The substrate specificity of β,β-carotene 15,15′-monooxygenase

Article abstract of DOI:10.1002/1522-2675(20010815)84:8<2301::AID-HLCA2301>3.0.CO;2-U

The synthesis of several substrate analogues of the enzyme β,β-carotene 15,15′-monooxygenase is reported. The substrate specificity of enriched enzyme fractions isolated from chicken intestinal mucosa was investigated. Regarding substrate binding/cleavage, these experiments demonstrate that i) any deviation from the rod-like β,β-carotene structure is not tolerated, ii) one 'natural', unsubstituted β-ionone ring is required, iii) the position and presence of the Me groups attached to the polyene chain is significant. These results suggest a hydrophobic barrel-like substrate binding site in which the protein's amino acid residues through interaction with the Me groups, direct the central C=C bond in binding distance to the active site's metal-oxo center, supporting the unique regiospecificity of cleavage to retinal (provitamin A).

Full text of DOI:10.1002/1522-2675(20010815)84:8<2301::AID-HLCA2301>3.0.CO;2-U

Helvetica Chimica Acta ± Vol. 84 (2001)
2301
The Substrate Specificity of b,b-Carotene 15,15'-Monooxygenase
by Gabriele M. Wirtza), Claus Bornemanna), Alfred Gigerb), Robert K. Müllerb), Heinz Schneiderb), Götz
Schlotterbeckc), Gerhard Schieferb), and Wolf-Dietrich Woggon*a)
a) Institut für Organische Chemie der Universität Basel, St. Johanns-Ring 19, CH-4056 Basel
b) F. Hoffmann-La Roche Ltd., Vitamins and Fine Chemicals Division,
c) F. Hoffmann-La Roche Ltd., Pharmaceutical Division, CH-4070 Basel
The synthesis of several substrate analogues of the enzyme b,b-carotene 15,15'-monooxygenase is reported.
The substrate specificity of enriched enzyme fractions isolated from chicken intestinal mucosa was investigated.
Regarding substrate binding/cleavage, these experiments demonstrate that i) any deviation from the ꢀrod-likeꢁ
b,b-carotene structure is not tolerated, ii) one ꢀnaturalꢁ, unsubstituted b-ionone ring is required, iii) the position
and presence of the Me groups attached to the polyene chain is significant. These results suggest a hydrophobic
barrel-like substrate binding site in which the proteinꢁs amino acid residues through interaction with the Me
groups, direct the central CˆC bond in binding distance to the active siteꢁs metal-oxo center, supporting the
unique regiospecificity of cleavage to retinal (provitamin A).
1. Introduction. ± b,b-Carotene (1) is the parent structure and most important
compound of the ꢀorange pigments of lifeꢁ that comprise more than 650 carotenoids,
which are abundant in photosynthetic bacteria, and in the plant and animal kingdoms
[1][2]. Both plants and bacteria can biosynthesize carotenoids [3], but mammals rely
on extraction from the diet [4]. The significance of b,b-carotene (1) in humans concerns
its antioxidant activity and its enzymatic conversion to retinal (2) (pro-vitamin A). To
date, two modes of cleavage of 1 have been proposed: the central cleavage of 1, which
provides two moles of 2 [5], and the more recently discovered excentric cleavage which
yields first apocarotenals, such as 8-apocarotenal (3) that may be degraded to 2 by b-
oxidation (Scheme 1) [6].
Central cleavage of 1 seems to be the most important metabolic pathway, and
enzymatic activity has been detected in various tissues since its discovery in the mid-
1950s. Since then many attempts to purify and characterize this enzyme have failed.
Nevertheless, and despite the lack of solid information regarding the enzymatic
reaction mechanism or the nature of the co-factor, the enzyme was termed b,b-carotene
15,15'-dioxygenase (EC 1.13.11.21).
Only very recently have we been able to establish a protocol for purification of the
enzyme isolated from chicken intestinal mucosa, which led to the identification of the
catalytically active protein [7]. The enzyme was enriched 226-fold to a specific activity
1
of 2500 pmol ´ h 1 ´ (mg protein) , and the band eluting from the final gel filtration that
best correlated with b,b-carotene 15,15'-monooxygenase activity (Fig. 1) was se-
quenced, and complimentary oligonucleotides were designed.
With RT-PCR, a cDNA fragment was obtained, labelled, and subsequently used to
screen a chicken duodenal expression library. The isolated full-length cDNA of b,b-
carotene 15,15b-monooxygenase contains a sequence of 1578 bp that codes for a

2302
Helvetica Chimica Acta ± Vol. 84 (2001)
Scheme 1. Metabolism of b,b-Carotene (1)
3'
15
13
8'
13'
15'
7'
3
1
central cleavage
CHO
excentric
cleavage
2
CHO
3
protein of 526 amino acids. Expression of the hexa-His-tailed protein in E. coli and
BHK (baby hamster kidney) cells gave, after affinity chromatography, a catalytically
active, cytosolic enzyme (60.3 kDa) that cleaves b,b-carotene (1) to retinal (2) as the
only reaction product [8]. At about the same time, another research group published
the expression of the b,b-carotene 15,15'-monooxygenase derived from Drosophila
melanogaster [9].
Parallel to our efforts to purify the native protein [7] and to determine the reaction
mechanism [10], we developed a synthetic supramolecular enzyme model that
catalyzes carotenoid cleavage [11][12], and we also investigated the substrate
specificity of b,b-carotene 15,15'-monooxygenase from chicken intestinal mucosa using
enriched enzyme fractions. We wish to report here the synthesis of several potential
substrates and incubation experiments with 14 substrate analogues.
2. Results and Discussion. ± 2.1. Synthesis of Substrates. The surprisingly
regioselective cleavage of b,b-carotene was often attributed to the central CˆC bond of
b,b-carotene being the sterically least-hindered compared to all other CˆC bonds of 1
[13]. To investigate this point, we decided to synthesize the norcarotenoids 4 and 5, such
that the enzyme could cleave three double bonds of 4, and five double bonds of 5 (Fig. 2).
The synthesis of 20,20'-dinor-b,b-carotene (5) by a C18 ‡ C2 ‡ C18 ˆ C38 approach
has been reported earlier [14]. Synthesis of 4 and 5 by C20 ‡ C19 ˆ C39 and C19 ‡ C19 ˆ
C38 strategies is claimed in an US patent [15], however, with low yields and rather
scarce experimental data. The ready availability of the C15-Wittig salt 6 [16][17] and the
prospect to prepare additional analogues with different end groups prompted us to
synthesize 4 and 5 by Wittig condensation reactions according to the building principle

Helvetica Chimica Acta ± Vol. 84 (2001)
2303
Fig. 1. Identification of the enzyme (see Dioxy25) that catalyzes the central cleavage of 1. Upper panel: SDS gels
of fractions of the final gel filtration chromatography. Lower panel: activity profile of the corresponding
fractions.
C15 ‡ C9 ‡ C15 ˆ C39 and C15 ‡ C8 ‡ C15 ˆ C38 with the dialdehydes 7 and 8 as central
components (Scheme 2). The symmetrical C8-dialdehyde 8 was readily prepared from
1,1,4,4-tetramethoxybut-2-ene and divinyl ether by published procedures [18].
The synthesis of a mixture of the (all-E)- and of (2E,4Z,6E)-stereoisomers of C9-
dialdehyde 7 (Scheme 3) could be achieved by Wittig condensation of the protected C5-
aldehyde Wittig salt 9 [19] with fumaraldehyde mono(dimethyl acetal) 10 [20][21]
followed by Amberlyst-15-catalyzed deacetalization [22] of the intermediate diacetal.
Repeated crystallization of (2E,4E/Z,6E)-7 yielded the higher-melting (2E,4Z,6E)-7
stereoisomer. The corresponding (all-E)-7 isomer was obtained by an analogous route
starting from the six-membered-ring acetal phosphonium salt 11 [23], taking advantage
of the nicely crystalline intermediate (all-E)-12 (Scheme 3). The reaction of (all-E)-7
with the phosphonium salt 6 gave an isomeric mixture that isomerized in hot EtOH to

2304
Helvetica Chimica Acta ± Vol. 84 (2001)
4
5
Fig. 2. Two substrate analogues lacking one (4) or two (5) Me groups in the polyene chain
Scheme 2. Synthesis of Norcarotenoids 4 and 5 with the Wittig Salt 6 as a Common Precursor
O
O
5
4
8
+
PPh3Cl
2
O
O
6
7
furnish (all-E)-20-nor-b,b-carotene (4) (95% pure by HPLC). In contrast, isomer-
ization of the (11Z)-, (11Z,11'Z)-, and (all-E)-isomeric mixture of bisnorcarotenoid 5,
obtained by condensation of 6 and 8, proved to be difficult; in particular structure
elucidation of the three isomers of 5 was not facile. Therefore, a protocol for the
separation of this mixture by HPLC was developed (see Exper. Part), and the 1H-NMR
spectra of the isolated isomers were compared with those measured by direct LC-NMR
analysis. Finally, a sample of (all-E)-5 was obtained (97% pure by HPLC) for
incubation experiments.
Another aim of this study was to investigate whether the enzyme b,b-carotene
15,15'-monooxygenase tolerates deviations from the ꢀrod-likeꢁ structure of the standard
substrate. This can be achieved by replacing the central (E)-CˆC bond of 1 with, e.g., a
(Z)-configured CˆC bond, a triple bond, or a single bond. These modifications (see
Fig. 3) result in bent carotenoids, which, if not cleaved, could be inhibitors of the
enzyme. The (15Z)-b,b-carotene (13) as well as 15,15'-dehydro-b,b-carotene (14) were
available, while for 15,15'-dihydro-b,b-carotene (15), a synthetic procedure had to be
developed. First attempts to employ McMurry conditions [24] for the reductive
coupling of retinol failed, thus we decided on a double-Wittig reaction of 6 [25][26]
with the C10-dialdehyde 16, which was prepared from 17 via 18 and 19 according to [27]
in 15% overall yield (Scheme 4).

Helvetica Chimica Acta ± Vol. 84 (2001)
2305
Scheme 3. Synthesis of (all-E)-7, a Building Block for the Preparation of 20-Nor-b,b-carotene (4)
HO
O
PPh3Cl
PPh3Cl
9
H3CO
PPh)3Cl
H3CO
OCH3
OCH3
O
O
10
PPh3Br
+
O
11
10
OCH3
OCH3
H3CO
H3CO
OCH3
O
OCH3
O
O
O
O
O
(E/Z)-7
O
O
12
O
O
(all-E)-7
O
(2E,4Z,6E)-7
The Wittig reaction furnished a mixture of (all-E)-15, (11Z)-15, and (11Z,11'Z)-15,
which could be separated by semiprep. HPLC to yield 95.8% pure (all-E)-15
distinguishable by UV, MS, and 1H-NMR analysis from the two isomers. Isomerization
of (11Z)-15, and (11Z,11'Z)-15 ! (all-E)-15 under various conditions (I2 in heptane, or
photochemically) proved to be impossible.
The third aim of this work was to investigate whether modifications of the b-ionone
substructure of b,b-carotene are tolerated by the enzyme b,b-carotene 15,15'-mono-
oxygenase. For this purpose, we chosed easily available carotenoids (see below). This
aspect is of particular interest not only because we gain information on the active site,
but also because we expected to identify a nonsymmetric substrate analogue that could
be used to determine the enzymatic reaction mechanism [10]. It is important to note

2306
Helvetica Chimica Acta ± Vol. 84 (2001)
Scheme 4. Synthesis of 15,15'-Dihydro-b,b-carotene ((all-E)-15)
Br
BrPh3P
PPh3Br
Br
17
18
19
16
O
O
6
11
11'
(all-E)-15
that 18O2 experiments with b,b-carotene (1) as a substrate are not suitable to distinguish
between a monooxygenase and a dioxygenase mechanism since two identical molecules
are produced.
2.2. Incubations with Substrate Analogues of b,b-Carotene 15,15'-Monooxygenase.
Incubations were performed with enzyme fractions obtained from chicken intestinal
mucosa. The pellet from the 20 ± 458 (NH4)2SO4 fractionation was used to cleave b,b-
carotene (1) to retinal (2) in 30% yield [7] (378, pH 7.8, 3 h). Apo-b-carotenals that
could be derived from excentric cleavage were not observed. The HPLC system (diode-
array detection, internal standard) was optimized for clear separation of every possible
apocarotenyl metabolite and retinoic acid. Retinoic acid was detected only in the 45 ±
60% (NH4)2SO4 fraction, and it was shown in control experiments that this metabolite
derives from retinal [7]. The yield given for retinal does not include the fraction that is
converted to retinoic acid, hence the figures represent the lowest limit of enzymatic
carotenoid cleavage.
The incubation of (all-E)-20-nor-b,b-carotene (4) gave 45% yield (relative to 1) of
retinal and equal amounts of nor-retinal while, under the same conditions, the bis-nor-
carotenoid 5 was metabolized only 2% (Table 1). These results indicate the significance
of the Me groups of the polyene chain for substrate binding. Further, these experiments
show that the enzyme b,b-carotene 15,15'-monooxygenase is regiospecific with respect
to the central double bond even with substrate 4, which has three CˆC bonds of similar
reactivity and steric environment to the active site.
On incubation of the dihydro-b,b-carotene (all-E)-15, slow oxidation to b,b-
carotene (1) was observed followed by cleavage to retinal (2) (Table 2); 13 and 14 were
not substrates for b,b-carotene 15,15'-monooxygenase.
The three compounds (see Table 2) were also not competitive inhibitors of the
enzyme, in contrast to earlier reports for 14 [28]. Fig. 3 shows the structures of the three
substrates calculated with the TITAN force-field package; it is obvious that (all-E)-15
in its preferred diaxial conformation at the central C C bond displays ꢀa stepꢁ within
the carotenoid structure that leads to two separated chromophores.

Helvetica Chimica Acta ± Vol. 84 (2001)
2307
Table 1. Enzymatic Formation of Retinal (2) from 1 and Substrate Analogues 4 and 5
Substrate
Yield [%]
of retinal (2)
100a)
45b)
2b)
a) Absolute yield 30%. b) Relative to yield from 1, set at 100% for convenient comparison.
The (Z)-configuration of the central CˆC bond yields a completely bent
carotenoid (see 13), while the 15,15'-dehydrocarotene 14 is less bent, but the distance
between the Me groups at C(13) ! C(13') is larger by 1.5 Š than in 1. These results
suggest that the substrate binding pocket of the enzyme does not tolerate substrates
displaying any deviation from the ꢀrod-likeꢁ structure of b,b-carotene (1).
Results from incubations of substrates that have end groups other than those found
in 1 are shown in Table 3. It is interesting to note that substrates with only one b-ionone
ring substituted with a polar group like cryptoxanthin (20) or 3-oxoechinenone (21) are
readily cleaved to retinal, whereas zeaxanthin (22) is not a substrate, and a-carotene
(23), having a distorted end group, is only a poor substrate. When one b-ionone ring is
replaced by a more polar head, as in 2'-dehydroplectariaxanthin 24, only traces of
retinal were detected. These experiments suggest that the substrate binding site is not
uniform for both end groups, such that a substrate having polarity in one b-ionone ring
and the overall shape of the native substrate b,b-carotene is tolerated. In agreement
with these results are the results of incubations of the enzyme with canthaxanthin (4,4'-
dioxo-1), astaxanthin (3,3-dihydroxy-4,4'-dioxo-1), and lutein (3R,3'R-dihydroxy-23);
in all cases no cleavage to form retinal was detected.
In summary, the enzyme b,b-carotene 15,15'-monooxygenase has a rather rigid
binding site in which the substrate-protein interaction determines the absolute
preference to cleave the central C(15)ˆC(15') bond of those carotenoids that are at
least ꢀhalf-identicalꢁ to the native substrate 1.

2308
Helvetica Chimica Acta ± Vol. 84 (2001)
Table 2. Enzymatic Formation of Retinal (2) from the Structurally Diverse Carotenoids (all-E)-15, 13, and 14
Substrate
Yielda) [%]
of retinal (2)
15b)
0
0
a) Relative to yield from 1, set at 100%. b) Cleavage observed only subsequent to oxidation to 1.
Experimental Part
General. All reactions were performed under Ar. Column chromatography (flash chromatography, FC)
was performed with silica gel (SiO2, 0.040 ± 0.063 mm, 230 ± 400 mesh). HPLC: Hewlett-Packard HPLC system
Series 1050 with diode-array detector and Hewlett-Packard Chemstation. Method 1: Supelcosil LC-18 column,
250 Â 4.6 mm, 5 mm (Supelco), MeCN/H2O 25 :75 (v/v), flow 1.5 ml/min, 340 nm. Method 2: YMC Carotenoid
C30 column, 250 Â 4.6 mm, 5 mm (Stagroma), acetone, flow 1 ml/min, monitoring wavelength 450 nm; anal.
isolation of several isomers were carried out by this method (10-ml injection of a 60-mg/ml carotenoid soln.).
Method 3: YMC Silica 200A column, 250 Â 4.0 mm, 3 mm (Stagroma), hexane/0.02% (v/v) Et3N, flow 2 ml/min,
monitoring wavelength 400 nm. Method 4: for LC-NMR coupling: Prontosil C30 column, 250 Â 4.0 mm, 3 mm
1
(Bischoff), acetone/D2O 96 :4, flow 1 ml/min, monitoring wavelength 450 nm. UV: l in nm (e). IR: nÄ in cm
.
1H-NMR: Chemical shifts (d) in ppm, coupling constants J in Hz. EI-MS: Finnigan MAT SSQ-7000, direct
introduction, 70 eV, ion-source temp. 2408.
Enzymatic Reactions. All operation were carried out on ice, and refrigerated centrifuges (48) were used. All
chemicals were purchased in highest available purity from either Fluka AG (CH-Buchs), Sigma, Aldrich, or Bio-
Rad. Standards of carotenoids and retinoids were generous gifts from F. Hoffmann-La Roche, Basel.
Animals. Roughly four-week-old fattened chickens (Vedette) or 22-week-old laying hens (LSL Lohmann)
fed on a diet low in pigments (Kliba 556 SP18 feed) for two weeks were killed by decapitation. The first third of

Helvetica Chimica Acta ± Vol. 84 (2001)
2309
Fig. 3. Structures of carotenoids 13, 15, 14, and 1 generated by molecular-modelling studies (TITAN force-field
calculations)

2310
Helvetica Chimica Acta ± Vol. 84 (2001)
Table 3. Enzymatic Formation of Retinal (2) from Carotenoids 21 ± 25 with One or Both End Groups Modified
Substrate
Yielda) [%]
of retinal (2)
40
45
0
8
2
a) Relative to yield from 1, set at 100%.
the small intestine was removed, separated from the pancreas with scissors, and flushed with 0.9% NaCl soln.
The intestines were immediately placed on dry ice and stored at 808 until further use.
Enzyme Preparation. The intestine was thawed on ice and cut lengthwise. The mucosa was scraped off with
a glass slide and homogenized with a Teflon pestle (Heidolph RZR 2020, 6 passes at 200 rpm) in 4 vol of 0.1m
K3PO4 phosphate buffer, pH 7.8, containing 4 mm MgCl2 and 30 mm nicotinamide, and 5 mm soybean trypsin
inhibitor, 10 mm 6-aminocaproic acid, and 5 mm benzamidine as a protease inhibitor. After centrifugation for
1 h at 62000 g (Heraeus sepatech Suprafuge 22), the supernatant was subjected to 20 ± 45% (NH4)2SO4
fractionation (20 ± 45% AS). The pellet obtained by centrifugation (10 min at 5000 g (IEC Centra-7R)) was
stored in aliquots at 808 until further use.

Helvetica Chimica Acta ± Vol. 84 (2001)
2311
Incubation Procedure. In a standard incubation, the 20 ± 45% AS precipitate was redissolved in a small
volume (15 ml) of 0.1m phosphate buffer, pH 7.8, containing 4 mm MgCl2 and 30 mm nicotinamide to give a
protein concentration of ca. 5 mg/ml (by Bradford assay), and 20 mm glutathione (reduced, GSH) in a total
volume of 10 ml. The incubation was started by adding 250 ml of a soln. containing 8.8 nmol carotenoids as
mixed micelles, which were prepared in the following manner. A stock soln. of 80 mg/ml b,b-carotene in CHCl3
(235 ml) and of 10 mg/ml a-tocopherol in hexane (500 ml) were placed in a small glass vial wrapped in aluminum
foil. After removing the solvent under a faint stream of N2, the mixture was solubilized in 1 ml of micellar soln.
(prepared from glycocholic acid and lecithin: glycocholic acid (1.16 g) was dissolved in 5 ml of H2O under
stirring and by dropwise addition of 5n NaOH. After the pH was adjusted to 6.8 ± 7.2 with AcOH and the
volume brought to 10 ml with H2O, 80 mg of acolectin was added and dissolved under stirring). This b,b-
carotene micellar soln. was agitated on a vortex mixer and kept in the dark until use. Incubations were carried
out in reaction vials in a shaking water bath at 378 in the dark for 3 h. Control incubations without added
carotinoid or without added enzyme preparation were performed regularly.
Extraction and HPLC. The incubation was stopped by cooling on ice for 5 min and quenching with 6.5 ml of
MeCN, followed by vortexing. Extraction (3Â) with CHCl3 was performed by vortexing and centrifuging for
10 min at 5000 g at 48. The CHCl3 layer was removed, concentrated in vacuo, and the residue was redissolved in
200 ml of hexane/CHCl3 9 :1 containing 10 ng/ml retinyl acetate as internal standard. The final extract was passed
through a 0.45-mm mini-filter (Nylon, Schleicher & Schuell), and a 20-ml aliquot was analyzed by HPLC
(Hewlett-Packard 1100 system; Lichrospher 100-RP-18, 5 mm, 12.5 cm  4.6 mm (Bischoff); 258). The mobile
phase was MeCN/THF/H2O (‡1% ammonium acetate) (solvent A, 50 :20 :30 (v/v/v), or solvent B, 50 :44 :6 (v/
v/v)) at a flow rate of 1 ml/min; the elution program was 3 min at 100% A, followed by a 6-min linear gradient to
60% B, 12 min at 60% B, a 1-min linear gradient to 100% B, 9 min hold at 100% B, then a 7-min gradient back to
100% A. This program allowed separation of apocarotenals, which could have been formed after excentric
cleavage. Retinyl acetate was used as internal standard.
For analysis of cryptoxanthin, the gradient was changed to delay elution of the 3-hydroxyretinal by having
an initial 8 min of mobile phase C (solvent C, MeCN/THF/(H2O ‡ 1% ammonium acetate) 40 :10 :50 v/v/v)
before changing to 100% A in 3 min; the same gradient as described above was used for the remainder of the
elution. Detection with the photo-diode-array detector allowed collection of UV spectra (300 ± 550 nm). The
detector was set at 340 and 380 nm for retinoids, and 450 nm for carotenoids. Peaks were identified by coelution
with standard retinoids and quantified relative to the internal standard, with determination of peak areas
calibrated against known amounts of standards.
(2E,4E/Z,6E)-2-Methylocta-2,4,6-trienedial ((2E,4E/Z,6E)-7). To a stirred mixture of Wittig salt 9 (11.4 g,
30 mmol), methyl orthoformate (4.13 g, 4.28 ml, 39 mmol), and MeOH (11 ml) was added TsOH (70 mg,
0.366 mmol) at r.t. The mixture became a red-brown soln. and was stirred for 40 min; i-PrOH (75 ml), a few
drops (ca. 0.1 ml) of 5.27n MeONa in MeOH, fumaraldehyde mono(dimethyl acetal) (10; 4.15 g, 30 mmol)
were added, and, finally, over a period of 30 min, 5.27n MeONa in MeOH (2.3 ml, 12.1 mmol) was added. After
75 min, the mixture was poured on ice and conc. NaCl soln. and extracted with AcOEt (2 Â 200 ml). The
combined org. layers were washed with NaCl soln. (3 Â 50 ml), dried (Na2SO4), and concentrated in vacuo to
give 18.4 g of an orange semisolid. This crude product was taken up in 150 ml of hexane/AcOEt/Et3N 500 :500 :1
(v/v/v) and filtered through 300 g of silica gel, eluting with the same solvent mixture (1700 ml) to give crude (E/
Z)-condensation product as a yellow-orange oil (7.6 g). Acetone (160 ml), H2O (9.5 ml) and Amberlyst-15
(6.2 g) were added, and the mixture was stirred for 20 min at r.t. After 90 min, the soln. was decanted (the
residual Amberlyst-15 was washed with total 200 ml of CH2Cl2) and poured on ice/NaHCO3 soln. The CH2Cl2
layers were washed with NaHCO3 soln., dried (Na2SO4), and concentrated in vacuo to yield (E/Z)-7 (4.7 g).
Orange solid. HPLC (Method 1): 46.3% (all-E)-7, 46.3% (2E,4Z,6E)-7.
(all-E)-7-(5,5-Dimethyl[1,3]dioxan-2-yl)octa-2,4,6-trienal (12). Obtained from [3-(5,5-dimethyl[1,3]diox-
an-2-yl)but-2-enyl]triphenylphosphonium bromide (11) and fumaraldehyde mono(dimethyl acetal) (10)
according to the procedure described for the synthesis of (E/Z)-7. Compounds 12 crystallized from hexane as
light yellow needles. M.p. 88 ± 908. UV (EtOH) 320 (337). IR (KBr): 1675, 1657, 1609, 1152, 1098. 1H-NMR
(400 MHz, CDCl3): 9.57 (s, CHO); 7.18 (dd, J ˆ 15.2, 11.2, H C(3)); 6.93 (dd, J ˆ 14.8, 10.9, H C(5)); 6.49
(dd, J ˆ 11.2, 14.8, H C(4)); 6.36 (d, J ˆ 10.9, H C(6)); 6.16 (dd, J ˆ 7.9, 15.2, H C(2)); 1.93 (s, Me C(7)); 4.8
(s, H C(8)); 3.52 (s, H C(9), H C(11)); 3.68 (d, J ˆ 10.9); 1.23 (s, Me C(5')); 0.76 (s, Me C(5')). MS: 236
.
‡
(100, M ), 221 (16), 207 (57), 168 (21), 150 (21), 142 (66), 212 (65). Anal. calc. for C14H20O3: C 71.16, H 8.53;
found: C 71.05, H 8.48.
(all-E)-2-Methylocta-2,4,6-trienedial ((all-E)-7). Obtained by deacetalization of (all-E)-12 (2m H2SO4,
MeOH, 508). Crystallization from Et2O/hexane at 08. Light yellow powder. M.p. 84 ± 878. IR (KBr): 1676, 1619,

2312
Helvetica Chimica Acta ± Vol. 84 (2001)
1412, 1383, 1289, 1158, 1146, 1106, 1012, 995, 838, 617. 1H-NMR (400 MHz, CDCl3): 9.67 (d, J ˆ 7.8, H C(8));
9.54 (s, H C(1)); 7.26 (dd, J ˆ 11.2, 15.3, H C(6)); 7.11 (dd, J ˆ 11.6, 14.8, H C(4)); 6.95 (d, J ˆ 11.6,
.
‡
H
C(3)); 6.85 (dd, J ˆ 14.8, 11.2, H C(5)); 6.33 (dd, J ˆ 15.3, 7.8, H C(7)); 1.95 (s, Me). MS: 150 (75, M ),
‡
121 (100, [M CHO] ), 107 (13), 93 (25), 91 (48), 82 (20), 79 (25), 77 (60).
(2E,4Z,6E)-2-Methylocta-2,4,6-trienedial ((2E,4Z,6E)-7). Obtained from (2E,4E/Z,6E)-7 by repeated
cystallization from t-BuOMe/hexane. Light yellow crystals. M.p. 95 ± 978. HPLC (Method 1): 99.4%; 0.6% of
(all-E)-7. UV (2% CH2Cl2, cyclohexane): 328 (2320), 313 (2703), 207 (5283). IR (KBr): 1682, 1669, 1610, 1234,
1197, 1157, 1124, 1002, 986. 1H-NMR (400 MHz, CDCl3): 9.74 (d, J ˆ 7.6, H C(8)); 9.62 (s, H C(1)); 7.7
(dd, J ˆ 11.4, 15.2, H C(6)); 7.4 (d, J ˆ 10.9, H C(3)); 6.89 (dd, J ˆ 10.9, 11.4, H C(4)); 6.63 (dd, J ˆ 11.4, 11.4,
H
C(5)); 6.34 (dd, J ˆ 15.2, 7.6, H C(7)); 1.95 (s, Me C(2)). MS: Identical with the spectrum of (all-E)-7.
Anal. calc. for C9H10O2: C 71.98, H 6.71; found: C 71.99, H 6.61.
20-Nor-b,b-carotene (4). To the soln. of [(2E,4E)-3-methyl-5-(2,6,6-trimethylcyclohex-1-enyl)penta-2,4-
dienyl]triphenylphosphonium chloride (6; 6.013 g, 12.0 mmol), 7 (0.8 g of (2E,4E/Z,6E)-7 as described above,
content 94% by HPLC, 5.0 mmol), CH2Cl2 (40 ml), and H2O (5 ml) was added, at 378, 5.27n MeONa in MeOH
(2.3 ml, 12.1 mmol, 1 ml over 1 h, 1.3 ml over 2 h) with a Metrohm 665 Dosimat. The mixture was poured on ice
and extracted with CH2Cl2 (150 ml). The CH2Cl2 layer was washed with H2O, dried (Na2SO4), and evaporated
to yield 6.5 g of dark red crystals, which were taken up in EtOH (300 ml) and heated under reflux for 7 h. The
heating was removed, and crystallization was effected by stirring overnight, followed by cooling to 08. The
crystals were separated by filtration, washed with EtOH (250 ml, 08) and dried (r.t., pressure ca. 0.1 mm) to give
1.22 g (47%) of crude 4. Red solid. M.p. 133 ± 1348. HPLC (Method 3): 98.5%. Another sample of crude 4
(0.22 g, 96.7% by HPLC) was chromatographed on 300 g of silica gel (hexane/CH2Cl2 9 :1) to yield 0.16 g of red
oil, homogeneous by TLC. Crystallization from 20 ml of Et2O/EtOH 1:1 at 48 gave 0.12 g of (all-E)-4. Red shiny
plates. M.p. 154-1558. HPLC (Method 3): 95.2% (all-E)-4. UV/VIS (2% CH2Cl2, hexane): 473 (2495), 445
‡
‡
(2809), 269 (352). MS: 522 (100, M), 444 (2, [M C6H6] ), 430 (9, [M C7H8] ). Anal. calc. for C39H54
:
C 89.59, H 10.41; found: C 89.73, H 10.37.
20,20'-Dinor-b,b-carotene (5). To a soln. of 6 (6.013 g, 12.0 mmol) and (all-E)-octa-2,4,6-trienedial (8;
0.68 g, 5.0 mmol) in i-PrOH (60 ml) at 408 was added 5.27n MeONa in MeOH (2.3 ml, 12.1 mmol) over 3 h with
a Metrohm 665 Dosimat. After 1 h, the mixture was slowly cooled to 68, filtered, and dried to give crude 5 as red-
orange crystals (3.26 g, according to HPLC (Method 3) mainly a mixture of (11Z)-, (11Z,11'Z)-, and (all-E)-
isomers 24 :50 :24). FC of 0.22 g of the above mixture of stereoisomers on 300 g silica gel (hexane/CH2Cl2 9 :1)
yielded 120 mg of a viscous oil (Rf (silica gel; hexane/CH2Cl2 9 :1, 0.30), which was crystallized from 20 ml of
Et2O/EtOH 1:1 (overnight at 48) to give 42 mg of (all-E)-5. Shiny red-brown crystals. M.p. 183 ± 1848. Purity
.
‡
‡
97%. UV/VIS (hexane) 468 (2656), 442 (2994), 269 (328). MS: 508 (100, M ), 430 (4, [M C6H6] ).
(11Z)-20,20'-Dinor-b,b-carotene ((11/Z)-5) and (11Z,11'Z)-20,20'-Dinor-b,b-carotene ((11Z,11'Z)-5).
Structure elucidation of the isomers (11Z)-5 and (11Z,11'Z)-5 could be achieved either after repeated
2
chromatographic fractionation of crude 5 by HPLC (Method 2) and subsequent analysis by off-line D-NMR
experiments or by on-line stopped-flow HPLC/NMR coupling (HPLC, Method 4). Chromatography on C30-
columns yielded an excellent separation with very good selectivity for all occurring isomers (Fig. 4). UV/VIS
Absorption maxima recorded by a photodiode-array detector (Table 4) were also in agreement with the
proposed structures.
Interpretation of the 1H-NMR spectra of 5 was accomplished on the basis of the observed d and J values as
well as the 1H,1H connectivities in 2D-COSY spectra (not shown here). The resulting d values for the (all-E)-
and the (Z)-isomers are collected in Tables 4 and 5. Since the LC-NMR measurements were performed with
acetone/D2O as solvent, the d values differ from those measured in CDCl3.
(Butane-1,4-diyl)bis[triphenylphosphonium Bromide] (18). The salt 18 was synthesized according to [29].
1H-NMR (250 MHz, CDCl3): 7.85 ± 8.00 (m, 12 arom. H); 7.60 ± 7.80 (m, 18 arom. H); 3.90 ± 4.15 (m, 2 H C(1),
‡
2 H C(4)); 2.15 ± 2.35 (m, 2 H C(2), 2 H C(3)). EI-MS: 580 (2), 579 (5), 578 (2, [M HBr Br] ), 291
(15), 290 (100), 289 (45, [M 2Br]2 ).
2,7-Dimethylocta-2,6-diene (19). The salt 18 (23.7 g, 32 mmol) was dissolved in 150 ml of DMSO. Then,
73 ml (116 mmol, 2.2 equivalent) of 1.6m BuLi soln. in hexane was added under cooling with an ice bath so that
the reaction temp. did not exceed 108. The deep red-orange suspension was stirred for an additional 15 min.
Then, 5.17 ml (4.09 g, 70.4 mmol, 2.2 equiv.) acetone was added dropwise, and the mixture was stirred for
another 2 h, and an additional 5.17 ml (4.09 g, 70.4 mmol, 2.2 equiv.) acetone was added. Then, a 10-cm Vigreux
column was placed on the flask, and acetone and hexane were distilled off. Subsequently, the Vigreux column
was removed and the mixture distilled at 1 atm. A fraction that distilled in the range of 100 ± 1308 was collected
to which 70 ml of hexane was added, the org. layer was washed with 10 ml of H2O and dried (Na2SO4). After

Helvetica Chimica Acta ± Vol. 84 (2001)
2313
Fig. 4. HPLC Separation of the stereoisomers of 20,20'-dinor-b,b-carotene (5) on a C30-bonded column
(Method 4)
Table 4. UV/VIS Absorption Maxima of Stereoisomers of 5 (in acetone/D2O 96 : 4)
Stereoisomer
lmax [nm] (relative intensity)
445 (100)
(all-E)-5
471 (89)
467 (86)
462 (84)
(11Z)-5
(11Z,11'Z)-5
341 (20)
443 (100)
441 (100)
removal of the solvent, 19 (1.11 g, 25%) was isolated. Colorless oil. IR (film): 2966vs, 2923vs, 2855vs, 1660vw,
1451s, 1378s. 1H-NMR (250 MHz, CDCl3): 5.10 ± 5.20 (m, 2 olef. H); 1.95 ± 2.05 (m, 2 CH2); 1.69 (s, 2 Me), 1.61
‡
(s, 2 Me). EI-MS: 138 (6, M ), 123 (6), 95 (10), 82 (18), 70 (17), 69 (100), 67 (8), 41 (60).
(all-E)-2,7-Dimethylocta-2,6-dienedial (16) [27]. Diene 19 (497.3 mg, 3.60 mmol) was dissolved in 14 ml of
97% aq. EtOH and heated under Ar to 558. To this soln., 2.16 g (19.5 mmol, 5.4 equiv.) of SeO2 dissolved in
120 ml of 97% aq. EtOH was slowly added at such a rate that the reaction temp. was between 50 and 558. Then
the mixture was heated to reflux for 16 h under an inert atmosphere. After cooling to r.t., the mixture was
filtered through a plug of Celite, and the solvent was removed in vacuo at 258. The residue was dissolved in
250 ml of Et2O and washed with sat. aq. NaHCO3 and NaCl, then dried (Na2SO4), and the solvent was removed
to provide 626 mg of a brown oil. This oil was dissolved in 30 ml of AcOEt, and 2.0 g of MgO was added. The
suspension was stirred for 2 h at r.t. under Ar. The suspension was filtered through a plug of Celite. The filtrate
was concentrated on a rotary evaporator and the residue distilled at 0.15 mbar at 130 ± 1508 to provide 368 mg of
16. Light brown distillate. 1H-NMR (250 MHz, CDCl3): 9.43 (s, 2 CHO); 6.45 ± 6.55 (m, 2 olef. H); 2.59 (dd, J ˆ
3.6, 3.6, 2 CH2); 1.78 (s, 2 Me).
(all-E)-15,15'-Dihydro-b,b-carotene (15). Wittig salt 6 (1.76 g, 3.51 mmol, 2.2 equiv.) was dissolved in 80 ml
of EtOH, and a soln. of 264 mg (1.59 mmol) of 16 dissolved in 2.0 ml of EtOH was added. To this soln., 710 ml

2314
Helvetica Chimica Acta ± Vol. 84 (2001)
Table 5. 1H-NMR (acetone/D2O 96 : 4) Chemical Shifts (d [ppm]) of Olefinic Protons of Stereoisomers of 5
H-Atom
(all-E)-5
(11Z)-5
(11Z,11'Z)-5
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
C(7)
C(7')
C(8)
C(8')
6.12
6.12
6.04
6.04
6.08
6.08
6.62
6.62
6.29
6.29
6.34
6.34
6.30
6.30
6.30
6.30
6.16
6.16
6.12
6.12
6.6
6.10
6.31
6.62
6.04
6.29
6.85
6.34
6.31
6.31
6.37
6.30
6.16
6.16
6.12
6.12
6.6
C(10)
C(10')
C(11)
C(11')
C(12)
C(12')
C(13)
C(13')
C(14)
C(14')
C(15)
C(15')
6.6
6.31
6.31
6.05
6.05
6.85
6.85
6.31
6.31
6.37
6.37
Table 6. 1H-NMR Data (d [ppm], CDCl3) of Stereoisomers of 5
H-Atom
(all-E)-5
(11Z)-5
(11Z,11'Z)-5
Me C(1)
Me C(1')
CH2(2), CH2(2')
CH2(3), CH2(3')
CH2(4), CH2(4')
Me C(5)
1.02
1.02
1.47
1.61
2.02
1.71
1.71
1.94
6.19
6.19
6.10
6.10
6.10
6.10
6.61
6.61
6.31
6.31
6.32
6.32
6.32
6.32
6.32
6.32
1.03
1.02
1.47
1.60
2.02
1.72
1.71
1.94
6.18
6.21
6.11
6.17
6.53
6.11
6.35
6.61
6.09
6.31
6.78
6.32
6.32
6.32
6.32
6.32
1.03
1.03
1.47
1.62
2.02
1.72
1.72
1.94
6.23
6.23
6.18
6.18
6.53
6.53
6.37
6.37
6.10
6.10
6.79
6.79
6.37
6.37
6.38
6.38
Me C(5')
Me C(9), Me C(9')
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
C(7)
C(7')
C(8)
C(8')
C(10)
C(10')
C(11)
C(11')
C(12)
C(12')
C(13)
C(13')
C(14)
C(14')
C(15)
C(15')
5.4m MeONa in MeOH was slowly added under an inert atmosphere and cooling with a water bath (158). This
soln. was then stirred at r.t. for 2 h and subsequently heated to reflux for an additional 60 min. After cooling to
r.t., 300 ml of hexane were added. The org. layer was washed three times with a total of 240 ml 85% aq. MeOH
and dried (Na2SO4). After concentration, the residue was redissolved in 35 ml of hexane containing some Et2O.
The solution was then reduced to a total volume of ca. 15 ml with a rotary evaporator, cooled to 08 for 30 min
and filtered to remove precipitated Ph3PO. The filtrate was concentrated to give 823 mg of a red oil. This residue

Helvetica Chimica Acta ± Vol. 84 (2001)
2315
was chromatographed on an Alox II N column (hexane/benzene 4 :1) to give, after evaporation, 569 mg of an
orange oil. It was rechromatographed on an Alox II N column (hexane/benzene 5 :1) to yield, after evaporation,
a yellow oil (377 mg, 44%) of a mixture of three isomers of 151). Final purification was accomplished by semi-
prep. HPLC (Chromasil C8, 10 mm, MeOH/hexane 9 :1) to yield 5.6 mg (95.8% pure by HPLC) of 15. Slightly
yellow, amorphous solid. UV (cyclohexane), 332 (e ˆ 68700, A11 ˆ 1280). 1H-NMR (400 MHz, CDCl3): 6.50
(dd, J ˆ 11.2, 15.2, H C(11), H C(11')); 6.27 (d, J ˆ 15.2, H C(12), H C(12')); 6.11 (s, H C(7), H C(7'),
H
C(8), H C(8')); 6.08 (d, J ˆ 11.2, H C(10), H C(10')); 5.50 ± 5.55 (m, H C(14), H C(14')); 2.26
C(15), C(15')); 1.99 ± 2.03 (m, 2 H C(4), 2 H C(4')); 1.94 (s, 3 H C(19),
(2d, J ˆ 3.6, 3.6,
H
H
3 H C(19')); 1.81 (s, 3 H C(20), 3 H C(20')); 1.71 (s, 3 H C(18), 3 H C(18')); 1.60 ± 1.62 (m, 2 H C(3),
2 H C(3')); 1.47 (dd, J ˆ 6.0, 2.8, 2 H C(2), 2 H C(2')); 1.02 (s, 3H C(16), 3 H C(16'), 3 H C(17),
3 H C(17')). EI-MS: 538.6 (6, M ), 378.4 (6), 270.2 (20), 269.2 (100, M2‡), 145.0 (8), 119.0 (15), 95.1 (16), 93.1
‡
(13), 82.1 (15), 69.2 (10).
REFERENCES
[1] O. Straub, in ꢀKey to Carotenoidsꢁ, Ed. H. Pfander, Birkhäuser Verlag, Basel-Boston, 1987.
[2] ꢀCarotenoidsꢁ, Vol. 1a: Isolation and Analysisꢁ, Eds. G. Britton, S. Liaaen-Jensen, H. Pfander, Birkhäuser
Verlag, Basel, Boston, Berlin, 1995.
[3] ꢀCarotenoids, Vol. 3: Biosynthesis and Metabolismꢁ, Eds. G. Britton, S. Liaaen-Jensen, H. Pfander,
Birkhäuser Verlag, Basel, Boston, Berlin, 1998.
[4] R. Edge, T. G. Truscott, The Spectrum 2000, 12.
[5] J. A. Olson, J. Nutr. 1989, 119, 105.
[6] X.-D. Wang, G.-W. Tang, J. G. Fox, N. I. Krinsky, R. M. Russel, Arch. Biochem. Biophys. 1991, 285, 8.
[7] G. Wirth, Ph.D. Thesis, University of Basel, 1998.
[8] A. Wyss, G. Wirtz, W.-D. Woggon, R. Brugger, M. Wyss, A. Friedlein, H. Bachmann, W. Hunziker,
Biochem. Biophys. Res. Commun. 2000, 271, 334.
[9] J. von Lintig, K. Vogt, J. Biol. Chem. 2000, 275, 11915.
[10] M. Leuenberger, C. Jarret, W.-D. Woggon, Angew. Chem., Int. Ed. 2001, 40, 2613.
[11] R. R. French, J. Wirz, W.-D. Woggon, Helv. Chim. Acta 1998, 81, 1521.
[12] R. R. French, P. Holzer, M. G. Leuenberger, W.-D. Woggon, Angew. Chem., Int. Ed. 2000, 39, 1267.
[13] B. V. Milborrow, J. Devery, Br. J. Nutr. 1994, 72, 397.
[14] H. H. Inhoffen, F. Bohlmann, G. Rummert, Liebigs Ann. Chem. 1950, 569, 226.
[15] A. J. Chechak, C. D. Robeson, US Pat. 3,469,170 to Eastman Kodak Co., 1966.
[16] W. Sarnecki, H. Pommer, Ger. Patent 1060386 to BASF AG, 1957.
[17] H. J. Bestmann, P. Ermann, H. Rüppel, W. Sperling, Liebigs Ann. Chem. 1986, 479.
[18] S. M. Makin, N. I. Telegina, J. Gen. Chem. USSR 1962, 32, 1082 (translated from Zhurnal Obshchei Khimii
1962, 32 (4), 1104).
[19] H. Freyschlag, W. Reif, A. Nürrenbach, W. Sarnecki, US Pat. 3,449,438 to BASF AG, 1969.
[20] C. M. Cox, D. A. Whiting, J. Chem. Soc., Perkin Trans. 1 1991, 1907.
[21] H. Laatsch, A. Pudleiner, J. Prakt. Chem. 1994, 336, 663.
[22] G. M. Coppola, Synthesis 1984, 1021.
[23] J. Paust, W. Reif, H. Schumacher, Liebigs Ann. Chem. 1976, 2194.
[24] J. E. McMurry, M. Silvestri, J. Org. Chem. 1975, 40, 2687.
[25] W. Sarnecki, H. Pommer, Ger. Pat. 1060386 to BASF AG, 1959, US Pat. 2950321 to BASF AG, 1960.
[26] F. Lederer, M. Lalonde, Roche, US Pat. 4026778 to Roche, 1976.
[27] J. J. Plattner, U. T. Bhalero, H. Rapoport, J. Am. Chem. Soc. 1969, 91, 4933; U. T. Bhalero, J. J. Plattner, H.
Rapoport, J. Am. Chem. Soc. 1970, 92, 3429; U. T. Bhalero, H. Rapoport, J. Am. Chem. Soc. 1971, 93, 5311.
[28] A. A. Dmitrovskii, in ꢀNew Trends in Biological Chemistryꢁ, Ed. T. Ozawa, Springer Verlag, New York,
1991.
[29] L. Horner, H. Hoffmann, W. Klink, H. Ertel, V. G. Toscano, Chem. Ber. 1962, 95, 581.
Received March 9, 2001
1
)
The lmax of the different isomers, determined by a diode-array UV detector, were 334 for the (all-E)-, 328
for the (11Z)-, and 321 nm for the (11Z,11'Z)-15,15'-dihydro-b,b-carotene.