Dual organocatalysis in enantioselective synthesis of bicyclo[3.2.1]octanes

Get Complete Project Material File(s) Now! »

Single step enantioselective construction of bicyclo[3.2.1]octanes

A highlight-type article from our group has recently appeared in the literature on this topic,44 and only a synopsis of the work described therein, together with the most recent work will be described here.
Although several approaches to optically active bicyclo[3.2.1]octane derivatives have been described from the chiral pool or by multi-step synthesis,45 only a few approaches are available for their single-step enantioselective synthesis. Among these, we can mention the early work of Engler and co-workers, who achieved enantioselective 6 (5+2) cycloadditions of 1,4-benzoquinones with aromatic olefins using a titanium-TADDOLate complex catalyst (Scheme 16).46 A few other metal-catalyzed transformations were also developed.
As to organocatalytic approaches, a few systems were described early on over the 2000– 2007 period, but none of these were found both general and efficient. A significant step further was accomplished in 2009 with the work of Tang, Li and their co-workers who described a general approach to bicyclo[3.n.1] derivatives involving an interesting Michael / elimination / Michael cascade sequence catalyzed by a bifunctional aminocatalyst operating by enamine / H-bonding activation (Scheme 17). 47 Although a single example of a bicyclo[3.2.1]octane product was reported in the study, it was obtained in high yield, diastereo- and enantio-selectivity.

Enantioselective synthesis of bicyclo[3.2.1]octanes

From the above preliminary results, we selected the reaction between tert-butyl 2-oxo-cyclopentane carboxylate and crotonaldehyde with the Jørgensen catalyst 12 and the NHC catalyst IDipp (2) for the optimization study (Scheme 25).
As shown above, an oxidation step of the bicyclo[3.2.1]octanol product 7a into the diketone 16a is required to facilitate the stereochemical analyses. Previous conditions for this reaction using IBX (Scheme 23) were not satisfactory, affording only a moderate yield of the diketone product 16a. So, I started my work with the optimization of this oxidation reaction. Obviously, neutral or moderately acidic conditions are required to avoid the base-promoted epimerization of the product 7 by retro-aldol / retro-Michael processes. For this purpose, we selected pyridinium chlorochromate (PCC) as a potential oxidant candidate. We have rapidly found that in the presence of anhydrous magnesium sulfate, the desired diketone 16a could be obtained in nearly quantitative yield from the corresponding alcohol 7a (Scheme 26). Also, we have checked if the crude product 7a could be used directly in this reaction in order to save time and solvents. A comparison of the results obtained from crude and purified 7a afforded very similar results with a little preference for the reaction performed directly from crude 7a (Scheme 26).

Total syntheses of quadranes

Due to their potent biological activities and challenging molecular architectures, quadrane sesquiterpenes have soon attracted great attention from the synthetic community. This area of research was comprehensively reviewed recently,63 and only the most important work will be highlighted here.
Danishefsky first reported the total synthesis of (±)-quadrone (I) in 1980, just two years after its isolation (Scheme 29)72. The synthesis started from 4,4-dimethylcyclopentenone as the precursor of ring B with a consecutive copper(I)-promoted conjugate addition of a vinyl Grignard / enolate trapping with a functionalized alkyl iodide to give intermediate 19. Five additional steps were required to prepare the AB ring system 20, notably involving a base-promoted aldolization / crotonization sequence for the construction of ring A. The quaternary carbon atom at C1 was installed by a conjugate addition of a silyl ketene acetal, and some functional groups manipulations afforded intermediate 21. The elaboration of the six-membered C ring ensued from a remarkably diastereoselective cyclization of the lithium enolate derived from 21 to give the ABC ring system of quadranes, and a deprotection step afforded intermediate 22. The-alkylation at C5 of the ketone group in 22 proved very difficult due to regioselectivity issues, the3,4 enolate being formed preferentially. This problem was circumvented by the temporary introduction of the2,3 unsaturation with concomitant saponification of the methyl ester at C8 to give the enone intermediate 23. Then the desired4,5 enolate could be generated and trapped by formaldehyde, and the2,3 double bond diastereoselectively hydrogenated from the face to give the hydroxyacid intermediate 24. Amusingly, compound 24 was later isolated as a naturally occurring quadrane.73 The dehydration of 24 afforded (±)-terrecyclic acid A (II), while (±)-quadrone (I) was obtained under thermolytic conditions directly from 24. Overall, this total synthesis of (±)-quadrone (I) was achieved in 19 steps and 3.1% yield from 4,4-dimethylcyclopentenone.

Retrosynthetic analysis for (–)-isishippuric acid B (VI)

It was initially devised that (–)-isishippuric acid B (VI) could be obtained from the unsaturated diester 35 by a saponification / hydrogenation sequence (Scheme 32). Intermediate 35 would be obtained directly from the ketone 36 by a Wittig-like olefination, and the ketone 36 would derive from the product 7b obtained during the methodological studies presented in Chapter 1. While the above retrosynthetic analysis is very seducing on the paper, several of its features should be highlighted:
– the ketone group at C2 is to a quaternary center resulting in kinetically disfavored nucleophilic additions to this moiety.
– a severe 1,3-diaxial interaction exists between the methyl group at C8 and the electrophilic carbon atom at C2 rendering very difficult nucleophilic additions at C2 from the face.
– a bulky gem-dimethyl is present at C13 rendering very difficult nucleophilic additions at C2 from the face.
– the bicyclic ring system in 36 is rigid and shows no conformational flexibility.
Dr. Marc Presset, a former Ph.D. student in our group, has extensively studied the synthetic route presented in Scheme 32 in the racemic series.56 For the reasons mentioned above, all attempts to perform a Wittig-type olefination at C2 on intermediate 36 and its analogs failed, leaving the starting material unchanged in most cases. After considerable efforts several analogs of the unsaturated intermediate 35 could be prepared. Again, the reduction of the2,3 double bond in 35 and its analogs proved extremely challenging, and no conditions could be identified at the time to perform this reaction. Some nucleophilic substitutions at C2 were also examined from the activated alcohol (e.g. the corresponding mesylate) derived from 36, and in this case again, no suitable conditions could be identified.
From this accumulated knowledge, and despite the many unsuccessful attempts, it was decided to push forward our work on the total synthesis of (–)-isishippuric acid B (VI) taking advantage of some recently described methods for the manipulation of the sterically hindered double bonds as the2,3 double bond in 35 and its analogs. Our efforts in this direction are detailed in the next section.

READ  Hybrid Deep Recurrent Neural Networks / HMM for Handwriting Recognition 

Synthetic approach to (–)-isishippuric acid B (VI)

As detailed in Chapter 1, we have developed a method to prepare optically active bicyclo[3.2.1]octane derivatives based on a bicatalytic enantioselective organocascade combining iminium activation with NHC activation. Using this method on a 10-gram scale with the previously developed conditions, we were able to obtain compound 7b in 90% yield from the-ketoester 5b and crotonaldehyde using a combination of catalyst 12 and the NHC IDipp 2 in methanol (Scheme 33). The determination of the diastereoselectivity and enantioselectivity of the reaction was realized as before on the diketone product 16b obtained in 89% from 7b (80% over the two steps). Surprisingly, the diketone 16b was obtained as a mixture of two diastereomers with a reverse diastereoselectivity (dr = 1:4.4 Meax:Meeq), the diastereomer with the equatorial methyl group being largely major. Moreover, the desired diastereomer with the axial methyl group (the minor one in this case) proved almost racemic (er = 1.6:1). This unexpected difference between the results obtained on a 50–100 mg scale and on a 10 g scale forced us to look deeper to the reaction conditions, especially the evolution of the internal temperature during the reaction. It was rapidly found that the Michael addition step is relatively exothermic, and the observed loss of stereoselectivity was attributed to the uncontrolled elevation of the temperature during the early stages of the reaction producing less stereoselective Michael additions and probably some NHC-catalyzed retro-Michael / Michael processes leading to racemization. Accordingly, we modified our protocol as follow:
– the reaction mixture containing 5b and catalyst 12 was diluted to 0.5 M instead of 2.5 M, cooled down to –30 oC, and vigorously stirred to dissipate efficiently the evolved heat.
– the crotonaldehyde was added dropwise over 1h at –30 ºC and the resulting reaction mixture was kept at that temperature for 4 h and then allowed to warm slowly to 0 ºC.
Using this modified protocol on a 10-gram scale, we could reproducibly obtain the same results with small scale reactions, and the product 16b was obtained in good yield (80% over two steps) with acceptable stereoselectivities (dr Meax:Meeq = 2:1, er Meax = 8:1) to continue the synthesis.

Functionalization at C2

As mentioned above, Dr. Marc Presset has already accomplished a large body of work on these transformations in the racemic series,56 and my work builds up on his findings. Notably, it was previously found that the ketone group at C2 in 36 could undergo diastereoselective nucleophilic additions with vinyl Grignard reagents from the face to give the corresponding adduct 38 (Scheme 35, top). Surprisingly, and in apparent contradiction with earlier results (see for example the hydrogenation step in Scheme 29), this reaction indicates that the gem-dimethyl group at C13 induces less steric effect than the axial methyl group at C8. The reaction however required some thermal activation to proceed efficiently. Next, an allylic transposition from 38 allowed installing the desired double bond2,3 as in the products 39 and 40 (Scheme 35, bottom). A number of conditions were then tested to perform the diastereoselective hydrogenation, or more generally the reduction, of the double bond in 39 and its derivatives. A few attempts were also performed for the rhodium-catalyzed isomerization of the allylic alcohol moiety in 40 into the corresponding saturated aldehyde. Regrettably, no suitable conditions could be identified, compound 39 and its derivatives being extremely stable. This chemical inertness of these compounds toward hydrogenation and isomerization reactions was attributed, again, to the very crowded environment of the2,3 double bond.

The Arndt–Eistert homologation

The synthesis started with the ketone 36 obtained as described above (Scheme 34). The preparation of the retrosynthetic intermediate 61 from 36 requires the homologation of the ester moiety together with its conversion into a methyl ketone moiety. The Arndt–Eistert homologation,90 involving the reaction of an activated carboxylic acid with diazomethane and subsequent Wolff rearrangement91 of the intermediate diazoketone in the presence of a nucleophile, was selected for this purpose. Thus, the tert-butyl ester in 36 was saponified under acidic conditions to give the corresponding carboxylic acid 62 quantitatively (Scheme 50). The resulting clean crude material was then converted into the corresponding acyl chloride and allowed to react in situ with trimethylsilyldiazomethane to afford the diazoketone 63 in 75% yield. The silver catalyzed Wolff rearrangement of 63 at 80 ºC (no reaction occurred at lower temperature) in the presence of water unexpectedly afforded the ring-expended bicyclic product 64 and not the desired homologated carboxylic acid. A plausible mechanism for the formation of 64 is depicted in Scheme 50.

Table of contents :

Résumé du travail
Chapter 1 Dual organocatalysis in enantioselective synthesis of bicyclo[3.2.1]octanes .
1.1. Introduction
1.1.1. Organocatalysis: concepts, main activation modes & multi-catalysis
1.1.2. Objective of the research project
1.1.3. Single step enantioselective construction of bicyclo[3.2.1]octanes .
1.2. Original enantioselective bi-organocatalytic synthesis of bicyclo[3.2.1]octanes
1.2.1. Preparation of substrates
1.2.2. Early work
1.2.3. Enantioselective synthesis of bicyclo[3.2.1]octanes
1.2.4. Conclusion
Chapter 2 Total synthesis of quadrane sesquiterpenes
2.1. Introduction
2.1.1. Quadranes natural products
2.1.2. Total syntheses of quadranes
2.1.3. Retrosynthetic analysis for (–)-isishippuric acid B (VI)
2.2. Synthetic approach to (–)-isishippuric acid B (VI)
2.2.1. Construction of the bicyclo[3.2.1]octane
2.2.2. Functionalization at C2
2.3. Total synthesis of (+)-suberosanone (IV) and formal synthesis of (+)-suberosenone (III)
2.3.1. Retrosynthetic analysis
2.3.2. The Arndt–Eistert homologation
2.3.3. Completion of the syntheses
2.4. Summary and conclusion
Chapter 3 Enantioselective synthesis of glutarimides
3.1. Introduction
3.1.1. Glutarimides in drugs and natural products
3.1.2. Enantioselective routes to glutarimides
3.1.3. Our approach
3.2. Enantioselective synthesis of glutarimides
3.2.1. The racemic series
3.2.2. The optically active series
3.3. Conclusion
General conclusion
Experimental part
ES.0. General experimental information
ES.1. Chemistry of bicyclo[3.n.1]octanes
ES.2. Total synthesis of quadrane sesquiterpenes
ES.3. Chemistry of glurarimides

GET THE COMPLETE PROJECT

Related Posts