Tetrahedron Letters 52 (2011) 329–331
Contents lists available at ScienceDirect
Tetrahedron Letters journal homepage: www.elsevier.com/locate/tetlet
Asymmetric route to spirotetracyclic (1S)-50 ,110 -dihydro-3Hspiro[cyclopentane-1,100 -dibenzo[a,d]cyclohepten]-3-one derivatives Massimo Gianotti ⇑, Daniele Andreotti, Davide Casotto, Mario Mattioli, Anna Mingardi, Francesca Pavone, Roberto Profeta, Filippo Valente Neurosciences CEDD, GlaxoSmithKline, Medicines Research Centre, Via A. Fleming 4, 37135 Verona, Italy
a r t i c l e
i n f o
Article history: Received 15 October 2010 Revised 4 November 2010 Accepted 9 November 2010 Available online 27 November 2010
a b s t r a c t A new synthetic pathway to the final compound 1-[(1S)-5’,11’-dihydrospiro[cyclopentane-1,10’dibenzo[a,d]cyclohepten]-3-yl]-3-azetidinecarboxylic acid (1) was set up. The two preparative chiral HPLC separations needed in the first route were successfully avoided and replaced by two diastereomeric crystallizations of intermediate compounds 5 and 10. Ó 2010 Elsevier Ltd. All rights reserved.
Keywords: Diazo-insertion Diastereomeric crystallization Spirotetracyclic compounds Sleep disorders Zwitterion
Research aiming to identify new and potent dual H1/5-HT2A antagonists for the treatment of sleep disorders,1 led to new chemical entities characterized by a spirotetracyclic scaffold substituted with a cyclic aminoacid moiety. Compound 1 (Fig. 1) emerged as one of our most interesting and promising molecules, endowed as it was with appropriate, well balanced, dual H1/5-HT2A potencies and a good developability profile.2 Substance 1 and congeners appeared to be available by reductive amination of ketone (S)-9, which, therefore constitutes the major subgoal en route to the target molecules. It soon became apparent that the preparation of this spiroketone was an interesting chemical challenge, especially because no literature precedent could be found for the synthesis of such a novel structure. The first original route used to prepare 1 entailed the elaboration of 5,11-dihydro-10H-dibenzo[a,d]cyclohepten-10-one to (S)9 (Fig. 1).2 While this approach enabled the conduct of all planned structure–activity relationship (SAR) studies, it provided only limited amounts of material. The major problems with this route were that it required two preparative chiral HPLC separations, and that it involved steps that were difficult to scale-up. Significantly larger quantities of material were required to support the full preclinical characterization of the most promising compounds, including 1. Accordingly, a completely new enantioselective avenue to (S)-9 had to be developed. In this Letter, we describe a practical route
to 1 that is suitable for large scale production and that requires no preparative chiral HPLC. This new route relies on a key CH insertion reaction of a carbenoid. As detailed in Scheme 1, the precursor of the requisite carbenoid was diazoacetoacetate 7. The choice of 7, instead of a more straightforward diazoketone,3 was dictated by the fact that the Buchner reaction (namely the cyclopropanation of the aromatic ring) of the carbenoid derived from the latter would have been highly competitive with the desired CH insertion.4 The synthesis thus started from 10-bromo-5H-dibenzo[a,d]cyclohepten-5-one 2 (Scheme 1), easily obtained by bromination/dehydro-bromination of dibenzosuberenone.5 The bromo derivative 2 was then reacted with methylacrylate under Heck conditions to give compound 3 in very high yield and exclusively as the E-geometric isomer.
O
HO
O
N
O
1
⇑ Corresponding author. E-mail address:
[email protected] (M. Gianotti). 0040-4039/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.tetlet.2010.11.039
(S )-9
Figure 1. Key intermediate spiro tetracyclic (1S)-5’,11’-dihydro-3H-spiro[cyclopentane-1,10’-dibenzo[a,d]cyclohepten]-3-one (S)-9 in the previous retrosynthetic approach.
330
M. Gianotti et al. / Tetrahedron Letters 52 (2011) 329–331 OMe Br
OH
OMe O
O
O i
iii
ii
4
3 O
2 O
5 iv
OEt O
O
O
O N2
OEt
OEt
O
O O vii
v
vi
9
6
7
8
Scheme 1. Synthesis of the key Intermediate 9. Reagents and conditions: (i) DMF, methyl acrylate, TEA, PdCl2(PPh3)2, 75 °C; (ii) acetic acid, Pd/C, H2 10 atm, 90 °C; (iii) H2O/ THF, LiOH; (iv) MeCN, TEA, CDI, MgCl2, potassium monoethyl malonate; (v) p-acetamidobenzenesulfonyl azide, TEA; (vi) DCM, Rh2(OAc)4; (vii) toluene/H2O, DMAP, 92 °C.
All attempts to promote an enantioselective hydrogenation of 3 that would introduce the requisite chirality at the benzylic carbon met with failure. Even reduction in the presence of cinchonidine (CD)6 was unsuccessful. Therefore, compound 3 was hydrogenated over Pd(C) under more traditional conditions, resulting in saturation of the double bonds and hydrogenolysis of the aromatic keto function. This afforded ester 4, the hydrolysis of which to the corresponding acid 5 was straightforward. The latter, in turn, was converted into b-keto ester 6 under Masamune conditions,7 and then advanced to the diazo derivative 78 by reaction with p-acetamidobenzenesulfonyl azide.9 Exposure of diazoacetoacetate 7 to a catalytic amount of Rh2(OAc)4 resulted in a clean and smooth cyclization to the desired spiro intermediate 8, obtained as the exclusive product. A final decarbethoxylation gave the desired key intermediate, (±)-9. It is well established that carbenoid CH insertion reactions of the type leading to compound 8 proceed with retention of configuration.10 Consequently, the use of enantioenriched acid 5 in the above sequence materialized as a good strategy for the production of nonracemic 9. This prompted us to study the resolution of acid (±)-5 by crystallization of diastereomeric salts obtained by reaction with chiral, enantiopure amines. Table 1 reports a selection of amines and solvents used in representative crystallization experiments. All the precipitates were filtered and analyzed by chiral HPLC to evaluate the enantiomeric excess of the acid component. Best results were obtained with (2-naphthyl)ethylamine in both THF and dioxane. The desired (S) -5 was readily obtained by using (R)-(+)-1-(2-naphthyl)ethylamine and dioxane was the solvent of choice for the recrystallization, because of the better recovery obtained compared to THF. The requisite enantiomeric excess (98% ee) for progressing with the chemical route required up to five crystallizations.11 The sequence of Figure 2 permitted the efficient conversion of enantioenriched (S)-5 into (S)-7, which in turn underwent Rhcatalyzed cyclization to (S)-9. In accord with the literature,10 the Table 1 Resolution of 3-(10,11-dihydro-5H-dibenzo[a,d]cyclohepten-10-yl)propanoic acid (5) with chiral amines
OEt O
OH O
O
N2
O
S
(S )-9
(S )-7
(S )-5
Figure 2. Stereo-conservative diazo-insertion.
cyclization step via carbenoid CH insertion had occurred with complete retention of configuration. The final step on the way to 1 was the reductive amination of (S)-9 with methyl 3-azetidinecarboxylate, a reaction leading to the formation of the second stereogenic center of the molecule (Scheme 2).12 This transformation was carried out using the customary NaBH(OAc)3 as the reductant, resulting in formation of a favorable 80/20 diastereomeric ratio of products 10. Separation of the two diastereoisomers by conventional chromatography proved to be extremely tedious. In order to avoid resorting to the use of preparative HPLC, the separation of the two diastereomers of 10 by crystallization of salts prepared using chiral acids was investigated. As shown, in Table 2, best results were obtained using D-malic acid and acetone as solvent. Four consecutive crystallizations were required to obtain a single diastereoisomer.13 The last step, the hydrolysis to give the desired final compound 1, proceeded smoothly under standard conditions using potassium hydroxide in methanol/water (3:1). In conclusion, a new synthesis of lead compound 1 that avoids two preparative chiral HPLC separations was achieved. Key to the success of the new route were the resolution of (±)-5 with (R)(+)-1-(2-naphtyl)ethylamine, leading to an efficient multigram
O
O
O N
Chiral amines S-()-1-(2Naphtyl)ethylamine S-()-1-Aminotetralin S-()-1-Phenylpropylamine Cinchonidin R-(+)-1-(2-Naphtyl)ethylamine
Ethyl acetate 4% ee 4% ee Sol. 0% ee —
MeCN 2% ee 2% ee 10% ee 0% ee —
THF 55% ee Sol. Sol. Sol. —
Sol. solution (no precipitation was observed), — not performed.
Dioxane 47% ee — — — 46% ee
ii
i
1
S (S )-9 10
Scheme 2. Synthesis of the final compound 1: (i) methyl 3-azetidinecarboxylate, DCE, NaBH(OAc)3, rt, (ii) KOH, MeOH.
M. Gianotti et al. / Tetrahedron Letters 52 (2011) 329–331 Table 2 Crystallization of 10 with chiral acids 4. Chiral acids
Ethyl acetate
IPA
THF
Acetone
Tartaric acid (+) Tartaric acid ()
53% de 54% de 54% de
60% de 58% de 78% de
Sol. Sol. 84% de
Sol. Sol. 84% de
Sol.
35% de
Sol.
Sol.
Sol.
56% de
Sol.
Sol.
D-Malic
acid
L-Malic
acid Dibenzoyl-D-tartaric acid
5. 6. 7. 8.
Sol. solution (no precipitation was observed).
procedure to obtain enantiopure acid intermediate (S)-5, and the separation of the diastereomers of compound 10 by crystallization of the corresponding salts of D-malic acid. Besides enabling the conduct of the full preclinical characterization, the new synthesis demonstrates the yet undocumented resolution of substances of type 5 and the enantioselective assembly of spiroketones such as 9. These advances are likely to be of interest to medicinal and synthetic organic chemists alike.
9. 10. 11.
Supplementary data Supplementary data (all the experimental procedures, characterization, for compounds 3–9) associated with this article can be found, in the online version, at doi:10.1016/j.tetlet.2010.11.039. References and notes 12. 1. (a) Renger, J. J. Curr. Top. Med. Chem. 2008, 8, 937–953; (b) Gianotti, M.; Corti, C.; Delle Fratte, S.; Di Fabio, R.; Leslie, C. P.; Pavone, F.; Piccoli, L.; Stasi, L.; Wigglesworth, M. J. Bioorg. Med. Chem. Lett. 2010, 20, 5069–5073. 2. Gianotti, M.; Botta, M.; Brough, S.; Carletti, R.; Castiglioni, E.; Corti, C.; Dal-Cin, M.; Delle Fratte, S.; Korajac, D.; Lovric, M.; Merlo, G.; Mesic, M.; Pavone, F.; Piccoli, L.; Rast, S.; Roscic, M.; Sava, A.; Smehil, M.; Stasi, L.; Togninelli, A.; Wigglesworth, M. J. J. Med. Chem. 2010, 53, 7778–7795. 3. + N N Cl
5
O
O
i
ii
iii
9 + by-products
13.
331
(i) DCM, oxalyl chloride, DMF; (ii) (a) Et2O, TMSCHN2; (b) KF/MeOH; (iii) DCM, Rh2(OAc)4. (a) Maguire, A. R.; Buckley, N. R.; O’Leary, P.; Ferguson, G. J. Chem. Soc., Perkin Trans. 1 1998, 24, 4077–4091; (b) Maguire, A. R.; O’Leary, P.; Harrington, F.; Lawrence, S. E.; Blake, A. J. J. Org. Chem. 2001, 66, 7166–7177. Taljaard, B.; Taljaard, J. H.; Imrie, C.; Caira, M. R. Eur. J. Org. Chem. 2005, 12, 2607–2619. Nitta, Y.; Watanabe, J.; Okuyama, T.; Sugimura, T. J. Catal. 2005, 236, 164–167. (a) Brooks, D. W.; Lu, L. D.-L.; Masamune, S. Angew. Chem., Int. Ed. Engl. 1979, 18, 72–74; (b) Mansour, T. S.; Evans, C. A. Synth. Commun. 1990, 20, 773–781. The a-diazo-b-ketoester intermediate 7 demonstrated to be thermically stable up to 90 °C by means of DSC (Differential Scanning Calorimetry) analysis. Bollinger, F. W.; Tuma, L. D. Synlett 1996, 407–413. Muller, P.; Bolea, C. Helv. Chim. Acta 2002, 85, 483–494. Resolution of 3-(10,11-dihydro-5H-dibenzo[a,d]cyclohepten-10-yl)propanoic acid (5) by crystallization of the (R)-(+)-1-(2-naphtyl)ethylamine salt. 3(10,11-Dihydro-5H-dibenzo[a,d] cyclohepten-10-yl)propanoic acid (5, 1.22 Kg, 4.56 mol) was dissolved in dioxane (30 L) and (R)-(+)-1-(2naphthyl)ethylamine (781 g, 4.56 mol) was added portionwise. Soon a white precipitate appeared. The mixture was stirred for 2 h and then the precipitate, a white solid, was filtered and oven-dried at 55 °C under vacuum overnight. This afforded 1240 g of a white solid, which was suspended in 10 L of dioxane and heated to 95 °C in order to achieve complete dissolution. The solution was left to cool to room temperature and stirred for 1 h. The white precipitate was filtered and dried as described before to provide 994 g of a white solid, which was suspended in 10 L of dioxane and heated to 95 °C in order to achieve complete dissolution. The solution was cooled to about 40 °C and the white precipitate that formed was filtered and dried as detailed above to furnish 760 g of a white solid. Two additional recrystallizations (five crystallizations total) gave 580 g of (S)-3-(10,11-dihydro-5H-dibenzo[a,d]cyclohepten-10yl)propanoic acid, [(1R)-1-(2-naphthalenyl)ethyl]amine salt, as a white solid. A chiral HPLC assay indicated that the optical purity of the acid corresponded to 98% ee. Preparation of free acid (S)-5. The above salt (667.8 g, 1.53 mol) was dissolved in ethyl acetate (7 L) and treated with aqueous 1 M HCl solution (2 7 L and 1 4 L). Separation of the aqueous layers and evaporation of the organic phase afforded 337 g (83% yield) of (S)-3-((10S)-10,11-dihydro-5H-dibenzo[a,d] cyclohepten-10-yl)propanoic acid, (S)-5. The absolute configuration of the amino ester 10 and of the final amino acid 1 was not determined. During the exploration of the SAR, few attempts have been done to determine stereochemistry on the final amino acids or the precursor amino esters and we were successful only in one example, by VCD analysis, whose results are still under consideration. Separation of the diastereomers of 10 by crystallization of its salt with D-malic acid. Methyl 1-(50 ,110 -dihydrospiro [cyclopentane-1,10’-dibenzo[a,d]cyclohepten]-3-yl)-3-azetidine-carboxylate (10, 1.2 g, 3.32 mmol) was dissolved in acetone (12 mL), then D-malic acid (1 equiv) was added. The resulting light yellow solution was left at rt. overnight, whereupon a white precipitate formed. This solid was filtered and triturated four times with acetone (12– 15 mL) to give a diastereomerically pure malate salt (840 mg, white solid) of 10. This solid was then treated with aqueous saturated NaHCO3 solution and the aqueous phase was extracted with DCM to give pure 10 (592 mg, 49% yield).
