Conformational analysis and selection of odor-active conformers: Synthesis of molecules designed for the lily-of-the-valley(Muguet)-type odor

Article abstract of DOI:10.1002/1522-2675(20010711)84:7<2051::AID-HLCA2051>3.0.CO;2-H

The selection of odor-active conformers and the construction of a model for a targeted odor type, i.e., for the lily-of-the-valley odor, were examined. The disagreement of the odors of 1,3,4,5-tetrahydro-2-benzoxepin derivative 1 and 3-[4-(tert-butyl)phen

Full text of DOI:10.1002/1522-2675(20010711)84:7<2051::AID-HLCA2051>3.0.CO;2-H

Helvetica Chimica Acta ± Vol. 84 (2001)
2051
Conformational Analysis and Selection of Odor-Active Conformers:
Synthesis of Molecules Designed for the Lily-of-the-Valley(Muguet)-Type
Odor
by Fumiko Yoshiia)*), Tomoko Nakamurab), Shuichi Hironoc), Yumiko Shimizub),Takashi Hoshib),
Masayoshi Andob), and Hisahiro Hagiwaraa)*
a) Graduate School of Science and Technology, Niigata University, 8050 Ikarashi 2 nocho, Niigata 950-2181,
Japan
b) Faculty of Engineering, Niigata University, 8050 Ikarashi 2 nocho, Niigata 950-2181, Japan
c) Division of Physical Chemistry for Drug Design, School of Pharmaceutical Sciences, Kitasato University,
5-9-1 Shirokane, Minato-ku, Tokyo 108-0072, Japan
The selection of odor-active conformers and the construction of a model for a targeted odor type, i.e., for
the lily-of-the-valley odor, were examined. The disagreement of the odors of 1,3,4,5-tetrahydro-2-benzoxepin
derivative 1 and 3-[4-(tert-butyl)phenyl]-2-methylpropanal (2) is discussed in terms of their stable conformers.
The conformer active for the lily-of-the-valley odor was investigated by conformational analyses of several
related compounds. Based on the integrated model consisting of the assumed active conformers (Fig. 5),
compounds anticipated to possess the lily-of-the-valley odor were designed and synthesized. The odor of
synthetic 7-(tert-butyl)-1,2,4,5-tetrahydro-3H-benzocyclohepten-3-one (8) and 3-[4-(tert-butyl)phenyl]cyclo-
pentanone (13) were evaluated by perfumers to have a floral odor and to recall the lily-of-the-valley and lilac
odors, respectively. Our methodology to design new odoriferous compounds, based on conformational analysis,
selection of odor-active conformers, and construction of a model, proved to be satisfactory.
1. Introduction. ± Recent developments of odorant chemistry promoted a study for
structure-odor relationships (SOR) as described in comprehensive reviews by Rossiter
and by Chastrette [1][2]. At present, the ultimate goal of SOR studies is beyond our
reach, because odor types or odor categories are numerous, and understanding as well
as data concerning odor are still limited. However, elucidation of relationships between
odorant property and an odor type is significant to predict odor-reception mechanism
and to give indications for the synthesis of odoriferous compounds of current interest.
In triggering the specific odor sensations in human beings, it was assumed that ꢀan
odor-active conformerꢁ of the odoriferous molecule can fit and interact with an odor-
specific receptor protein. It is highly likely that there is a type of specific olfactory
receptor protein that specifically relates to the lily-of-the-valley-type odorant. This
assumption is based on the results by Buck and Axel [3] that putative olfactory receptor
proteins exist, and by Raming et al. that lyral (ˆ4-(4-hydroxy-4-methylpentyl)cyclo-
hex-3-ene-1-carbaldehyde), which has the odor of the lily of the valley, increases the
amount of the second messenger InsP3 at the OR5 putative olfactory receptor protein
[4]. Molecular modeling of the interaction between lyral and its receptor in the OR5
has been carried out by Singer and Shepherd [5], though specific interactions between
receptor proteins and the odorant have not yet been established. Therefore, it would be
promising to find the odor-active conformer of the lily-of-the-valley odorant, to

2052
Helvetica Chimica Acta ± Vol. 84 (2001)
construct a working model for odorants that have this specific odor, and to apply this
model to the design of new compounds.
Skouroumounis and Winter synthesized the 1,3,4,5-tetrahydro-2-benzoxepin deriv-
ative 1 as a conformationally restricted analogue of 3-[4-(tert-butyl)phenyl]-2-
methylpropanal (2; Lily Aldehydeꢂ, trade name by Soda Aromatic Co., Ltd.) with
the aim to develop new lily-of-the-valley-odor compounds [6] (Fig. 1). They examined
the relevance of the compact folded conformation of 2 to the lily-of-the-valley odor.
However, it was found that synthetic benzoxepin 1 had not even the slightest lily-of-
the-valley odor [6].
Fig. 1. 1,3,4,5-Tetrahydro-2-benzoxepin derivative 1 and LilyAldehyde ꢂ (2). Benzoxepin 1 is a conformation-
ally restricted analogue of aldehyde 2. The dihedral angles of the numbered bonds were changed for
conformational analysis.
In this work, the conformational analyses of benzoxepin 1 and aldehyde 2 were
carried out to re-examine this disagreement of the odor of 2 and of the proposed
conformationally restricted 1. Assumed conformers active for the lily-of-the-valley
odor were selected by conformational analyses of related compounds. Based on the
thus obtained model consisting of the odor-active conformers, we designed new
odoriferous compounds, synthesized them, and evaluated their olfactory properties.
2. Results and Discussion. ± 2.1. Computational Analysis. The (R)-aldehyde 2 was
constructed as an initial structure for computational analysis, because it has been
reported that the odor of the ( )-(R)-enantiomer is stronger than that of the (S)-
enantiomer [7], or that the (S)-enantiomer is odorless [8]. The (R)-benzoxepin 1 was
also constructed for similar reasons. Conformational analysis was then carried out by
rotating the dihedral angles of bonds 1, 2, 3, 4, 5, and 6 in benzoxepin 1 and bonds 1', 2',
3', and 4' in aldehyde 2 (Fig. 1). The dihedral angle of bond 4' in aldehyde 2 was rotated
for further comparison with other compounds in this work. The root-mean-square
(r.m.s.) threshold conditions of conformational analysis for aldehyde 2 and benzoxepin
1 were 0.3 and 0.1 Š, respectively.
As a result, 14 conformers of benzoxepin 1 and 22 conformers of aldehyde 2 were
selected (see Exper. Part). The features of the conformers are depicted in Fig. 2 by the
total energy [kcal/mol] vs. molecular size [Š] as estimated by the longest side length of
a circumscribed box (LLCB). The range of LLCB of the selected conformers of
aldehyde 2 and benzoxepin 1 were almost the same. As seen in these graphs, there were
two different groups of conformers, one being rather short (12 Š) and the other being
longer (13 Š). All 14 conformers of benzoxepin 1 were superimposed with all 22
conformers of aldehyde 2 as depicted in Fig. 3, showing the chosen conformers viewed

Helvetica Chimica Acta ± Vol. 84 (2001)
2053
Fig. 2. Size estimated by LLCB [Š] vs. total energy [kcal/mol]. Left: conformers of benzoxepin 1; right:
conformers of aldehyde 2.
Fig. 3. Conformers of benzoxepin 1 superimposed on the conformers of aldehyde 2. For convenience, the H-
atoms are omitted. Black, O-atom of 1; dark gray, O-atom of aldehyde 2; light gray, skeletons, i.e. C-atoms and
their connected bonds of 1 and 2.
from two different directions. The quaternary C-atom of the t-Bu group and six atoms
in the benzene ring of each compound 1 and 2 were selected for matching. In the case of
several possible superimpositions of conformers, the two O-functionalities were
arranged such as to be in the same direction and the two O-atoms such as to be as close
as possible to each other. It is apparent from Fig. 3 that no conformers of the
benzoxepin 1 match any conformers of aldehyde 2 in terms of the spatial position of the
O-functionalities. In other words, no conformer of benzoxepin 1 active for the type of
odor typical of 2 would be expected to be present. The study of Skouroumounis and

2054
Helvetica Chimica Acta ± Vol. 84 (2001)
Winter was based on the molecular-mechanics (MM2) calculation with the Macro
Model [6]. We employed the molecular-mechanics parameters of Sybyl for the
conformational search and the semi-empirical molecular-orbital method (PM3) for the
structure optimization. In addition, the atoms used for superimposition may be
different between the two studies. In our study, the benzene ring and the center of the
t-Bu group were matched, because they are common moieties occupying a large space
in benzoxepin 1 as well as in aldehyde 2. Thus, the differences of our results from those
of Skouroumounis and Winter are due to both the different calculation and the
different superimposition methods.
In the next stage, in addition to aldehyde 2, conformational analyses of three other
compounds having the lily-of-the-valley odor were carried out, namely of aldehyde 3
(Bourgeonalꢂ), alcohol 4 (Mugetanolꢂ), and aldehyde 5 as shown in Fig. 4. For alcohol
4, cis-Mugetanol ꢂ was selected because it has been reported that the lily-of-the-valley
aspect and intensity of ( )-cis-Mugetanol ꢂ are far superior to those of the other
isomers [9]. To gain a better superimposition with other molecules, both the (R) and
(S) configurations were constructed for alcohol 4 as initial structures. Aldehyde 5 is a
ring-opened analogue of the benzoxepin 1 [6].
Fig. 4. Compounds having the lily-of-the-valley odor
The number of conformers selected by conformational analyses were 28 for
aldehyde 3, 9 for the (R)- or (S)-configuration of alcohol 4, and 23 for aldehyde 5 (see
Exper. Part). The range of LLCB for the conformers of alcohol 4 was from 10.4 to
12.3 Š. The 12-Š conformers, the shorter LLCB conformers, were selected in the case
of aldehydes 2, 3, and 5 because we presumed the existence of a common conformer of
compounds 2 ± 5 as the odor-active conformer, and there were no conformers at ca.
13 Š for alcohol 4. The shorter conformers ( ꢀ 12 Š) were superimposed after PM3
optimization of the structures. Superimposition was again based on the O-function-
alities, including the bulky moiety of the t-Bu or i-Pr group and the six-membered
benzene or cyclohexane ring. Among the conformers selected and superimposed, there
were some conformers that showed a good agreement in the position or the direction of
the O-atoms (Fig. 5). The conformers in Fig. 5 were assumed to be close to the
conformer active for the lily-of-the-valley odor, because the compounds having this
odor are expected to adopt a similar conformation and to fit with the specific receptor
protein in the same manner. Since it is difficult to specify one active conformer for each
compound from the limited information here, all the conformers in Fig. 5 were kept as
active conformers. Fig. 5 represents an integrated model giving information regarding
the size and structure of molecules having the lily-of-the-valley odor. Models for the
lily-of-the-valley-type odor have been proposed by Pelzer, based on an AM1
calculation [9]. There are two types of fragment structures representing compounds

Helvetica Chimica Acta ± Vol. 84 (2001)
2055
Fig. 5. Superimposition of selected conformers of aldehyde 2, aldehyde 3, alcohol 4, and aldehyde 5, viewed from
two different directions. For convenience, H-atoms are omitted, except the H- atom of the OH group of the
conformers of alcohol 4.
with hydroxy and carbonyl functional groups, respectively. Fig. 5 was taken as a
common model for odorants that involve either fragment type.
The other molecular features of compounds 2 ± 5, such as the average electrostatic
potential (ESP) charges of the O-atoms in selected conformers and the logarithm of
the octanol/water partition coefficient (log P), were calculated, and the results are
listed in Table I. ESP Charge is an electrostatic parameter and log P is a hydrophobic
parameter. From odor-structure-relationship studies [1][2], electrostatic and hydro-
phobic factors seem to be important, in addition to steric or shape parameters. The size
of odoriferous molecules would be an important parameter when they are recognized
at specific olfactory receptors. In 1998, Zhao et al. have succeeded in driving the
expression of a particular receptor gene in the rat olfactory epithelium using a
recombinant adenovirus. The size specificity of odoriferous molecules was seen in the
infected animal by electrophysiological recording [10]. Molecular size might be a factor
to determine whether odorant-receptor interactions can be established.
Based on the superimposed model as shown in Fig. 5, we anticipated that
compounds of two categories, e.g. compounds 6 ± 9 of Type A and compounds 10 ± 14
of Type B (Fig. 6), are candidates to possess a lily-of-the-valley-type odor. The spatial
positions of O-functionalities of these compounds were focused on at first, and the t-Bu
Table 1. Other Molecular Features of Aldehyde 2, Aldehyde 3, Alcohol 4, and Aldehyde 5
Average of ESPa) charge
Calculated log Pb)
No of conformers
of the O-atom
selected for Fig. 5
aldehyde 2
aldehyde 3
alcohol 4
0.26
0.28
0.33
0.28
3.56
3.30
2.79
3.56
6
4
2
4
aldehyde 5
a) ESP, electrostatic potential. b) P, octanol/H2O partition coefficient.

2056
Helvetica Chimica Acta ± Vol. 84 (2001)
Fig. 6. Candidate compounds 6 ± 14 having potentially the lily-of-the-valley odor. The two rings are connected
via a dashed line in Type B compounds.
and benzene moieties were kept as is. Carbonyl or OH groups were connected to rings
to restrict the position of the O-atoms. Both (R)- and (S)-configurations of the
compounds in Fig. 6 carrying a stereogenic center were considered since the spatial
position and direction of O-functionalities are important. Conformational analyses and
structure optimizations of all the candidate compounds 6 ± 14 were then carried out.
The selected conformers of each compound 6 ± 14 were superimposed on the model
(see Fig. 5) to compare them with the active conformers of compounds 2 ± 5 (see Exper.
Part).
After extensive investigations, benzocycloheptenone 8 and the (S)-enantiomer of
cyclopentanone 13 were selected, because, among the other candidates, the conformers
of compounds 8 and 13, both fit well with the conformers in Fig. 5. Superimposed
conformers of benzocycloheptenone 8 and cyclopentanone 13 are depicted in Figs. 7
and 8, respectively. The position and direction of the O-atoms in compounds 8 and 13
fit well with those of the conformers in Fig. 5, though not completely. The other
molecular features of the best-fit conformers of benzocycloheptenone 8 and cyclo-
pentanone 13 were also calculated and are shown in Table 2. Molecular size, three-
dimensional structure (especially accordance with the position of the O-atoms), ESP
charges of O-atoms, and the calculated log P of benzocycloheptenone 8 and
cyclopentanone 13 were not significantly different from those of compounds 2 ± 5
shown in the model of Fig. 5 and in Table I.
2.2. Synthesis of the Designed Molecules Benzocycloheptenone 8, Cyclopentanone
13, and Benzocycloheptenone 23. Benzocycloheptenone 8 was prepared according to
Table 2. Other Molecular Features of the Designed Compounds Benzocycloheptenone 8 and Cyclopentanone 13
ESPa) charge of the O-atom
Calculated log Pb)
Benzocycloheptenone 8
Cyclopentanone 13
0.29
0.29
3.82
3.43
a) ESP, electrostatic potential. b) P, octanol/H2O partition coefficient.

Helvetica Chimica Acta ± Vol. 84 (2001)
2057
Fig. 7. Superimposition of conformers of benzocycloheptenone 8 on the conformers in Fig. 5, viewed from two
different directions. For convenience, the H-atoms are omitted.
Fig. 8. Superimposition of conformers of cyclopentanone 13 on the conformers in Fig. 5, viewed from two
different directions. For convenience, the H-atoms are omitted.
the procedure of Mataka and co-workers [11] as shown in Scheme 1. Bromination of
diol 16, prepared by lithium aluminium hydride reduction of 4-(tert-butyl)phthalic
anhydride (15) was carried out with CBr4 and triphenylphosphine to give dibromide 17
in 93% yield. Alkylation of diethyl 3-oxopentanedioate with dibromide 17 provided
oxobenzocycloheptanedicarboxylate 18 in 70% yield under phase-transfer conditions,
and decarboxylation of 18 in aqueous alkaline media furnished the desired
benzocycloheptenone 8 in 84% yield.

2058
Helvetica Chimica Acta ± Vol. 84 (2001)
Scheme 1. Synthesis of Benzocycloheptenones 8 and 23
i ) LiAlH4, Et2O; 92% (16), 90% (25). ii) CBr4, PPh3, CCl4; 93% (17), 92% (26). iii) Bu4NI, 5% aq. NaHCO3
soln., (EtOOCCH2) 2CO, CH2Cl2; 70% (18, 55% (27). iv) KOH, H2O / EtOH; 84% (8), 88% (23).
Cyclopentanone 13 had been used to investigate the chromatographic resolution
and ability of chiral recognition on methylbenzoylcellulose beads [12]. Cyclopentanone
13 was also the intermediate of a new fungicide [13]. However, its odor was not
described. The first step of its synthesis, i.e. 1,4-conjugated addition of 4-(tert-
butyl)phenylmagnesium bromide to cyclopentanone 20 in the presence of copper
reagent did not give reproducible results. Thus, an alternative route was employed
involving the Grignard reaction of bromide 19 with cyclopentenone 20, which provided
the tertiary alcohol 21 in 85% yield. The latter was treated with pyridinium dichromate
(PDC) in CH2Cl2 to give cyclopentenone 22 in 82% yield. Catalytic hydrogenation of
22 provided the desired cyclopentanone 13 in 50% yield. The volatility of cyclo-
pentanone 13 was responsible for the moderate yield after hydrogenation. Attempted
reductive Heck reaction [14] of cyclopentenone 20 with of 4-(tert-butyl)phenyl
bromide 19 leading to cyclopentanone 13 failed.
To increase the volatility of tert-butyl-substituted benzocycloheptenone 8, the
methyl-substituted benzocycloheptenone 23 was prepared via 24 ± 27 according to the
same synthetic procedure as described for 8 (see Scheme 1).
2.3. Olfactory Evaluation of the Designed Molecules Benzocycloheptenone 8,
Cyclopentanone 13, and Benzocycloheptenone 23. The odors of the three compounds
were evaluated by eight perfumers as shown in Fig. 9. According to the comments of
the perfumers, benzocycloheptenone 8 recalls the note of Cyclamenaldehyde ꢂ and Lily
Aldehyde ꢂ, and cyclopentanone 13 recalls the note of lilac. The synthetic compounds 8
and 13 possess the targeted lily-of-the-valley-type odor, though their intensities are
weak and not adequate for perfume materials. The additional synthetic benzocyclo-

Helvetica Chimica Acta ± Vol. 84 (2001)
Scheme 2. Synthesis of Cyclopentanone 13
2059
i) Mg, THF; 85%. ii) PDC, CH2Cl2; 82%. iii) Pd/C, H2, EtOH; 50%.
Fig. 9. Intensity of eight odor notes of the synthetic benzocycloheptenone 8, cyclopentanone 13, and
benzocycloheptene 23, evaluated by eight perfumers. Intensity 1: the degree of the detected feature is weak;
intensity 2: the degree of the feature is rather weak; intensity 3: the degree of the feature is medium.
heptenone 23 also recalls the note of Cyclamenaldehyde ꢂ and Lily Aldehyde ꢂ, and also
has a distinctive marine note. As a novel perfume material, the value of benzocyclo-
heptenone 23 is rather low, because there are many other compounds having a marine
note.
Concerning the odor type or odor note, the designed molecules benzocyclohepte-
none 8 and cyclopentanone 13 were confirmed to have floral, lily-of-the-valley-type
odor by olfactory evaluation. It is worth to note that the benzocycloheptenone 8 and
cyclopentanone 13 were proposed only by computational analyses. Both compounds
have the lily-of-the-valley odor and did not have other strong odor notes. Therefore,
our methodology to design new odoriferous compounds based on conformational

2060
Helvetica Chimica Acta ± Vol. 84 (2001)
analysis proved to be satisfactory. The model obtained in this work also would predict
the conformation of odorants having this odor at the receptor protein.
Concerning the intensity of odor, the model is still insufficient. Analysis of a variety
of compounds having this odor may lead to a better model to design compounds having
the lily-of-the-valley odor. A series of pyranyl and furanyl ether type compounds, as
well as compounds having a hydroxy or carbonyl group, have been reported by Pelosi et
al. [15][16]. Furthermore, the discovery of additional molecular parameters related to
odor is expected to lead to improved models.
3. Conclusion ± By intensive conformational analyses of 1,3,4,5-tetrahydro-2-
benzoxepin derivative 1 and 3-[4-(tert-butyl)phenyl]-2-methylpropanal (2), the dis-
agreement of odor of 1 and 2 was explained based on their stable conformers.
Additional analyses of compounds 3 ± 5 having the lily-of-the-valley odor established
that some conformers of these compounds had a common molecular size and position
of the O-functionalities. The corresponding conformers were assumed to be odor-active
conformers. The synthesis of designed compounds based on our integrated model (see
Fig. 5), i.e. of benzocycloheptenone 8 and cyclopentanone 13, revealed that they
possess the targeted odor of the lily of the valley. This verifies the choice of the
conformers active for the lily-of-the-valley odor, as selected by conformational analysis.
Our methodology to design new odoriferous compounds was successful in terms of
odor type. However, in terms of odor intensity, the model has to be improved.
Investigations on the relation between odor type and odor intensity and on the
additional molecular features related to odor are underway.
We thank Soda Aromatic Co., Ltd. for the odor evaluation of the synthetic benzocycloheptenone 8,
cyclopentanone 13, and benzocycloheptenone 23, and for the mass-spectral measurements. A scholarship for
F. Y. by Sasaki Environmental Technology Foundation is also gratefully acknowledged.
Experimental Part
1. Computational Analysis. A flow-chart for the general calculation is shown in Fig. 10. 1) Starting structures
were constructed and minimized by molecular-mechanics calculations with Tripos force field parameters in the
SYBYL QSAR module (version 6.2 and 6.3) on a Silicon Graphics IRIS 4D/420GTX. 2) Conformational
analyses were performed by random search [17]. Standard criteria for the search were as follows: maximum
cycle was from 3000 to 30000 times, depending on the structural flexibility; r.m.s. threshold was between 0.1 and
0.5 Š, according to the molecular structures; energy convergence was 0.005 kcal/mol. After the conforma-
tional analysis, the conformer with the lowest energy and those with the energy up to 5 kcal/mol from the lowest
energy were selected. 3) The size and compactness of molecules were checked by LLCB. The LLCB was
calculated by our self-written program on the same workstation. The LLCB was first introduced as ꢀthe longest
side length of a hexahedron that circumscribes the van der Waals surface of a moleculeꢁ in our previous work
[18]. The circumscribed hexahedron is made by the best plane, the intermediate plane, and the worst plane. The
sum of squares of distance between the best plane and each atom of the molecule is the smallest. In other words,
we can put atoms in a molecule exactly on or nearest to the best plane. The LLCB has been established to be
useful in previous works [19][20][21] as a parameter for estimating molecular size. This parameter was used to
select or eliminate conformers for further optimization. 4) Structure optimizations of the conformers based on
bond lengths, bond angles, and dihedral angles, by means of the PM3 method [22], were carried out in
succession. 5) Graphic images of conformers were superimposed using the least-squares-fit method for
comparison of the 3-D structures in the SYBYL. 6) Other molecular features were considered: log P, as a
parameter of the hydrophobicity of the whole molecule, was calculated by MlogP [23] on an Apple-Macintosh
LC520 computer. The MlogP does not depend on the conformation of a molecule. The electrostatic-potential
(ESP) charges of O-atoms of selected conformers were also calculated with KEY WORD, ꢀESPꢁ, by the

Helvetica Chimica Acta ± Vol. 84 (2001)
2061
PM3 method, because an O-atom is considered to be a requisite for the lily-of-the-valley odor. The coordinates
of conformers selected in this work are available. Please contact F. Yoshii (fuyoshii@mb.infoweb.ne.jp) or
H. Hagiwara (hagiwara@gs.niigata-u.ac.jp).
Fig. 10. Calculation method
2. General: Medium-pressure liquid chromatography (MPLC): GL-Science PU-612 instrument, silica-gel
packed column. CC ˆ Column chromatography. M.p.s: Yanaco MP hot-stage apparatus; uncorrected. IR
Spectra: Shimadzu FT/IR-4200 spectrophotometer, CCl4 solns.; in cm 1. 1H-NMR Spectra: Varian Gemini 200 H
(200 MHz) instrument; CDCl3 solns.; d in ppm rel. to SiMe4 as internal standard (ˆ0 ppm), J in Hz. 13C-NMR
Spectra: Varian Gemini 200 H (50 MHz) instrument. MS: Hitachi M-4100 spectrometer; recorded at Soda
Aromatics Co., Ltd.; m/z (rel. %). Microanalyses were carried out in the microanalytical laboratory of the
Instrumental Analysis Center for Chemistry, Tohoku University.
3. Syntheses. 4-(tert-Butyl)benzene-1,2-dimethanol (16). To a soln. of 4-(tert-butyl)phthalic anhydride (15;
4634 mg, 22.7 mmol) in Et2O (80 ml), LiAlH4 (2161 mg, 56.9 mmol) was added at 788. After 30 min stirring at
788, the mixture was stirred for 3 h at r.t.. The reaction was quenched by addition of Et2O and sat. aq. NH4Cl
soln. and the org. layer dried (Na2SO4) and evaporated CC(AcOEt/hexane 1:2) provided 16 (4050 mg, 92%).
IR (CCl4): 3535, 3080, 2966, 2907, 2872, 1614, 1477, 1020. 1H-NMR (200 MHz, CDCl3): 1.32 (s, 9 H); 3.04 (s,
2 H); 4.71 (s, 2 H); 4.74 (s, 2 H); 7.30 ± 7.37 (m, 3 H). 13C-NMR (50 MHz, CDCl3): 31.2; 34.4; 63.3; 64.1; 76.4;
‡
77.0; 77.6; 125.0; 126.6; 129.4; 136.3; 138.8; 151.3. EI-MS: 194 (4, M ), 179 (28), 176 (62), 161 (100), 133 (28),
‡
‡
117 (7), 105 (22), 91 (16), 77 (7), 57 (8) and 41 (8) HR-MS: 194.1348 (C12H18O2 , M ; calc. 194.1307).
Diethyl 7-(tert-Butyl)-2,3,4,5-tetrahydro-3-oxo-1H-benzocycloheptene-2,4-dicarboxylate (18). To a stirred
suspension of Bu4NI (1190 mg, 3.22 mmol) in 5% aq. NaHCO3 soln. (43 ml) and CH2Cl2 (10 ml), a soln. of
dibromide 17 (1714 mg, 5.36 mmol) and diethyl 3-oxopentanedioate (1.27 ml, 6.97 mmol) in CH2Cl2 (10 ml) was
slowly added, and the resulting suspension was stirred vigorously at r.t. for 21 h. After addition of sat. aq. NH4Cl
soln., the org. layer was extracted with AcOEt (2Â), washed with H2O and brine, dried (Na2SO4), and
evaporated. CC(AcOEt/hexane 1:7) provided 18 (1354 mg, 70%). IR (CHCl3): 2966, 1743, 1711, 1566, 1466,
1441, 1387, 1356, 1266, 1146, 1022. 1H-NMR (200 MHz, CDCl3): 1.23 (t, J ˆ 7.0, 3 H); 1.30 (t, J ˆ 7.0, 3 H); 1.32 (s,

2062
Helvetica Chimica Acta ± Vol. 84 (2001)
9 H); 3.03 ± 3.32 (m, 4 H); 3.48 ± 3.74 (m, 2 H); 4.05 ± 4.33 (m, 4 H); 7.08 ± 7.25 (m, 3 H). 13C-NMR (50 MHz,
CDCl3): 14.0; 14.1; 31.3; 32.9; 33.8; 34.4; 58.2; 59.5; 61.3; 61.4; 124.5; 127.0; 129.8; 133.8; 136.3; 150.6; 169.4;
‡
202.3. EI-MS: 360 (37, M ), 315 (54), 314 (100), 286 (62), 241 (35), 240 (69), 213 (71), 173 (33), 129 (29), 58
‡
‡
(74). HR-MS: 360.1947 (C21H28O5 , M ; calc. 360.1937).
7-(tert-Butyl)-1,2,4,5-tetrahydro-3H-benzocyclohepten-3-one (8). A soln. of 18 (1354 mg, 3.76 mmol) in 8%
aq. KOH soln. (20 ml) and EtOH (30 ml) was heated to reflux for 1.5 h. After addition of 1n HCl, the mixture
was extracted with CHCl3 (2Â) and the combined org. layer washed with H2O and brine, dried (Na2SO4), and
evaporated. CC(MeOH/CHCl3 1 :2) afforded 8 (684 mg, 84%), which was recrystallized from hexane/CHCl3.
M.p. 54.8 ± 55.18. IR (CHCl3): 2964, 2909, 2864, 1709, 1442. 1H-NMR (200 MHz, CDCl3): 1.32 (s, 9 H); 2.58 ± 2.66
(m, 4 H); 2.86 ± 2.95 (m, 4 H); 7.13 ± 7.28 (m, 3 H). 13C-NMR (50 MHz, CDCl3): 30.1; 31.0; 31.4; 34.4; 44.6; 44.6;
123.8; 126.1; 128.9; 137.4; 140.0; 150.0; 211.6. Anal. calc. for C15H20O (216.32): C 83.29, H 9.32; found: C 83.19, H
9.46.
1-[4-(tert-Butyl)phenyl]cyclopent-2-en-1-ol (21). To a stirred slurry of Mg (294 mg, 12 mmol) in THF
(10 ml) under N2, 1-bromo-4-(tert-butyl)benzene (1.75 ml, 10 mmol) and dibromoethane (4 ml) were added.
Stirring was continued for 3 h at r.t. Then, a soln. of cyclopent-2-en-1-one (20) (0.57 ml, 7.0 mmol) in THF (3 ml)
was added at 08, and the resulting soln. was stirred at r.t. for 1 h. The reaction was quenched by addition of sat.
aq. NH4Cl soln. and the org. layer extracted with AcOEt (2Â), washed with H2O and brine, dried (Na2SO4), and
evaporated. MPLC (AcOEt/hexane 1:2) purification of the residue provided 21 (1291 mg, 85%). IR (CCl4):
3599, 2945, 2854, 1603, 1493, 1359, 1284, 1111. 1H-NMR (200 MHz, CDCl3): 1.32 (s, 9 H); 2.25 (m, 2 H); 2.33 (m,
2 H); 5.92 (td, J ˆ 5.6, 2.4, 1 H); 6.15 (td, J ˆ 5.6, 2.2, 1 H); 7.42 (m, 4 H). 13C-NMR (50 MHz, CDCl3): 31.3; 31.5;
‡
34.4; 41.7; 86.7; 124.5; 125.0; 134.5; 136.6; 144.0; 149.6. EI-MS: 215 (33, [M 1] ), 199 (43), 198 (58), 183 (100),
‡
‡
161 (28), 159 (35), 155 (15), 115 (11), 57 (14). HR-MS: 215.1443 (C15H20O , [M 1] ); calc. 215.1436).
3-[4-(tert-Butyl)phenyl]cyclopent-2-en-1-one (22). A slurry of PDC (568 mg, 1.51 mmol), NaOAc (126 mg,
1.53 mmol), and 4 Š molecular sieves (100 mg) in CH2Cl2 (6 ml) was stirred for 30 min under N2. Then a soln. of
the 21 (113 mg, 0.53 mmol) in CH2Cl2 (2 ml) was added, and stirring was continued for 30 min. The mixture was
passed through a short SiO2 column, the filtrate evaporated, and the residue purified by MPLC (AcOEt/hexane
1:2): 22 (90 mg, 82%), which was recrystallized from hexane/CHCl3. M.p. 108.3 ± 110.18. IR (CCl4): 2966, 2909,
2870, 1705, 1600, 1464, 1412, 1296, 1184. 1H-NMR (200 MHz, CDCl3): 1.35 (s, 9 H); 2.65 (td, J ˆ 2.8, 1.9, 2 H);
3.05 (m, 2 H); 6.56 (t, J ˆ 1.7, 1 H); 7.52 (m, 2 H); 7.61 (m, 2 H). Anal. calc. for C15H18O (214.31): C 84.07, H 8.47;
found: C 83.80, H 8.77.
3-[4-(tert-Butyl)phenyl]cyclopentan-1-one (13). A slurry 22 (728 mg, 3.4 mmol) and Pd/C (49 mg) in EtOH
(13 ml) was stirred at r.t. 20.5 h under H2. Filtration and evaporation gave a residue, which was purified by
MPLC (AcOEt/hexane 1:2): 13 (365 mg, 50%), which was recrystallized from hexane/CHCl3. M.p. 68.48. IR
(CCl4): 3057, 2966, 2870, 1748, 1462, 1408, 1329. 1H-NMR (200 MHz, CDCl3): 1.32 (s, 9 H); 1.95 (m, 2 H); 2.45
(m, 4 H); 3.40 (m, 1 H); 7.20 (m, 2 H); 7.38 (m, 2 H). Anal. calc. for C15H20O (216.32): C 83.29, H 9.32; found: C
82.83, H 9.57.
Diethyl 2,3,4,5-Tetrahydro-7-methyl-3-oxo-1H-benzocycloheptene-2,4-dicarboxylate (27). As described for
18, with Bu4NI (663 mg, 1.79 mmol), 5% aq. NaHCO3 soln. (15 ml), CH2Cl2 (10 ml), dibromide 26 (471 mg,
1.69 mmol), diethyl 3-oxopentanedioate (0.43 ml, 2.34 mmol), and CH2Cl2 (3 ml), at 408 for 3.5 h. CC (AcOEt/
hexane 1:3) provided 27 (294 mg, 55%). IR (CHCl3): 3028, 2980, 2966, 1743, 1711, 1566, 1466, 1441, 1201.
1H-NMR (200 MHz, CDCl3): 1.23 ± 1.32 (m, 15 H); 3.03 ± 3.32 (m, 4 H); 3.48 ± 3.74 (m, 2 H); 4.05 ± 4.33 (m,
‡
4 H); 7.08 ± 7.25 (m, 3 H). EI-MS: 318 (31, M ), 272 (100), 244 (51), 198 (76), 171 (68), 143 (88), 131 (66) and
115 (30), 91 (16).
1,2,4,5-Tetrahydro-7-methyl-3H-benzocyclohepten-3-one (23). As described for 8, with 27 (294 mg,
0.92 mmol), 6.5% aq. KOH soln. (7 ml), and EtOH (7 ml), under reflux for 2 h: 23 (140 mg, 88%), which
was recrystallized from hexane/CHCl3. M.p. 61.8 ± 63.68. IR (CHCl3): 3024, 2964, 2924, 2909, 1711, 1442.
1H-NMR (200 MHz, CDCl3): 1.32 (s, 9 H) ; 2.58 ± 2.66 (m, 4 H); 2.86 ± 2.95 (m, 4 H) and 7.13 ± 7.28 (m, 3 H). EI-
‡
MS: 174 (100, M ), 159 (20), 146 (21), 132 (59), 131 (64), 117 (52), 91 (20), 77 (8). Anal. calc. for C15H20O
(216.32): C 82.72, H 8.11; found: C 82.50, H 8.32.
4. Olfactory Evaluation of the Synthetic Molecules. Benzocycloheptenones 8 and 23 and cyclopentanone 13
were dissolved in dipropylene glycol to give a 10% soln. Eight perfumers evaluated their odor notes using paper
strips. The odor note is judged by 8 features, which are fruity, floral, woody, musty, animal, musky, amber, and
fatty. Each perfumer ranked the intensity of the 8 features in a scale from 0 to 5 :0, not detecting the specified
feature at all; 1, the degree of the detected feature is weak; 2, the degree of the feature is rather weak; 3, the
degree of the feature is medium; 4, the degree of the feature is rather strong; 5, the degree of the feature is
strong. The average intensity (A. I.) of the eight features is calculated according to the following equation.

Helvetica Chimica Acta ± Vol. 84 (2001)
2063
A. I. ˆ {SNi (number of perfumer who ranked the sample as i) ´ i (ranking point, i ˆ 0,1,2,3,4,5)} / 8 (total number
of perfumers). We also asked the perfumers to give the names of perfume materials recalled by each synthetic
compound.
REFERENCES
[1] K. J. Rossiter, Chem. Rev. 1996, 96, 3201.
[2] M. Chastrette, SAR QSAR Environ. Res. 1997, 6, 215.
[3] L. Buck, R. Axel, Cell 1991, 6, 175.
[4] K. Raming, J. Krieger, J. Strotmann, I. Boekhoff, S. Kubick, C. Baumstark, H. Breer, Nature (London)
1993, 361, 353.
[5] M. S. Singer, G. M. Shepherd, Neuro Report 1994, 5, 1297.
[6] G. Skouroumounis and B. Winter, Helv. Chim. Acta 1996, 79, 1095.
[7] D. Enders, H. Dyker, Liebigs Ann. Chem. 1990, 11, 1107.
[8] A. Mosandl, Z. Lebensm.-Unters. Forsch. A 1997, 205, 76.
[9] R. Pelzer, U. Harder, A. Krempel, H. Sommer, H. Surburg, P. Hoever, in ꢀRecent Developments in Flavor
and Fragrance Chemistryꢁ, Proceedings of the 3rd International Haarmann & Reimer Symposium, Eds. R.
Hopp and K. Mori, VCH Verlagsgesellschaft, Weinheim, 1993, p. 48.
[10] H. Zhao, L. Ivic, J. M. Otaki, M. Hashimoto, K. Mikoshiba, S. Firestein, Science (Washington, D.C.) 1998,
279, 237.
[11] S. Mataka, Y. Mitoma, T. Thiemann, T. Sawada, M. Taniguchi, M. Kobuchi, M. Tashiro, Tetrahedron 1997,
53, 3015.
[12] E. Francotte, R. M. Wolf, J. Chromatogr. 1992, 595, 63.
[13] A. Huxley-Tencer, E. Francotte, M. Bladocha-Moreau, Pestic. Sci. 1992 34, 65.
[14] H. Hagiwara, Y. Eda, K. Morohashi, T. Suzuki, M. Ando, N. Ito, Tetrahedron Lett., 1998, 39, 4055.
[15] C. Anselmi, M. Centini, M. Mariani, A. Sega, P. Pelosi, J. Agric. Food Chem. 1992, 40, 853.
[16] C. Anselmi, M. Centini, M. Mariani, A. Sega, P. Pelosi, J. Agric. Food Chem. 1993, 41, 781.
[17] M. Saunders, J. Am. Chem. Soc. 1987, 109, 3150.
[18] F. Yoshii, Q. Liu, S. Hirono, I. Moriguchi, Chem. Senses 1991, 16, 319.
[19] F. Yoshii, S. Hirono, I. Moriguchi, Quant. Struct.-Act. Relat. 1994, 13, 144.
[20] F. Yoshii, S. Hirono, Chem. Sens. 1996, 21, 201.
[21] F. Yoshii, S. Hirono, I. Moriguchi, Helv. Chim. Acta 1993, 76, 2279.
[22] J. J. P. Stewart, ꢀMOPAC Ver. 6,ꢁ QCPE #455.
[23] I. Moriguchi, S. Hirono, I. Nakagome, H. Hirano, Chem. Pharm. Bull. 1994, 42, 976.
Received March 1, 2001