was assigned to it on the basis of extensively interpreted CD spectra and of
its further transformation (Scheme 7). Since the fungus also produced b,b-carotene as
well, the authors viewed 42 as a degradation product of b,b-carotene.
After addition of [2-14C]mevalolactone (41) to the culture medium, it proved
possible to isolate radiolabeled trisporic acid C, with an incorporation rate of 0.6%.
The decisive experiment to explain the distribution of radioactivity in the prochiral Me
groups at C(1) was carried out by selective hydrogenation of the side-chain CˆC bond,
followed by thermal decarboxylation of the vinylogous b-keto carboxylic acid 43 to 44.
The captured CO2 showed an activity of 1.4% of the total activity of 43, against 99.3%
for 44. From this, it was possible to infer a sterically specific folding of the 1,5-diene in
chair-like manner, as anticipated in formula 45. By our nomenclature, this would
correspond to folding in C1-(E,E) mode [14] with si(2),re(1),si(6) concerted electron

Helvetica Chimica Acta ± Vol. 83 (2000)
2047
Scheme 7
migration. After oxidation of the b,b-carotene Mepro-S group (ˆ Me(16)) to a CˆO
group, introduction of the oxo group at C(4) and degradation of the side chain at C(1),
the result is trisporic acid C (42)8).
Using a completely different experimental approach, Britton et al. [5] later found a
quite different result with Flavobacterium R 1560. Incorporation of [2-13C]mevalolac-
tone in the presence of the cyclase inhibitor nicotine produced a labeled lycopene in
which, according to NMR analysis, enhanced labeling had occurred at Me(16) and
Me(16'), but not at Me(17) and Me(17'). Nicotine-free cultures provided zeaxanthin
(2), likewise with labeling at Me(16) and Me(16'). From these results, Britton et al.
inferred a folding of the lycopene (as the zeaxanthin precursor) by the C2-(E,E) mode
(Scheme 2). Further incorporation experiments made it possible to confirm and
enlarge on these results [24].
Our findings (Sects. 3.1 and 3.2) supplement the conclusions of Britton et al. [5]. To
interpret the apparently contradictory findings of BuꢀLock and co-workers [6], we tend
to favor the assumption of an organism-specific cyclization. It is true that, to date, only a
few confirmed instances of enantiomeric folding of the 1,5-diene are known in
carotenoid biogenesis, quite in contrast to the many examples in the diterpenoid series.
8
)
The Mepro-S group in the b-ring is given the locant 16 according to carotenoid nomenclature. Hence, it
corresponds biogenically to the same numbering in the precursor lycopene, provided that its folding follows
the C2-(E,E) mode (see Schemes 2 and 3). However, an inversion in the numbering in fact occurred in the
C2-(E,E) mode (see Scheme 7).

2048
Helvetica Chimica Acta ± Vol. 83 (2000)
They mostly involve the variable sense of chirality at C(6) in carotenes with an e end
group9).
Although the (6R,6'R)-e,e-carotene is found in tomatoes [25], the (6S,6'S)-
enantiomer is found in algae [26]. We have postulated a (5Z)-lycopene precursor as
a prerequisite for the formation of an e- rather than a b-ring [27].
We thank the Schweizerische Nationalfonds zur Förderung der wissenschaftlichen Forschung for financial
support of our work. We are grateful to Dr. P. Bramley, then of Universität Konstanz, for his advice on
incorporation experiments with cultures of Phycomyces blakesleanus, Drs. E. K. Weibel and E. Hochuli, F.
²
Hoffmann-LaRoche AG, Basel, for cultures of Flavobacterium R 1560, Dr. A. Hofer for test samples of
[16,16,16,16',16',16'-2H6]lycopene, Dr. H.P. Märki for b,b-[16,16,16,16',16',16'-2H6]carotene and reference
spectra, Dr. A. W. Albert, Merck, Sharp & Dohme, Rahway, USA, for a generous test sample of mevinolin,
and Dr. R. K. Müller, F. Hoffmann-LaRoche AG, Basel, for providing us with pure geranylgeraniol. For in-depth
discussions and valuable suggestions, we are also grateful to Drs. G. Britton and M. Brown, University of
Liverpool, UK, and to the following colleagues from the Organisch-Chemischen Institut, Universität Zürich:
Dr. Heidi Meier, Dr. A. Zumbrunn, Prof. Dr. W.-D. Woggon, Prof. Dr. P. Rüedi, and Ms. Dipl. Chem. Edith
Märki-Fischer for their support. For NMR spectra, we thank Prof. W. von Philipsborn and co-workers, for mass
spectra Prof. M. Hesse and co-workers, for FT-IR spectra Dr. M. Rey, for IR spectra and analyses Mr. H.
²
Frohofer and co-workers, and for CD spectra Dr. K. Noack, F. Hoffmann-LaRoche AG, Basel. Finally, C. H. E.
thanks Prof. H.-J. Hansen for positions and facilities in his research group.
Experimental Part
General. See [28]. Solvents used for reactions were of anal. quality and were dried carefully before use. For
particulars of biological procedures see [1], p. 62 ± 65 and 83 ± 84.
1. Incorporation of [3',3',3'-2H3]MVA ((Æ)-14). The culture of bacteria in liquid medium was carried out in
modified 2-l Erlenmeyer flasks, each containing 500 ml of culture medium in an orbital incubator at 160 rpm at
r.t., under light, over 30 h. The ensuing sedimentation of bacteria was accomplished at 10000 Â g, followed by 3-
fold washing with 0.1m Tris ´ HCl buffer, pH 7.0. The bacterial mass from 4 l of culture was suspended in 0.1m
Tris ´ HCl buffer (320 ml) and divided into 40-ml portions in 8 Erlenmeyer flasks. To each incubation experiment
was added the K salt of (Æ)-14 (40 mg), hen egg albumin lysozyme (60 mg), MnSO4 ´ 4 H2O (22 mg), MgSO4 ´ 7
H2O (75 mg), and ATP (100 mg) dissolved in Tris ´ HCl buffer (10 ml), and the incubations were carried out as
for the bacterial cultures over 24 h.
2. Isolation of Zeaxanthin (15a). The combined products were filtered over Celite, then extracted with
acetone until colorless, the filtrate diluted with Et2O and thoroughly washed with H2O. After drying and
removal of solvent, saponification was performed with 10% KOH in MeOH. The usual extraction was
performed, followed by phase separation between 95% MeOH and hexane. CC of hypophasic carotenoids on
silica gel with acetone/hexane 4 :1 gave 40 mg of zeaxanthin, which was acetylated with Ac2O/pyridine.
Purification of the acetate was performed by prep. HPLC. CD: see Fig. 2. FT-IR (Fig. 6; CHCl3; difference
spectrum): 2241 (sh), 2214s, 2173w, 2106w, 2063s. 1H-NMR (400 MHz, CDCl3): 1.074 (s, Me(17,17')); 1.105
(s, Me(16,16')); 1.583 (t, 3J ˆ 12, HÀC(2,2')); 1.721 (s, Me(18,18')); 1.775 (dd, 3J ˆ 12, HÀC(2,2')); 1.966, 1.972
(2s, Me(19,19',20,20')); 2.049 (s, 2 Ac); 2.108 (dd, HÀC(4,4')); 2.439 (dd, HÀC(4,4')); 5.059 (m, HÀC(3,3'));
6.101 (m, HÀC(7,7')); 6.114 (m, HÀC(8,8')); 6.160 (m, HÀC(10,10')); 6.257 (m, HÀC(14,14')); 6.365
(m, HÀC(12,12')); 6.607 ± 6.995 (m, HÀC(11,11',15,15')). 2H-NMR (Fig. 1; 61.4 MHz, CDCl3): 1.072
(s, CD3(17,17')); 1.720 (s, CD3(18,18')); 1.963 (s, CD3(19,19',20,20')).
3. Preparation of a Cell-Free System from Flavobacterium R 1560. The bacterial mass from 4 l of culture was
isolated by centrifugation, washed several times with 0.1m Tris ´ HCl buffer (pH 7), and finally suspended in the
same buffer (40 ml). With a ꢁFrench pressꢀ (Aminco), the cells were homogenized 3 times at 48 and 10000 psi.
After each pass, a few mg of DNAase I were added to the soln. Non-homogenized material and residual cell wall
and cell membrane were then separated by centrifugation for 2 h at 6000 Â g. The supernatant was used as a cell-
free system for incubation experiments.
9
)
The stereochemistry of folding of the precursor may also be inferred from the sense of chirality at C(6) [25].

Helvetica Chimica Acta ± Vol. 83 (2000)
2049
4. Incorporation of [16,16,16,16',16',16'-2H6]Lycopene (6). Each incubation run contained a soln. of 6
(20 mg) in EtOH (0.1 ml), mevinolin (16) (10 mg), and the cell-free system (3 ml; containing ca. 500 mg of
protein), made up to 4 ml with 0.1m Tris ´ HCl buffer (pH 7.0). Four hundred such incubation runs were
continually illuminated and shaken over a 20 h period at 308.
5. Isolation and Characterization of Labeled b,b-Carotene (17). After termination of the incubation by
addition of a quantity of MeOH and thorough mixing with a Vortex mixer, the contents were filtered over Celite,
and the filter cake was extracted with acetone until no more color was apparent. The isolated carotenoids were
saponified, separated into epi- and hypophases, and the carotene hydrocarbons were separated by CC on Al2O3
(Woelm, neutral; hexane/Et2O 19 :1). Further purification from the (Z)-isomer by HPLC (Spherisorb S5, NH2;
hexane/EtN(i-Pr)2 99.9 :0.1; 0.5 ml/min; detection at 445 nm (Fig. 3) provided crystalline 17 (5 mg). CD (Fig. 5;
EPA; 5 :5 :2): qualitative. 1H-NMR (400 MHz, CDCl3): 1.029 (s, Me(16,16',17,17')); ca. 1.47 (m, H2ÀC(2,2'));
ca. 1.63 (m, H2ÀC(3,3')); 1.720 (s, Me(18,18')); 1.97 (s, Me(19,19',20,20')); 2.02 (m, H2ÀC(4,4')); 6.11 ± 6.19
(m, Me(7,7',8,8',10,10')); 6.265 (m, HÀC(14,14')); 6.355 (d, J ˆ 15, HÀC(12,12')); 6.618 ± 6.683
2
(m, HÀC(11,11',15,15')). H-NMR (Fig. 4; 61.4 MHz, CDCl3): 1.035 (s, CD3(1,1')).
6. (Æ)-[3',3',3'-2H3]MVA (14). 6.1. Dimethyl 2,2'-(1,3-Oxathiolane-2,2-diyl)bisacetate Dimethyl Ester (8). To
a soln. of dimethyl 3-oxopentanedioate (43.87 g) in CH2Cl2 (250 ml) was added, with stirring, first 2-mercapto-
ethanol (29.68 g) and then BF3 ´ Et2O (42.38 g). Stirring was continued for 8 h at r.t. The mixture was then
diluted with Et2O (400 ml) and H2O (200 ml). After 15 min, the aq. phase was separated, and the Et2O layer
washed with aq. NaHCO3 and then with aq. NaCl until neutral. The oil obtained, after drying (Na2SO4) and
solvent removal, crystallized after a short time: 8 (52.5 g, 89%). M.p. 458. IR (CHCl3): 2958w, 1740s, 1438s,
1342m, 1195w, 1175m, 1080. 1H-NMR (80 MHz, CDCl3): 3.06 (t, J ˆ 6.2, 6H, 2 HÀC(4)); 3.2 (AB, J ˆ 18, 4 H);
‡
3.69 (s, 2 MeO); 4.19 (t, J ˆ 6, 2 HÀC(5)). MS: 234 (6, M ), 203 (3), 175 (11), 161 (33), 143 (16), 101 (25), 60
(100), 59 (35), 45 (14). Anal. calc. for C9H14O5S (234.27): C 46.14, H 6.02, S 13.69; found: C 46.21, H 5.90, S 13.69.
6.2. 2,2'-(1,3-Oxathiolane-2,2-diyl)bisethanol (9). LiAlH4 (10 g) was added to THF (120 ml) under an inert
atmosphere. A soln. of 8 (30.86 g) in THF (250 ml) was added dropwise with stirring. Stirring was continued for
12 h, and then hydrolysis was performed with MeOH/H2O 1:1 (30 ml), followed by 15% aq. NaOH (15 ml).
After addition of H2O (45 ml), stirring was continued for a further 30 min, and the mixture was then filtered
over Celite; washing with THF. After evaporation of the org. phase, the residue obtained was dried in vacuo.
1
Compound 9 was obtained as an oil in quant. yield. H-NMR (80 MHz, CDCl3): 2.17 (t, J ˆ 6, 2 HÀC(2',2''));
3.07 (s, 2 OH); 3.08 (t, J ˆ 6, 2 HÀC(4)); 3.81 (t, J ˆ 6, 2 HÀC(1',1'')); 4.19 (t, J ˆ 6, 2 HÀC(4)). MS: 196 (6,
‡
‡
[M ‡ 18(NH3)] ),179 (24, [M ‡ 1] ), 169 (18), 136 (25), 120 (5), 102 (27), 85 (6), 35 (12). Anal. calc. for
C7H12OS (178.25): C 47.17, H 6.79, S 17.99; found: C 47.28, H 7.01, S 17.76.
6.3. 2,2-Bis[2-(tetrahydropyran-2-yloxy)ethyl]-1,3-oxathiolane (10). 3,4-Dihydro-2H-pyran (33.73 g) and
TsOH (0.5 g) were added at 08, with powerful stirring, to a soln. of 9 (23.52 g) in CH2Cl2 (200 ml). After the
reaction was complete, the soln. was diluted with Et2O (300 ml), and then washed with aq. NaHCO3 and aq.
NaCl. After drying, solvent removal, and filtration through deactivated silica gel (hexane/Et2O 3 :2), 10 was
obtained as a colorless oil (42 g, 92%). 1H-NMR (80 MHz, CDCl3): 1.5 ± 1.8 (m, 6 CH2 (tetrahydropyran)); 2.2
(t, 2 HÀC(2',2'')); 3.05 (t, 2 HÀC(4)); 3.4 ± 4.0 (m, 2 HÀC(1',1''), 2 CH2 (tetrahydropyran)); 4.15
(t, 2 HÀC(5)); 4.55 (m, 2 CH (tetrahydropyran)).
6.4. 1,5-Bis(tetrahydropyran-2-yloxy)pentan-3-one (11). A soln. of 10 (14 g) in MeCN/H2O 4 :1 (150 ml)
was treated with CaCO3 powder (8 g) and then with HgCl2 (24 g), stirred for 2 h at r.t., and filtered over Celite.
The filtrate was treated with aq. NH4OAc and extracted with Et2O. After the usual workup, 11 (10.41 g, 90%)
was obtained. IR (film): 2940s, 1715s, 1350m, 1200m, 1120m, 1035m. 1H-NMR (80 MHz, CDCl3): 1.48 ± 1.80
(m, 6 CH2 (tetrahydropyran)); 2.75 (t, 2 HÀC(2,4)); 3.5 ± 4.15 (m, 2 HÀC(1,5), 2 CH2 (tetrahydropyran)); 4.55
(m, 2 CH (tetrahydropyran)).
6.5. 1,5-Bis(tetrahydropyran-2-yloxy)-3-[2H3]methylpentan-3-ol (12). To a Grignard soln. prepared from Mg
(1 g) and CD3I (5 g) in ether (70 ml) was added rapidly at 08 a soln. of 11 (7.5 g) in Et2O (50 ml), and the
mixture was refluxed for 2.5 h. After hydrolysis with aq. NH4Cl and the usual workup, the product was
chromatographed (silica gel; hexane/Et2O 9 :1) to give 12 (7.2 g, 90%). Colorless oil. 1H-NMR (80 MHz,
CDCl3): 1.49 ± 1.76 (m, 6 CH2 (tetrahydropyran)); 1.85 (t, J ˆ 6, 2 HÀC(2,4)); 3.52 ± 3.93 (m, 9 H); 4.59
(m, 2 CH (tetrahydropyran)).
6.6. 3-[2H3]Methylpentane-1,3,5-triol (13). A soln. of 12 in MeOH (20 ml) was treated with TsOH (0.23 g)
and heated under reflux for 4 h. After cooling, NaHCO3 (0.21 g) was added, and the mixture was diluted with
CHCl3 (30 ml). After stirring for 20 min, the soln. was filtered through a layer of MgSO4 and solvent was
removed. CC on silica gel (CHCl3) provided 13 (0.55 g, 85%). Viscous oil. 1H-NMR (90 MHz, C5D5N): 2.10
(m, 2 HÀC(2,4)); 4.15 (t, 2 HÀC(1,5)); 5.75 (s, 3 OH).

2050
Helvetica Chimica Acta ± Vol. 83 (2000)
6.7. (Æ)-4-Hydroxy-4-[2H3]methyltetrahydropyran-2-one (14, [2H3]MVL). A soln. of 13 (0.77 g) in MeOH
Â
(7 ml) was added to Ag2CO3/Celite (Fetizon et al. [29]) in benzene (300 ml). MeOH was distilled off, and H2O
was removed azeotropically with benzene. After refluxing for 6 h, the mixture was filtered, solvent was
removed, and the product was distilled in a Kugelrohr at 1108/0.2 Torr to yield 14 (0.5 g). Clear liquid. IR
(CHCl3): 3603w, 3016s, 2227w, 1728s, 1404w, 1265m, 1234m, 1076m. 1H-NMR (80 MHz, CDCl3): 1.83 ± 1.98
(m, 2 HÀC(4)); 2.56 ± 2.62 (m, 2 HÀC(2)); 4.25 ± 4.69 (m, 2 HÀC(5)).
7. [2H3]DMAA (21) and [2H3]DMAPP (23). 7.1. [4,4,4-2H3]But-2-yn-1-ol (19). LiNH2 (12 g) was placed in a
flame-dried four-necked flask under Ar, and dried NH3 (50 ml) was condensed onto it at À708. Compound 18
was added dropwise over 15 min. After 45 min, CD3I (25 g) was slowly added with continual stirring. After 3 h,
the NH3 was allowed to evaporate, and the residue was taken up in H2O (70 ml), sat. with NaCl, and
continuously extracted with Et2O in a Soxhlet apparatus. After drying and distillation through a 10-cm Vigreux
column, 19 (6.3 g, 50%) was obtained. B.p. 828/100 Torr. 1H-NMR (90 MHz, CCl4): 3.6 (s, OH); 4.1
(s, 2 HÀC(2)).
7.2. 3-Iodo[4,4,4-2H3]but-2-en-1-ol (20). A suspension of LiAlH4 (4.85 g) in THF (1 l) was treated with
NaOCH3 (97%, 14.2 g) and heated under reflux for 30 min. It was then allowed to cool to r.t., and a soln. of 19
(4.66 g) in THF (20 ml) was added dropwise. Heating under reflux was resumed for 3 h. After cooling to À708, it
was treated slowly with I2 (62 g) in THF (140 ml). After stirring for 90 min at the same temp., the cooling bath
was removed, and the mixture was hydrolyzed with H2O and half-sat. NaCl soln. After extraction with H2O and
washing with thiosulfate soln. and sat. NaCl soln., it was dried (MgSO4), filtered, and evaporated. CC on silica
gel (CH2Cl2/Et2O 9 :1) produced 20 (11.18 g, 87%).
7.3. (E)-3-Methyl[4,4,4-2H3]but-2-en-1-ol (21). A suspension of CuI (28 g) in Et2O (300 ml) was slowly
treated with MeLi soln. in Et2O (1.6m, 185 ml) at À158 over a 1 h period. A clear soln. resulted. To this was
added dropwise under Ar a soln. of 20 (11.9 g) in Et2O (30 ml), and the mixture was allowed to stir for 4.5 d at
08. Workup was by careful addition of MeOH (4.5 ml) and H2O (10.5 ml). The precipitate was separated by
filtration over Celite. After extraction, washing, and solvent removal, the product was chromatographed (silica
gel; CH2Cl2/Et2O 4 :1). One distillation of the product in a Kugelrohr at 1008/100 Torr gave 21 (3.72 g, 71%).
7.4. (E)-1-Bromo-3-methyl[4,4,4-2H3]but-2-ene (22). A soln. of 21 (3 g) in Et2O (120 ml) was treated at 08
with PBr3 (1.7 ml) and stirred for 1 h at r.t. After addition of dil. aq. NaHCO3, it was washed until neutral, dried
(MgSO4), filtered, and concentrated. The subsequent Kugelrohr distillation provided 22 (3.4 g, 66%). 1H-NMR
(200 MHz, CDCl3): 1.734 (s, MeÀC(3)); 4.019 (d, J ˆ 8.5, 2 HÀC(1)); 5.530 (t, J ˆ 7, H ÀC(2)).
7.5. [2H3]DMAPP (23). Prepared by the method of [18]. 1H-NMR (200 MHz, D2O/NH4OD): 1.73
(s, MeÀC(3)); 4.46 (br. s, 2 HÀC(1)); 5.46 (br. s, HÀC(2)).
8. [2H3]GA (25) and [2H3]GPP (27). 8.1. Methyl (2E,6E)-3,7-Dimethyl[8,8,8-2H3]octa-2,6-dienoate (24). To
a suspension of NaH (55% in mineral oil, 459 mg) in THF (25 ml) at 08 was added, dropwise, methyl
acetoacetate (1.08 ml). After 10 min, the clear soln. obtained was treated with BuLi soln. in hexane (1.6m,
6.4 ml). After 10 min, bromide 22 (1.6 g) was added dropwise. After 10 min stirring at r.t., it was once more
cooled to 08 and treated with diethyl chlorophosphate (1.52 ml). After 2 h stirring at r.t., the mixture was added
dropwise to a soln. of Me2CuLi at À708. (This had been prepared previously by dropwise addition of MeLi soln.
(25 ml) to a suspension of CuI (3.8 g) in Et2O (25 ml).) After allowing the mixture to warm to 08, it was treated
with aq. NH4Cl and extracted with Et2O. After washing with a small quantity of NH3/aq. NaCl, drying, filtering,
and CC (silica gel; hexane/Et2O 7:3), the oil obtained was distilled in a Kugelrohr at 1308/10 Torr to yield 24
(1.35 g, 73%). Colorless oil. IR (CHCl3): 2956m, 2222w, 2186w, 1715s, 1650s, 1438s, 1150s. 1H-NMR (200 MHz,
CDCl3): 1.60 (s, MeÀC(7)); 2.16 (br. s, MeÀC(3), 2HÀC(4,5)); 3.69 (s, MeO); 5.07 (br. s, HÀC(2)); 5.67
(s, HÀC(6)).
8.2. (2E,6E)-3,7-Dimethyl[8,8,8-2H3]octa-2,6-dien-1-ol (25). An ice-cooled soln. of 24 (1.134 g) in THF
(120 ml) was treated with a soln. of diisobutylaluminium hydride (DIBAH, 1m, 17 ml) in hexane, and stirred for
10 min. After hydrolysis with an equiv. volume of H2O and further stirring for 1 h, the mixture was diluted with
Et2O, saturated with brine, and the phases were separated. The org. layer was washed thoroughly with dil., aq.
NaCl; drying (MgSO4), solvent removal, and distillation in a Kugelrohr provided 25 (1.145 g). Colorless oil. B.p.
1408/10 Torr. IR (CHCl3): 3610m, 2922s, 2222w, 2190w, 1668w, 1415w, 1380m, 980s. 1H-NMR (200 MHz, CDCl3):
1.60, 1.68 (2s, MeÀC(3,7)); 2.08 (s, 2HÀC(4,5), OH); 5.09 (br. s, HÀC(2)); 5.42 (t, J ˆ 7, H ÀC(6)). MS: 126 (5,
‡
[M À CH2OH] ), 93 (13), 72 (75), 40 (100).
8.3. (2E,6E)-1-Chloro-3,7-dimethyl[8,8,8-2H3]octa-2,6-diene (26). A soln. of 25 (1.1 g) and thoroughly
dried Ph3P (2.42 mg) in CCl4 (7 ml) was heated under reflux for 2 h. After cooling, it was diluted with Et2O
(15 ml) and cooled with stirring to 08, followed by filtration, solvent removal, and Kugelrohr distillation to yield
26 (820 mg, 78%). B.p. 608/0.2 Torr. 1H-NMR (90 MHz, CCl4): 1.60, 1.70 (2s, MeÀC(3,7)); 2.10 (s, 2 HÀC(4,5));

Helvetica Chimica Acta ± Vol. 83 (2000)
2051
4.00 (d, J ˆ 7, 2 H ÀC(1)); 5.05 (br. s, HÀC(2)); 5.40 (t, HÀC(6)). 2H-NMR (61.4 MHz, CHCl3): 1.67
(s, CD3(8)).
8.4. (2E,6E)-3,7-Dimethyl[8,8,8-2H3]octa-2,6-dienyl Diphosphate ([2H3]GPP; 27). Preparation analogous
to that of 23. 1H-NMR (200 MHz, D2O/NH4OD): 1.64 (s, MeÀC(7)); 1.73 (s, MeÀC(3)); 2.14 (m, 2HÀC(4,5));
4.48 (dd, J ˆ 6.6, 2 HÀC(1)); 5.24 (m, HÀC(6)); 5.48 (br. t, J ˆ 7, H ÀC(2)). 31P-NMR (81 MHz, D2O/ND4OD):
À9.51 (d, J ˆ 21, P(1)); 5.70 (d, J ˆ 22, P(2)).
9. [2H3]FA (29) and [2H3]FPP (31). 9.1. Methyl (2E,6E,10E)-3,7,11-Trimethyl[12,12,12-2H3]dodeca-2,6,10-
trienoate (28). A suspension of NaH (55% in mineral oil, 103 mg) in THF (5.5 ml) was added with ice-cooling to
methyl acetoacetate (250 mg) and then stirred for 10 min, after which BuLi in hexane (1.6m, 1.35 ml) was added.
After a further 10 min, 26 (280 mg) was added, and the mixture was stirred for 1 h at 08 and then for 1 h at r.t.
After hydrolysis with a little H2O and the usual workup, CC was performed (silica gel; hexane/Et2O 4 :1). It was
possible to recover unreacted 26 (84 mg, 30%). The enolized b-keto ester (245 mg, 64%) was isolated by
Kugelrohr distillation at 1008/0.01 Torr. The product was then added dropwise to a suspension of NaH (48 mg) in
THF (3.5 ml). After 20 min at 08, it was treated with ClPO(OEt)2 (194 mg). After stirring for 2 h at r.t., it was
then cooled to À788 and treated with the soln. prepared from CuI (430 mg) and MeLi (2.82 ml) (as described for
24). After an analogous workup, we obtained 28 (157 mg, 61%). Colorless oil. B.p. 1008/0.1 Torr. 1H-NMR
(90 MHz, CDCl3): 1.60 (s, MeÀC(7,11)); 2.0 ± 2.18 (m, MeÀC(3), 2 HÀC(4,8,9)); 3.68 (s, MeO); 5.10
(m, HÀC(6,10)); 5.58 (br. s, HÀC(2)).
9.2. (2E,6E,10E)-3,7,11-Trimethyl[12,12,12-2H3]dodeca-2,6,10-trien-1-ol (29). DIBAH reduction of 28
(157 mg) as described for 25 produced 29 (147 mg).
9.3. [2H3]Farnesyl Chloride (30). A soln. of N-chlorosuccinimide (98 mg) in CH2Cl2 (3 ml) was treated at
À308 with Me2S (0.54 ml). After cooling to À408, a soln. of 29 (140 mg) in CH2Cl2 (1 ml) was added to it
dropwise. The mixture was allowed to warm to 08 over 1 h. After a further 1 h stirring at this temp., it was diluted
with Et2O and worked up in the usual manner. Isolated 30 was used without further purification in the following
step.
9.4. (2E,6E,10E)-3,7,11-Trimethyl[12,12,12-2H3]dodeca-2,6,10-trienyl Diphosphate ([2H3]FPP; 31). Pre-
pared as described for 23. 1H-NMR (400 MHz, D2O): 1.62 (s, MeÀC(7,11)); 1.72 (s, MeÀC(3)); 2.04 ± 2.17
(m, 2 HÀC(4,5,8,9)); 5.20 (m, HÀC(6,10)); 5.50 (m, HÀC(2)). 31P-NMR (81 MHz, D2O/ND4OD): À9.28
(br. s, P(1)); À5.58 (br. s, P(2)).
9.5. Synthesis of 29 (via Scheme 5). (E)-3-Methyl-1-(phenylsulfonyl)[4,4,4-2H3]but-2-ene (32). A soln. of
21 (1.24 g) in Et2O (60 ml) was converted to bromide 22 with PBr3 (1.6 g) in Et2O (30 ml) as described under 7.4.
It was then treated immediately with NaSO2Ph (2.3 g) in DMF (12 ml). After stirring overnight at r.t. and the
usual workup, crystalline 32 (800 mg) was obtained. CC of the mother liquors (silica gel; CH2Cl2/Et2O 9 :1) gave
a further 152 mg of 32 (total yield 32%). 1H-NMR (200 MHz, CDCl3): 1.31 (s, MeÀC(3)); 3.79 (d, J ˆ 8,
2 HÀC(1)); 5.19 (d, J ˆ 8, HÀC(2)); 7.55 ± 7.61 (m, 3 H); 7.86 ± 7.89 (m, 2 H).
9.6. Methyl 4-Methyl-2-(phenylsulfonyl)[5,5,5-2H3]pent-3-enoate (33). A soln. of 32 (866 mg) in THF
(40 ml) was treated dropwise at À208 with a soln. of BuLi in hexane (1.6m, 3.6 ml). After 10 min, ClCOOMe
(1 ml, 1.23 g) was added. After stirring for 2 h at r.t., it was worked up in the usual manner. CC of the product
(silica gel; Et2O/hexane 7:3) gave 33 (760 mg). 1H-NMR (90 MHz, CDCl3): 1.46 (s, MeÀC(3)); 3.70 (s, MeO);
4.75 (d, J ˆ 9, HÀC(1)); 5.28 (d, J ˆ 9, HÀC(2)).
10. [2H3]GGA (39). 10.1. Methyl (2E,6E,10E,14E)-13-(Methoxycarbonyl)-3,7,11,15-tetramethyl-13-(phe-
nylsulfonyl)[16,16,16-2H3]hexadeca-2,5,10,14-tetraenoate (37). Prepared as described in [21]. To the p-allyl
complex of methyl farnesate (814 mg) in Et2O (90 ml) was added Ph3P (394 mg). After 10 min stirring, Et2O
was evaporated with N2, the residue dried in vacuo, taken up in THF (45 ml), and renewed addition of Ph3P
(394 mg) gave Soln. I. To a suspension of NaH (in mineral oil, 136 mg) in THF (40 ml) was added dropwise a
soln. of 33 (814 mg) in THF (8 ml) at r.t. This was stirred until a clear soln. had been produced, giving Soln. II,
which was added in one portion to Soln. I, and the mixture was stirred at r.t. for 5 h. The mixture was worked up
by pouring into half-sat. aq. NaCl and extracting with Et2O. The combined org. fractions were washed with H2O,
dried (MgSO4), filtered, and the solvent was removed. CC (silica gel; hexane/Et2O 3 :2) yielded 37 (547 mg,
34%). Colorless oil. 1H-NMR (90 MHz, CDCl3): 1.25 (t, J ˆ 7.5, MeO); 1.40 (s, MeÀC(15)); 1.56
(s, MeÀC(7,11)); 1.93 (br. s, 2 CH2); 2.13 (s, MeÀC(3), 2 CH2); 2.95 ± 3.05 (2 br. s, 2 HÀC(12)); 3.66
(s, MeO); 4.13 (q, J ˆ 7.5, CH2O); 5.10 (m, HÀC(6,10)); 5.30 (s, HÀC(14)); 5.63 (s, HÀC(2)), 7.36 ± 7.63
(m, 2 H), 7.76 ± 8.0 (m, 2 H).
10.2.
(2E,6E,10E,14E)-3,7,11,15-Tetramethyl-13-(phenylsulfonyl)[16,16,16-2H3]hexadeca-2,6,10,14-tetra-
enoate (38). A soln. of 37 (547 mg), 4-aminophenol (270 mg), and Cs2CO3 (120 mg) in DMF (6 ml) was
stirred at 858 for 1.5 h. After cooling, the mixture was poured into half-sat. aq. NaCl, extracted with Et2O, and

2052
Helvetica Chimica Acta ± Vol. 83 (2000)
the combined org. phases were washed with dil. HCl and aq. NaCl, and dried (MgSO4). After filtration and
solvent removal, the product was chromatographed (silica gel; hexane/AcOEt 3 :1) to yield 38 (310 mg, 64%).
IR (CHCl3): 2935m (br.), 2984m, 1708s, 1648m, 1450m, 1385w, 1305s, 1145s, 1085m. 1H-NMR (200 MHz,
CDCl3): 1.14 (s, MeÀC(15)); 1.28 (t, J ˆ 7.5, MeO); 1.51 (s, 3 H); 1.57 (s, 3 H); 1.88 ± 2.12 (m, 2 HÀC(4,5,8,9));
2.15 (s, MeÀC(3)); 2.85 (d, J ˆ 15, 2 HÀC(12)); 3.84 (m, HÀC(13)); 4.14 (q, J ˆ 7.5, CH2O); 4.89 (d, J ˆ 10,
HÀC(14)); 5.60 (m, HÀC(6,10)); 5.66 (s, HÀC(2)); 7.4 ± 7.7 (m, 3 H); 7.8 ± 7.9 (m, 2 H).
10.3. (2E,6E,10E,14E)-3,7,11,15-Tetramethyl-13-(phenylsulfonyl)[16,16,16-2H3]hexadeca-2,6,10,14-tetraen-
1-ol (39). A soln. of 38 (600 mg) in THF (40 ml) was treated dropwise at À208 with a soln. of DIBAH in hexane
(1m, 3 ml). After stirring for 1 h at this temp., it was hydrolyzed with H2O and extracted with Et2O and worked
1
up as usual. Chromatography (silica gel; Et2O/hexane 7:3) yielded 39 (495 mg, 91%). Colorless oil. H-NMR
(200 MHz, CDCl3): 1.13 (s, MeÀC(15)); 1.43 (s, OH); 1.51 (s, MeÀC(7)); 1.57 (s, MeÀC(13)); 1.68
(s, MeÀC(11)); 1.89 ± 2.139 (m, 2 HÀC(4,5,8,9)); 3.86 (m, HÀC(13)); 4.16 (d, J ˆ 7, 2 H ÀC(1)); 4.89 (d, J ˆ
11, HÀC(14)); 5.00 ± 5.20 (m, 2 HÀC(12)); 5.41 (t, J ˆ 7, H ÀC(2)); 7.47 ± 7.62 (m, 3 H); 7.81 ± 7.86 (m, 1 H).
10.4. (2E,6E,10E,14E)-3,7,11,15-Tetramethyl[16,16,16-2H3]hexa-2,6,10,14-tetraen-1-ol ([2H3]GGA; 40). The
desulfonation of 39 was carried out with Li/EtNH2 as described in [26]. 1H-NMR (200 MHz, CDCl3): 1.60
(s, MeÀC(7,11,15)); 1.68 (s, MeÀC(3)); 4.15 (d, J ˆ 6.8, 2 HÀC(1)); 5.11 (br. s, HÀC(6,10,14)); 5.42
‡
‡
(m, HÀC(2)). 2H-NMR (61.4 MHz, CHCl3): 1.65 (s, CD3(16)). MS: 293 (9, M ), 277 (27, [M À H2O] ), 276
(100), 220 (23), 208 (84), 194 (30), 189 (19), 180 (29), 166 (35), 152 (59), 149 (58), 140 (69), 135 (42), 126 (41),
123 (53), 121 (62), 109 (40), 95 (43), 81 (70), 72 (45).
REFERENCES
[1] S. S. Mohanty, Dissertation, University of Zürich, 1988.
[2] ꢁCarotenoidsꢀ, Eds. G. Britton, S. Liaaen-Jensen, H. Pfander, Birkhäuser, Basel, 1998, Vol. 3, p. 385.
[3] a). H. P. Märki, Dissertation, University of Zürich, 1980; b) H. P. Märki, C. H. Eugster, J. Chem. Soc.,
Chem. Commun. 1980, 527; c) H. P. Märki, C. H. Eugster, Helv. Chim. Acta 1981, 64, 1257; d) C. H.
Eugster, Pure Appl. Chem. 1979, 51, 498.
[4] A. Hofer, Dissertation, University of Zürich, 1981; A. Hofer, C. H. Eugster, Helv. Chim. Acta 1982, 65, 365.
[5] G. Britton, T. W. Goodwin, W. J. S. Lockley, A. P. Mundy, N. J. Patel, G. Englert, J. Chem. Soc., Chem.
Commun. 1979, 27.
[6] D. J. Austin, J. D. BuꢀLock, D. Drake, Experientia 1970, 26, 348; J. D. BuꢀLock, D. J. Austin, G. Snatzke, L.
Hruban, J. Chem. Soc., Chem. Commun. 1970, 255.
[7] C. H. Eugster, E. Märki-Fischer, Angew. Chem. 1991, 103, 671.
[8] S. C. Kushwara, C. Subbarayan, D. A. Beeler, J. W. Porter, J. Biol. Chem. 1969, 244, 3635; S. C. Kushwaha,
G. Suzue, C. Subbarayan, J. W. Porter, J. Biol. Chem. 1969, 245, 4708; J. W. Porter, S. L. Spurgeon, Pure
Appl. Chem. 1979, 51, 609; C. Papastephanou, F. J. Barnes, A. V. Briedis, J. W. Porter, Arch. Biochem.
Biophys. 1973, 137, 415.
[9] H. M. Hill, S. K. Calderwood, L. J. Rogers, Phytochemistry 1971, 10, 2051.
[10] P. M. Bramley, A. Than, B. H. Davies, Phytochemistry 1977, 16, 235.
[11] T. K. Chakraborty, V. Bhushan, S. Chandrasekaran, Indian J. Chem. 1983, 22B, 9.
[12] G. Englert, in ꢁCarotenoid Chemistry and Biochemistryꢀ, Eds. G. Britton, T. W. Goodwin, Pergamon Press,
Oxford, 1982, p. 107.
[13] K. Noack, in ꢁCarotenoid Chemistry and Biochemistryꢀ, Eds. G. Britton, T. W. Goodwin, Pergamon Press,
Oxford, 1982, p. 135.
[14] C. H. Eugster, Pure Appl. Chem. 1979, 51, 490.
[15] G. Britton, D. J. Brown, T. W. Goodwin, F. J. Leuenberger, A. J. Schocher, Arch. Microbiol. 1977, 113, 33.
[16] A. W. Alberts, J. Chen, G. Kuron, V. Hunt, J. Huff, C. Hoffman, J. Rothrock, M. Lopez, H. Joshua, E.
Harris, A. Patchett, R. Monaghan, S. Currie, E. Stapley, G. Albers-Schönberg, O. Hensens, J. Hirshfield, K.
Hoogsteen, J. Liesch, J. Springer, Proc. Natl. Acad. Sci. U.S.A. 1980, 77, 3957.
[17] E. J. Corey, J. A. Katzenellenbogen, G. H. Posner, J. Am. Chem. Soc. 1967, 89, 4245.
[18] V. M. Dixit, F. M. Laskovics, W. I. Noall, C. D. Poulter, J. Org. Chem. 1981, 46, 1967; V. J. Davisson, A. B.
Woodside, C. D. Poulter, Methods Enzymol. 1984, 110, 130; V. J. Davisson, T. M. Zabriskie, C. D. Poulter,
Bioorganic Chem. 1986, 14, 46; V. J. Davisson, A. B. Woodside, T. R. Neal, K. E. Stremler, M. Mühlbacher,
C. D. Poulter, J. Org. Chem. 1986, 51, 4768.

Helvetica Chimica Acta ± Vol. 83 (2000)
2053
[19] S. N. Huckin, L. Weiler, J. Am. Chem. Soc. 1974, 96, 1082; F. W. Sun, L. Weiler, Tetrahedron, Woodward
Memorial Issue 1981, 37, 303.
[20] J. B. Lee, T. J. Nolan, Can. J. Chem. 1966, 44, 1331; R. Appel, Angew. Chem. 1975, 87, 863.
[21] B. M. Trost, P. E. Strege, L. Weber, T. J. Fullerton, T. J. Dietsche, J. Am. Chem. Soc. 1978, 100, 3407; B. M.
Trost, P. E. Strege, L. Weber, T. J. Fullerton, T. J. Dietsche, J. Am. Chem. Soc. 1978, 100, 3416; B. M. Trost,
P. E. Strege, L. Weber, T. J. Fullerton, T. J. Dietsche, J. Am. Chem. Soc. 1978, 100, 3426.
[22] E. Keinan, D. Eren, J. Org. Chem. 1986, 51, 3165.
[23] T. W. Goodwin, in ꢁCarotenoidsꢀ, Eds. O. Isler, H. Gutmann, U. Solms, Birkhäuser, Basel, 1971, p. 577 ff.
[24] G. Britton, Pure Appl. Chem. 1985, 57, 701.
[25] R. Buchecker, C. H. Eugster, Helv. Chim. Acta 1971, 54, 327; R. Buchecker, C. H. Eugster, Helv. Chim.
Acta 1973, 56, 1124; R. Buchecker, C. H. Eugster, Helv. Chim Acta 1971, 54, 327.
[26] T. Bjùrnland, S. Liaaen-Jensen, J. Throndsen, Phytochem. 1989, 28, 3347.
[27] C. H. Eugster, Pure Appl. Chem. 1979, 51, 463.
[28] A. Baumeler, W. Brade, A. Haag, C. H. Eugster, Helv. Chim. Acta 1990, 73, 700.
Â
[29] M. Fetizon, M. Golfier, J.-M. Louis, Tetrahedron 1975, 31, 171.
Received February 2, 2000