Synthesis and Olfactory Evaluation of Bulky Moiety-Modified Analogues to the Sandalwood Odorant Polysantol®

Five new bulky moiety-modified analogues of the sandalwood odorant Polysantol® have been synthesized by aldol condensation of appropriate aldehydes with butanone, deconjugative α-methylation of the resulting α,β-unsaturated ketones, and reduction of the corresponding β,γ-unsaturated ketones. The final compounds were evaluated organoleptically and one of them seemed to be of special interest for its natural sandalwood scent.


Introduction
(-)-(Z)-β-Santalol (1), the main constituent of natural sandalwood oil, is an odour compound with typical sandalwood fragrance and is described as warm-woody, creamy and sweet with an animalic tonality [1,2]. It consists of a bulky bicyclic moiety separated from the hydroxyl group by an unsaturated 5 C-atoms spacer [3]. The best synthetic substitutes for this noble perfumery raw material are a series of trimethylcyclopentenyl alkenols, such as 2−5 [4,5], derived from campholenic aldehyde. The structural similarities between β-santalol and these substitutes, regarding the bulky lipophile, the spacer and the osmophoric polar hydroxyl group, seems to be clear (Figure 1), Polysantol ® (2) being OPEN ACCESS the most expensive and appreciated by perfumers [3]. The structure-odour properties of this compound and a series of derivatives have been studied [6][7][8]. The structure of olfactory receptors and the corresponding mechanism of interaction between receptor proteins and odour molecules, rewarded by the 2004 Nobel Prize [9,10], are still little known. Therefore, the determination of essential structural elements responsible for the sandalwood-type sensation can be only performed by molecular similarity studies within a series of sandalwood odour compounds and structurally similar, but odourless, molecules. As is well known [3,11], three subunits are important for the sandalwood odour impression (Figure 1), which correspond to the hydroxyl group (A), a lipophilic substituent (B) in the neighbourhood of this hydroxyl group, and a bulky rigid hydrophobic moiety (C). This set of structural features constitutes the sandalwood olfactophore. In this way, some fragrance chemists assumed that the vicinity of the osmophore must be crucial for the odour and this flexible spacer became the main object of structure-odour sandalwood studies [12]. On the other hand, the analysis of structure-odour relationship (SOR) data allowed to postulate that the geometry of the immediate proximity of the osmophoric hydroxyl group tolerates less variations than the orientation of the more distant lipophilic bulky group [3]. For that reason the bulky moiety of the trimethylcyclopentenyl group in campholenal derivatives 2−5 has been replaced by structures of similar steric bulk (6 [13], 7 [14], 8 [15], 9 [16], 10 [17], 11 [18]).
As a continuation of our previous studies on the synthesis of odorants [19][20][21], we have developed a collection of several substitutes of sandalwood scent [22][23][24]. As other authors have done [7,8,25], we have studied the influence of the global shape of the hydrophobic moiety C, and for the refinement of the olfactophore model on compounds structurally similar to Polysantol ® , five new compounds 33−37 ( Figure 3) have been synthesized for this work and their odour evaluated. These molecules have been obtained from the aldehydes 1215 and 18, respectively, through a straightforward process involving the aldol condensation of each starting aldehyde with butanone, the deconjugative α-methylation of the respective enones and the reduction of the corresponding -unsaturated ketone to yield every alcohol analogue to the odorant Polysantol ® (see Scheme 2 below).
2.1.1. Conversion of nopol (16) into dihydronopal (18) According to the findings of Heitmann and Mätzel [27], the use of Adams' catalyst in methanol with a low hydrogen pressure allowed us to obtain cis-dihydronopol (17) [28,29] in good yield (90%) and with high diastereoselectivity. Therefore, the substituent in the 2 position is cis with respect to the gem-dimethyl bridge in 17 (Scheme 1). Scheme 1. Selective hydrogenation of nopol and oxidation of cis-dihydronopol under mild conditions. a d because of a single geminal coupling. The different resonance signals of the two methyl groups on C-6', due to the magnetic anisotropy of the cyclobutane ring, is also characteristic of this skeleton. Hence, Me-8' (equatorial) always appears ca 0.4 ppm deshielded respect to Me-9' (axial) in 2-αpinene derivatives (such as 16). Nevertheless, in 2-αH-pinane derivatives (such as 17) the less rigid geometry compared to the saturated system produces an appreciable change in the resonance position of the equatorial and axial methyl groups on C-6', the Δδ between them being now ca. 0.1 ppm. The highly overlapped region of δ 1.80−2.00, corresponding to the 1', 3's, 4'a, 4's and 5' protons was too poorly separated for determination of coupling constants. However, the chemical shifts of such protons were obtained from the 2D NMR shift correlations (HSQC, HMBC, COSY and NOESY). In addition, homodecoupling experiments were also performed to obtain some coupling constants. In general, the relationship δHa < δHe is valid, except for H-4, where 4a (an equatorial proton) resonates at higher field than 4s (an axial proton), which is in accordance with the finding for protons attached to a cyclohexane ring [31]. The conversion of dihydronopol (17)  The starting aldehydes 1215 and 18 were reacted with butanone by aldol condensation. Thus, isovaleraldehyde (12) yielded the -unsaturated ketone 19 using potassium hydroxide as catalyst [34] (Scheme 2). The intermediate β-hydroxyketone was directly dehydrated by azeotropic distillation in dry toluene and p-toluenesulfonic acid [23,35]. The crude 19 obtained was purified by flash chromatography to afford pure 19 in 88% yields.
For the synthesis of 20 [36], heptanal (13) was reacted in a similar manner; aldol reaction with butanone followed by p-toluenesulfonic acid-assisted dehydration. The α,β-unsaturated ketone 20 was obtained in 71% yield [38]. The spectroscopic properties of 20 and 19 are alike with respect to the synthon C1-C5, and regarding 21, the aldol self-condensation by-product derived from 13, the NMR data agree with those already reported [39].
When citronellal (14) was used as starting aldehyde, the aldol condensation with butanone to obtain the α,β-unsaturated ketone 22 was performed using a basic thermal dehydration instead the acid dehydration [40]. This was necessary because although all the aldol condensation attempts via acid dehydration led to the desired 22, it immediately underwent an intramolecular Michael-type reaction that led first to a six-membered ring closure, and then, after subsequent capture of the emerging tertiary cation by the enol oxygen, to a second ring closure [41].
As in the case described by Sasaki [41], we only obtained the trans-fused hexahydroisochromene 23. This stereoselective nonsynchronous bicyclization may be rationalized taking into account the rule for 1,2-disubstituted cyclohexane compounds. According to that, the thermodynamically more stable conformation is that with more alkyl groups adopting the equatorial position. As displayed in Scheme 3, rotamer I leads to a cis-1,2-disubstituted cyclohexane (one axial and the other equatorial) whereas rotamer II leads to a trans-1,2-disubstituted cyclohexane with both groups in equatorial position. In addition to this energetic argument, it seems that in a trans-1,2-disubstituted eq-eq conformation the oxygen of the enol function and the carbocation centre are likely closer for the second cyclization than that in a cis-1,2-disubstituted eq-ax conformation. The structure of compound 23 was assigned by standard spectroscopic techniques (IR, MS, 1 H-NMR, 13 C-NMR, 2D NMR). It is worth underscoring some details of its 1 H NMR like the upshielded resonance of H-5ax as a q ( 0.59, J 5ax-5eq-4a-6 =12 Hz), due to the magnetic anisotropy of the double bond (Δ 3 ). Furthermore, the chemical shift assigned to H-4a (δ 1.55-1.68) seems to be a br t, where the highest coupling constant is ca. 12 Hz. This is sufficiently consistent with both dihedral angles H4a-C-C-H8a and H4a-C-C-H5ax of 180º, what are the corresponding angles of a trans-fused bicyclic system in which the two hydrogen atoms of the junction carbons are both axial.
In the synthesis of the α,β-unsaturated ketone 27, cis-dihydronopal (18) was reacted in a similar way -aldol reaction with butanone followed by acid catalyzed dehydration. The crude obtained was purified by flash chromatography to afford 27 in 56% yields. With respect to the spectroscopic data of 27, no dramatic changes occur either the C1-C5 moiety, with respect to the analogues19, 20, 22 and 24, or the 2-αH-pinane moiety, with respect to the precursors 17 and 18. A C'-2 epimer of 27 was prepared by Mookherjee and co-workers and described as possessing a powerful sandalwood aroma with urine [47], sweet and floral undertones. In this patent the inventors claimed its use for enhancing the aroma or taste of smoking tobacco and tobacco articles. The ketones 19, 20, 22, 24 (+25) and 27 could be converted into the corresponding ,γ-unsaturated ketones 28-32 by a deconjugative -methylation reaction [23,48,49]. This procedure relies on the initial formation of an enolate, using a slight stoichiometric excess of potassium t-butoxide, followed by the methylation of the resulting ion under conditions that provided the kinetically favoured product in excess over the thermodynamically favoured product. A ten molar excess of cooled iodomethane was added quickly over the cooled (0 ºC) solution of the referred enolate in DMF. The configuration about the double bond in compounds 28-32 was E, as indicated, and the procedure provided those five new enones, which after chromatographic purification yielded pure compounds 28 (77%), 29 (42%), 30 (55%), 31 (84%) [50] and 32 (85%).

Odour evaluation
The independent odour evaluation of each bulky moiety-modified Polysantol ® analogue 33-37 (each over 97% pure according to GC) was carried out by a group of perfumers using two different protocols: (a) at three times from impregnated blotting paper strips (see section 3.6) ( Table 1), (b) after injecting them, separately, onto a GC fitted with sniffing port (Table 2).
Thus, the profile of the (E)-3,3-dimethyl-5-((1S,2S,5S)-6,6-dimethylbicyclo[3.1.1]hept-2-yl)pent-4en-2-ol (37) was identified as the most interesting and promising of the series because of it is full of qualities and it directly emulates the natural sandalwood odour instead of that of synthetic Polysantol ® . For that reason, this compound has recently been claimed as a potential useful odorant [24]. Furthermore, it is noteworthy that 34 and 35 display fairly good behaviour as woody and sandalwood odorants, a fact that supports the hypothesis that structurally rigid molecules interact with a smaller number of olfactory receptor proteins than fairly flexible molecules, which can be assumed to interact with the proteins involved in a more complex manner [52].  Table 2. Odour evaluation of alcohols 33-37 using a GC fitted with sniffing port.

Compounds Odour
Borneol, balsamic, camphoraceus woody notes, but not sandalwood. Also fencholic, slightly valerianic with a note which remembers to wet mossy forest soil at the end.
Woody notes with dryness and amber nuances Iso E Super-type. Also fatty, green, floral and soapy notes, with an animalic and valerianic tone at the end.
Very clean and natural sandalwood note, Polysantol-type and as intense as this.
The woody bouquet is harmonized with amber, balsamic, animalic, sweet, green and a slightly cresolic background.
Woody and mild sandalwood scent. It is also slightly greasy with burn, moist and green nuances, and a reminiscent of citrus fruits at the end.
Very clean, intense and rounded sandalwood note, more natural scent than Polysantol-type. It is also woody, vetiver, green in mossy-type, with animal and vanilla notes at the end.
Chemical shift values are reported in parts per million (ppm,  scale) and coupling constants (J) are in hertz (Hz). All described coupling constants refer to a three-bond coupling distance ( 3 J). 13 C-NMR spectra were recorded on the same instruments (75 or 100 MHz, CDCl 3 , TMS). Chemical shifts are also reported in ppm and carbon substitution degrees were established by DEPT multipulse sequence. 2D NMR experiments (DQF-COSY, HSQC, HMBC, NOESY) were carried out for all compounds of dihydronopol series (17, 18, 27, 32, 37) and for the isochromene 23, on the same instrument. Infrared (IR) spectra were recorder on a FT-IR Perkin-Elmer 1760X spectrometer using a thin film between KBr plates (neat). Mass spectra (MS) were obtained in all cases by GCMS analysis carried out on a Hewlett-Packard 5990 A II gas chromatograph coupled to a Hewlett-Packard 5989B mass spectrometer using the electron impact (EI) ionization method (70 eV); the parameters for the GC unit were the same as those described previously for the GC analyses. High-resolution mass spectra (HRMS) were obtained on a trisector EBE Waters Micromass AutoSpect NT spectrometer using EI (70 eV).

Aldol condensation of 1215 and 18 with butanone to give 19, 20, 22, 24 and 27
(a) With subsequent acidic catalyzed dehydration: a 6.0 M solution of starting aldehydes (1215 and 18) in MeOH (1.0 mL, 6.0 mmol) was added dropwise to a stirred solution of butanone (1.73 g, 24.0 mmol) and KOH (15 mg, 0.25 mmol) in MeOH (1.5 mL) at 0 ºC for 1 h. Then, the mixture was allowed to warm to room temperature and stirring was continued for a further 8 h. The reaction was quenched with a 1N aqueous solution of AcOH (100 mL), the solvent was then partially evaporated in vacuo and the resulting crude diluted with Et 2 O (25 mL) and washed with 1 N AcOH solution (25 mL) and brine (3×25 mL). The crude was dried over anhyd Na 2 SO 4 and evaporated to yield a yellow residue, which was used in the next reaction without further purification. Then, a DeanStark apparatus was fitted to a flask containing a solution of the above aldol crude reaction and TsOH·H 2 O (40 mg, 0.2 mmol) in dry toluene (10 mL), and the mixture was refluxed for 90 min. The solution was allowed to cool down and washed with an aqueous saturated NaHCO 3 solution (3×25 mL), 1N AcOH solution (25 mL) and brine (3×25 mL), dried over anhyd Na 2 SO 4 and evaporated in vacuo. The aldol condensations of butanone (a) with 12 afforded (E)-3,6-dimethylhept-3-en-2-one (19)  (b) With subsequent basic catalyzed dehydration: a solution of 14 (2.0 g, 12.9 mmol) in MeOH (4.0 mL) was added dropwise to a stirred solution of butanone (3.74 g, 51.9 mmol) and KOH (30 mg, 0.5 mmol) in MeOH (2.0 mL) at 0 ºC for 1 h. Then, the mixture was allowed to warm to room temperature and stirring was continued for a further 8 h. A condenser was then fitted to the flask and the mixture heated at ca. 50 ºC for 2 h. The solution was allowed to reach room temperature and quenched with 1N AcOH solution (25 mL). The mixture was extracted with Et 2 O (3×25 mL), and the combined organic extracts were neutralized by washing with brine (3×25 mL). The crude was dried over anhyd Na 2 SO 4 and evaporated in vacuo to afford (E)-3,6,10-trimethylundeca-3,9-dien-2-one (22) in a 52% yield.   (25)

Sensory evaluation
Direct smelling analysis. Blotting paper strips were impregnated with compounds 33-37, previously diluted with Et 2 O (25 mg/200 L), and smelt by perfumers at that moment (after solvent evaporation), 3 h and 24 h later. The olfactory description in each session therefore corresponded to the top, heart and base notes, respectively.
GC sniffing analysis. Odour assessment of compounds 33-37 was achieved by a group of perfumers using a Hewlett-Packard Model 5890 Series II gas chromatograph equipped with a thermal conductivity detector (TCD) and handmade sniffing port. Separation was done with a 10% Carbowax 20M over Chromosorb W/AW 80100 mesh packed column (1.8 m×6 mm OD×2.2 mm ID); injector temperature: 250 ºC; detector temperature: 250 ºC, oven temperature program: 60 ºC (0 min) to 240 ºC (20 min) at 4 ºC/min. Sample size for each injection was approximately 1 L in a 1:10 split mode.

Conclusions
The literature SOR data on the sandalwood olfactophore seem to point that the bulky hydrophobic moiety of odorants such as -santalol (1) and campholenal derivatives 2-5 could be replaced by substructures of similar steric bulk. Thus, new five bulky moiety modified analogues 33-37 of the commercial sandalwood odorant Polysantol ® (2) have been synthesized. Starting from the aldehydes isovaleraldehyde (12), heptanal (13), citronellal (14), phenylacetaldehyde (15) and dihydronopal (18), and by an expeditious sequence of aldol condensation with butanone, deconjugative -methylation of the resulting -unsaturated ketones, and reduction of the corresponding -unsaturated ketones, the new five analogues were prepared in good yield. These compounds 33-37 were organoleptically evaluated and one of them (compound 37) seemed to be of special interest due to its natural sandalwood scent, which means that the dihydronopyl group is able to mimic the bulky hydrophobic center C of the sandalwood olfactophore. The other synthesized alcohols do not seem to be of interest as odorants, although the branched-chain citronellal derivative 35 and the aromatic-ring phenylacetaldehyde derivative 36 have some sandalwood notes, at least according to the GC sniffing odour evaluation.