Efficient synthesis of 4-halo-D 3 -trishomocubane derivatives

Reaction of C s -trishomocubane-8,11-diol with hydrohalic acids results in haloalcohols – 4,7-disubstituted derivatives of D 3 -trishomocubane. The reaction involves the rearrangement of the trishomocubane skeleton and the stereoselective formation of 4,7-D 3 -trishomocubanediol and 5‐oxahexacyclo[5.4.1.0 2,6 .0 3,10 .0 4,8 .0 9,12 ]- dodecane as side products. The mechanism of the rearrangement was proposed. 7-Halo-D 3 -trishomocuban-4- ols were oxidized to the corresponding haloketones. The structure of the major isomers of the haloketones was confirmed through X-ray analysis


Introduction
The chemistry of polycyclic cage compounds remains fascinating and continues to draw the attention of organic chemists since the middle of the last century.Interest in the pharmacology of polycyclic cage compounds was stimulated when the wide range of pharmacological properties of adamantane derivatives have been discovered. 1 Since then, plenty of polycyclic hydrocarbon three-dimensional scaffolds were developed.Some of them, like bicyclo[1.1.1]pentane or cubane are considered to be key bioisosteres of traditional flat aromatic systems. 2,3 3 -Symmetrical trishomocubane (pentacyclo[6.3.0.0 2,6 .0 3,10.0 5,9]undecane) has attracted particular attention, because it has specific properties valuable for novel drug design, viz., a relatively large cage size, high lipophilicity, and a conformational rigidity.5][6] Furthermore, unlike the above cages, D 3 -trishomocubane is internally chiral, which may also considerably increase the drug efficiency. 7owever, despite the fact that D 3 -trishomocubane derivatives have great potential for drug discovery, their syntheses often remain tricky and the mechanisms of the transformations are still unclear.In particular, rearrangement of C s -trishomocubane-8,11-diol 1 in acidic media, which is the main method of synthesis of 4,7disubstituted D 3 -trishomocubanes, leads to a mixture of products, and the ratio of them strongly depends on the conditions [8][9][10] .Thus, the result of the interaction of diol 1 with hydrobromic or hydroiodic acid at 100 o C is the formation of a mixture of hexacyclic ether 2 and 7-halo-D 3 -trishomocubane-4-ol diastereomers 3a,b or 4a,b.At the same time, reaction of C s -diol 1 with HI under harsh conditions (160°C, 20 h) 8 or of the tosylate of this diol with NaI in HMPA 10 results in a mixture of diastereomers of diiodide 5, an important compound for the synthesis of homohypostrophene, D 3 -trishomocubane, and their derivatives.Another problem is the instability of the yields of 7-iodo-D 3 -trishomocubane-4-ols (4a,b, Scheme 1), which are varying in a very wide range (from 44% up to 86%) for unclear reasons (stated by Kent 8 and Helmchen 9 , confirmed by us 11 ).Scheme 1. Reaction of C s -trishomocubane-8,11-diol (1) with HI and HBr.Note, 7-halo-D 3 -trishomocubane-4-ols are starting materials for the synthesis of D 3 -trishomocubanone and chiral D 3 -trishomocubane-carboxylic acid 11 , valuable intermediates for a wide range of D 3 -trishomocubane derivatives.Because much in the C s -trishomocubane-8,11-diol rearrangement under the acidic conditions is still unclear, we decided to reinvestigate this reaction more thoroughly and gain a deeper insight into its mechanism.Optimization of the rearrangement step would increase the overall yield for these compounds, too.
First, we carefully investigated the rearrangement of C s -trishomocubane-8,11-diol 1 in hydroiodic and hydrobromic acids.In our hands, the reaction of 1 with hydroiodic acid gave nearly the same results (with insignificant differences in yields and product ratios) as reported previously. 8,9,11In the case of aqueous hydrobromic acid, the rearrangement of 1 (Scheme 1) with usual work-up of the reaction mixture (poured into water, extracted by an organic solvent that was then dried and evaporated) results in the mixture of two diastereomeric 7-bromo-D 3 -trishomocubane-4-oles 3a,b in 3:1 ratio (63% yield) with admixture of ether 2 (11% yield).Evaporation of the water layer under normal pressure results in the formation of a chloroform soluble mixture of diastereomeric bromoalcohols 3a,b and dibromides 7 (approx.7% yield) (Scheme 2).We suggested that D 3 -trishomocubane-4,7-diol 6 can be formed as a by-product of the rearrangement, be wellsoluble in aqueous acidic media (unlike C s -diol 1, 7-bromo-D 3 -trishomocubane-4-oles, and hexacyclic ether 2) and under heating undergo nucleophilic substitution of the hydroxyl group(s) with bromine.This suggestion can explain the solubility of initially insoluble residue in organic solvent after further boiling in HBr.Scheme 2. Reaction of D 3 -trishomocubane-4,7-diol (6) with HBr.
To prove our assumption about the structure of the water-soluble admixture, we attempted the rearrangement of the C s -diol 1 while boiling in hydrochloric acid (Scheme 3), followed by the treatment as discussed above (extraction with CHCl 3 , evaporation, etc.).Analysis of the products from the organic phase revealed the presence of the hexacyclic ether 2 and two chloroalcohols 8 in a 5:1 ratio (these diastereomers were separated by crystallization).At the same time, evaporation of the aqueous phase under high vacuum resulted in the formation of a white hygroscopic precipitate.After drying, the precipitate was subjected to NMR and GS/MS analysis and was identified as compound 6.The yield of 4,7-D 3 -trishomocubanediol 6 is ca.20-25%.Difference in ratios of stereoisomers of halo-alcohols (3:1 for 3a,b and 5:1 for 8a,b), formed after the reaction of Cs-trishomocubane-8,11-diol with hydrohalic acids, might be explained with different nucleophilicity of the halide anions.Scheme 3. Reaction of C s -diol 1 with hydrochloric acid.
Despite the fact that the synthesis of 4,7-D 3 -trishomocubanediol 6 was described earlier in a few articles, there is not much physico-chemical data on compound 6 -only IR spectra was reported by Barborak 12 and later by Naemura. 13Based on 13 C NMR, it was concluded that while compound 6 can exist as three different isomers (one С 1 -and two С 2 -symmetrical, Figure 1), in the course of the reaction formation of only one, 6a, takes place.While structures 6b and 6c would have only 6 signals (or 12 in the case of their mixture), in the 13 C spectra there are 11 signals, that correspond to structure 6a.It turned out that upon continuous reflux in hydrochloric acid of hexacyclic ether, 2 also undergoes rearrangement to give a mixture of chloroalcohols 8a,b in approx.4:1 ratio.However, the reaction proceeds slower than in the case of diol 1. Treatment of ether 2 with AlCl 3 in dichloromethane leads to the same results.Obviously, the mechanism of rearrangement of the C s -diol 1 in hydrohalic acids suggested earlier by Kent (via a concerted halide-displacement/protonation), 8 does not explain the formation of all the observed products.It predicts the formation of only one stereoisomer of 7-halo-D 3 -trishomocubane-4-oles while 2 isomers are observed in the mixtures and does envisage the formation of 4,7-D 3 -trishomocubanediol 6.We suggested, that the mechanism can have additional alternative pathways (Scheme 4), which can be realized via formation of a nonclassical + C 4 D 3 -trishomocubane (type A).In our previous work 14 we have suggested that a similar rearrangement occurs via the formation of an analogous nonclassical cation.We assume that the formation of cation A from diol 1 involves protonation of 1, dehydration, and further rearrangement of the protonated hexacyclic ether 2. Nucleophilic attack of A by a halogen leads to the formation of the major isomer of haloalcohol (3, 4, or 8).The competing reaction of A with water affords protonated diol 6, where the hydroxyl groups can be further substituted with halogen(s) to furnish the minor isomer of haloalcohol and dihalide.

Scheme 4. Proposed rearrangement pathway.
Hydrochloric acid is seen to be the most convenient among the hydrohalic acids for the C strishomocubane-8,11-diol rearrangement because of its high thermal stability and relatively low boiling point.This makes chloroalcohol 8 the most easily accessible haloalcohol and, in advance, the most preferable starting material for further syntheses.
Therefore, we decided to develop the procedure for the chiral resolution of its enantiomers.So, the major isomer of chloroalcohol 8a, obtained after the fractional crystallization of 8a,b from acetone, was treated sequentially with phthalic anhydride and (R)-(+)-α-phenylethylamine to obtaine a mixture of diastereomeric esters 10 (Scheme 5).The mixture was separated by fractional crystallization from acetone and the isolated major diastereomer was subjected to basic hydrolysis to furnish the pure enantiomer of 7-chloro-D 3 -trishomocubane-4-ol ( =+69.3(c 1.17, CHCl 3 ) in 35% yield.The ratio of the haloketone diastereomers was 2:1 for the chloroketone and 4:1 for the bromo-and iodoketone, the major isomers were isolated from the mixture by crystallization.2D-NMR of the main diastereomers of the haloketones show an anti-orientation of the halogen.Such a structure was confirmed through X-ray analysis of the bromoketone.The bromide substituent has an exo-position in relation to the cubane fragment (the C2-C3-C4-Br1 torsion angle is -179.5(1)°).

Conclusions
We have reported a new synthetic pathway leading to the stereoselective formation of 7-halo-D 3trishomocuban-4-ones and 4,7-D 3 -trishomocubanediol.On the basis of the experimental results, a possible reaction mechanism has been proposed.The obtained 7-halo-D 3 -trishomocuban-4-ols were oxidized to the corresponding haloketones.The structure of the major isomers of the haloketones was confirmed through Xray analysis.

Experimental Section
General. 1 H and 13 C NMR spectra were recorded using BrukerAvance NMR spectrometers operating at 400 and 500 1 H frequency (101 and 126 MHz for 13 C experiments).Chemical shifts are reported relative to internal TMS ( 1 H) standard.Melting points are uncorrected.Solvents were dried before use according to standard methods.Elemental analysis was carried out in the analytical laboratory of Institute of Organic Chemistry, NAS of Ukraine.

Figure 2 .
Figure 2. Molecular structure of compound 6a according to X-ray diffraction data.Thermal ellipsoids are shown at 50% probability level.

Figure 3 .
Figure 3. Molecular structure of compound 11 according to X-ray diffraction data.Thermal ellipsoids are shown at 50% probability level.