|Home | About | Journals | Submit | Contact Us | Français|
Short, practical and scalable syntheses of (±)-7-methylomuralide and (−)-7-methylomuralide have been developed. Three consecutive tandem reaction pairs establish all the carbons and the stereochemistry of the target molecule, vastly simplifying the synthetic scheme from N-trichloroethoxycarbonyl glycine. The chiral directing group controls the absolute stereochemistry of the key aldol reaction.
Lactacystin (1) was discovered by Õmura et al. using an in vitro screen for secondary metabolites that exhibit nerve growth factor-like activity against a murine neuronal cell line.1 Subsequently, it was determined that the observed stimulation of neurite outgrowth shown by 1 stemmed from its inhibition of the 20S proteasome by acylation mainly at the N-terminal threonine of the β5 (chymotryptic) subunit.2 The active small molecule in this process is the β-lactone omuralide3 (clasto-lactacystin, 2), of which 1 is actually a precursor.2,4
Proteasome inhibition constitutes a useful means for modulating and studying a variety of cellular processes, including mitosis, heat shock response, and antigen presentation.5 Of current clinical interest is the effect of proteasome inhibition on cancer cells, as it has been demonstrated to lead to cell cycle arrest and apoptosis by preventing the degradation of I-κ B and ultimately hindering transcription.6 Consequently, several proteasome inhibitors are being evaluated in the clinic for the treatment of hematological malignancies. Velcade® (bortezomib) is one such drug currently prescribed for multiple myeloma.7
Both 1 and 2 are commercially available through purveyors of fine chemicals. However, their high cost ($1,500/mg and $2,360/mg, respectively; Sigma-Aldrich), and the lengthy syntheses required for their preparation underscore the need for simpler and more economical routes to this class of proteasome inhibitor. Herein, we report a concise synthesis of 7-methylomuralide (3), a simpler, but nearly equipotent analogue3 of omuralide (2), using a route which allows for the rapid and inexpensive preparation of this important molecule.
The potency of 3 has inspired two prior syntheses from these laboratories (13–15 steps).8 However, to maximize material through-put, a shorter route to 3 was envisaged whereby intermediate 4 could be accessed through only two stereogenic reactions: directed reduction of C6, and a stereoselective aldol reaction to establish C5 and C9. Unfortunately, initial attempts at direct bond formation with isobutyraldehyde failed due to the instability of aldol-product 6, which undergoes rapid reversion to starting material (vide infra) under a variety of conditions (Figure 1).
However, it was possible to override this tendency for reverse aldolization by using as intermediates the imides 7a–c (Fig. 2A). These were synthesized in one step by treating the corresponding N-carbamoyl glycine O-methyl esters with 3.1 equiv. LHMDS and 1.1 equiv. dimethylmalonyl dichloride (cf. Scheme 1 and Supporting Information). Reaction of the Bocimide 7a with base and isobutyraldehyde gave solely and in excellent yield aldol ester 9a (mp 136–138 °C), the structure of which was determined by X-ray crystallographic analysis (see insert 9a, in Fig. 2A). The diastereoselectivity of this aldol reaction is opposite to that required for the synthesis of 3. This problem was overcome by the use of the trichloroethoxycarbonyl (Troc)-imide 7c and a 25-fold excess of isobutyraldehyde, which afforded the required aldol carbonate ester 8c (mp 116–118 °C) in 70% isolated yield and >10 : 1 predominance over the diastereomer 9c. In the case of benzyl carbamate 7b, the aldol/acyl migration reaction9 is not strongly diastereoselective (Fig. 2A) between the 8 and 9 diastereomers. The X-ray crystal structure of 8b (mp 136–138 °C) is shown in Fig. 2A; aldol esters 8b and 8c were correlated by conversion to the acid8a corresponding to (±)-3 (vide infra). The dependence of stereoselectivity on aldehyde concentration for 7 c is a good indication that the rate-determining step in formation of 8c is the aldol reaction, while the rate-determining step in formation of the 9c is the acyl-transfer (Figure 2B).
In accord with previous observations, the alcohol derived from 8c by Troc removal with zinc metal proved unstable due to retro-aldol cleavage and was never detected; only lactam 7d (Scheme 1A) could be isolated from the reaction mixture. However, if the same reaction was conducted in the presence of sodium triacetoxyborohydride (10 equiv.), the nascent alcohol was efficiently trapped as the boronate by displacement of an acetate ligand on boron. This borohydride derivative then internally reduced the adjacent ketone function, producing dihydroxy ester 4 diastereoselectively and in excellent yield (91%). Saponification and lactonization4,8 yielded (±)-3, whose spectroscopic characteristics were identical to those previously reported.8
In order to synthesize enantiomerically pure lactone (−)-3, it was necessary to render the tandem aldol/acyl-transfer step asymmetric. To this end, we turned to the practical chiral directing group 10 (Scheme 1B), previously developed in these laboratories for enantioselective Diels-Alder reactions of unsaturated esters and enantioselective alkylation of ester enolates.10 It was anticipated that the van der Waals interactions similar to those described in previous studies would organize the pre-transition state assembly for aldol bond formation. Use of the controller 10 in the key aldol reaction was predicted to provide the required absolute configuration for the synthesis of (−)-3.
The controller was appended by carbodiimide-mediated coupling of enantiomerically pure alcohol 1010 with N-Troc glycine. Lactam formation (11 → 12) was effected in even higher yield (89%) than in the route shown in Scheme 1A, via a deep red intermediate (possibly the dianion of 11).
The 10-based controller effectively screened one face of enolate 13 in the aldol reaction in favor of diastereomer 14 (8 : 1 selectivity at C5). This selectivity can be attributed to the previously proposed10 van der Waals attraction between the electron-rich potassium enolate and the appendant electron-poor 2-naphthyl-sulfone, which favors the pre-transition state assembly 13. Even though the selectivity due to differentiation of the prochiral aldehydic carbon was 4 : 1, overall the reaction was efficient enough to provide isomer 14 (mp 178–180 °C) in 61% yield after recrystalization from methanol, which removed the three minor diastereomeric impurities.
The subsequent tandem Troc removal/intramolecular reduction also proceeded in high yield (95%) to afford diol 15, which could be saponified to (−)-1 6 (1 0 recovered quantitatively) and lactonized to form optically pure lactone (−)-3.8b
In summary, short, practical and scalable syntheses of (±)-3 and also (−)-3 have been developed. The approach relies on three consecutive tandem reaction pairs to establish all the carbons and the stereochemistry of the target molecule, vastly simplifying the synthetic scheme to a mere five or six steps from glycine carbamates. Of note is the use of chiral directing group 10 to control the absolute stereochemistry of the key aldol reaction, suggesting that this easily prepared chiral controller may be considerably more useful than previously indicated.
Experimental procedures and characterization. This material is available free of charge via the Internet at http://pubs.acs.org.
We thank the National Institutes of Health for a Postdoctoral Fellowship to R.A.S., Dr. Douglas Ho for X-ray diffraction analysis, and the Saghatelian and Jacobsen labs for assistance in collecting mass and IR spectral data, respectively.