Table of Contents:
At the beginning of the present work, syntheses of highly functionalized cyclohexenon
derivatives by alkylation were examined, following procedures of J. BAUMEISTER. The first
test system for this alkylation reaction was cyclohexenon derivative 84, familiar from the
literature, which was prepared in three steps with an overall yield of 24% (Scheme 1).
Scheme 1: Preparation of cyclohexenon derivative 84.
Furthermore the silylenol ether 95 was synthesised by a solvent-free DIELS-ALDER-reaction in
a very good yield of 95% (Scheme 2). Proceeding from the precursor 95 preparation of
cyclohexenon derivatives 85 was achieved in 96% yield, 86 in 46% over three steps and 87
in 43% over three steps respectively.
With those four cyclohexenon derivatives in hand the regioselective α-alkylation was
investigated (Scheme 3). Various reaction conditions (bases, temperature, solvents and
additives) were tested; however the desired product could not be isolated. The
corresponding double alkylated side product was sometimes obtained in moderate yields.
Scheme 3: Desired reaction of the cyclohexenon derivatives to the corresponding α-alkylated products 82.
Since no suitable alkylation method could be found, a synthesis following SCHEIDT’s
ocilactomycin synthesis for the preparation of the highly functionalized cylohexenon
derivative by DIELS-ALDER-reaction as the key step was envisioned. (R)-Citronellal (78) was
chosen as starting material for the preparation of the diene, and the isolated stereogenic
centre is therefore derived from the chiral pool (Scheme 5). In a four-step sequence starting
from 78 (25 g), mesitylate 110 was obtained in 67% overall yield by addition of MeMgBr,
benzoyl protection, ozonolysis (with reductive work-up) and by transforming the primary
alcohol into the desired mesitylate (Scheme 4).
By reductive benzoyl deprotection, PCC oxidation and TEBBE-reaction, the mesitylate 114
was obtained in three steps with 48% overall yield (Scheme 5). Up to this step all reactions
were performed in a multi-gram scale. SN2-Reaction with sodium cyanide and subsequent
reduction of the nitrile gave access to the aldehyde 116 in 88% yield.
HORNER-WADSWORTH-EMMONS-Reaction under PATERSON-conditions followed by
WITTIG-reaction finally led to diene 102 in 78% yield over two steps and starting from
(R)-Citronellal (78) in 22% over eleven steps.
Diene 102 was used in a Me2AlCl-mediated DIELS-ALDER-reaction together with the
dienophile EVANS-auxilary 55 and the desired product 101 was obtained in 75% yield and
high diastereoselectivity (Scheme 6).
The auxiliary was cleaved reductively by LiBH4, and, after DMP-oxidation, aldehyde 124 was
obtained in nearly quantitative yield.
Starting from commercially available carboxylic acids 228, the methyl β-ketoesters 230 were
prepared first. Via REGITZ diazotransfer and oxidation with tert-BuOCl the hydrates 278 could
be obtained (Scheme 7). The corresponding cyclohexyl β-ketoesters 264 were prepared from
230 by transesterification. By diazotransfer and oxidation the hydrates 279 were accessible.
These three or four steps respectively could be executed in a multi-gram scale with good
overall yields, and storage of the hydrates was possible.
The hydrates 278 or 279 were dehydrated to the corresponding α,β-diketoesters 282 and
283 by bulb-to-bulb distillation or by molecular sieves (Scheme 8). These vicinal tricarbonyl
compounds (VTC) were freshly prepared prior to use in alkylation/ arylation reactions with
ZnMe2, ZnEt2 or ZnPh2. A selection of prepared substances is shown together with the yields
obtained. On the whole good yields were obtained for those alkylation/ arylation reactions,
whereby substituents in γ-position have a great impact on the reaction. It becomes clear that
good yields could be obtained in the case of electronic neutral or electron-rich
α,β-diketoesters, whereas low yields were obtained for electron-poor derivatives. Sterical
influences appear to have a minor impact on the yields and no significant differences were
observed between methyl esters and the corresponding cyclohexyl esters.
The alkenylation of VTC proved to be difficult, and many conditions had to be tested before
this reaction was successfully accomplished using zincates. Alkenyl-dimethyl-zincates
enabled the desired reaction, however methyl transfer was also observed. In order to force
back the side reaction as much as possible, sterically demanding neopentyl rests were used
instead of methyl groups. As shown in Scheme 9 and starting from different vinyl iodides
(307, 309 or 311), lithiated species were prepared first by iodine lithium exchange. Mixed
alkenyl-dineopentyl-zincates 328 were obtained next and by addition of the VTC 282 or 283
respectively their alkenylation could be achieved.
Selected examples of substances prepared via this way are shown together with the
corresponding yields in Scheme 10. Substituents in γ-position once again have a great
impact on the reaction, and the best yields were obtained for α,β-diketoesters without
enolizable protons adjacent to the outer carbonyl group. Good yields were obtained for
electronic neutral or electron-rich VTCs, whereas reactions with electron-poor derivatives did
not yield the desired products. No significant differences were observed between methyl
esters and the corresponding cyclohexylesters. For the vinyl iodides good yields were
obtained in the case of the 2,2-disubsitituted derivative 307, whereby E-vinyliodide 309 gave
moderate to good yields, while reactions with Z-vinyl iodide 311 gave low yields only. In the
case of vinyliodide 307, products with two stereogenic centres were obtained. Interestingly,
the diastereomeric ratio is unequal 1:1. Depending on the substituents separation of both
diastereomers via column chromatography was possible; nevertheless elucidation of the
relative orientation of the stereo centres was not possible by NMR. It did, however, succeed
in one case using X-ray analysis (Abbildung 8/ Figure 7). Reactions with 309 and 311
respectively showed that this reaction is stereospecific in regard to doublebond geometry.
For broader applicability this reaction should be transferred to trisubstituted vinyliodides, for
example. Vinyliodides with functional groups might as well be of interest as a transfer of
alkyn anions. It might be accomplished by lithiation and subsequent transformation into the
In future the alkylation/arylation as well as alkenylation reactions should be designed
enantioselectively, in order to utilise them for the synthesis of complex target molecules. One
possibility could be the use of chiral ligands (Scheme 11).
The starting point for the development of suitable ligands might be chiral amino-alcohols,
which are shown in Schema 77. These ligands are used for enantioselective addition
reactions of dialkylzinc reagents to aldehydes, and a transfer to VTC appears possible.