Abstract
Highly torquoselective electrocyclizations of chiral 1-azatrienes are described. These 1-azatrienes contain an allylic stereocenter that is substituted with a silyl group and are derived in situ from condensation of γ-silyl-substituted enals with vinylogous amides. The ensuing stereoselective ring-closures are part of a tandem sequence that constitutes an aza-[3 + 3] annulation method for constructing 1,2-dihydropyridines. Several mechanisms for the formal 1,7-hydrogen shift of these 1-azatrienes were evaluated computationally.
Electrocyclizations represent an important pericyclic process in organic synthesis. Our aza-[3 + 3] annulation1-3 method involving chiral enals 1 and vinylogous amides 2 is a powerful strategy for total syntheses of alkaloids4 and a unique platform for studying the torquoselectivity of electrocyclizations of 1-azatreienes 35 (Scheme 1). Despite its significance in constructing chiral 1,2-dihydropyridines, efforts to develop and understand torquoselective ring-closures of 1-azatrienes 3 have lagged behind with the sole exceptions of Tanaka and Katsumura’s elegant work.6 Although we have developed highly torquoselective electrocyclizations of a chiral auxiliary substituted 1-azatrienes,7 a more general and practical approach employing chiral enals has yielded diastereoselectivity of 83:17 at best (see aza-electrocyclization of 3a in Scheme 2).8.9 Recently, our collaborative efforts to understand the origins of the stereoselectivities of a number of pericyclic reactions10 have led us to model these stereoselective ring closures computationally. A complete stereochemical model for these electrocyclic reactions is still being developed. In the course of our studies, we predicted that the stereochemical outcomes of these electrocyclizations depend on the electronic nature of the allylic substituent X. As shown in Scheme 2, if X is a σ donor such as SiR3 instead of a σ acceptor such as OAc, a reversal of stereoselectivity is predicted (4 versus 5). We have now shown that such a reversal occurs and that the electrocyclizations of these silyl-substituted 1-azatrienes are highly torquoselective.
Scheme 1. Torquoselective Electrocyclizations of 1-Azatrienes in Aza-[3 + 3] Annulations.
Scheme 2. A Prediction of Reversal of the Torquoselectivity.
We commenced our investigation by examining aza-[3+3] annulations of vinylogous amides 6 and 7 with γ-silyl-substituted enal 811,12, and quickly found that the respective desired aza-annulation products 11/11′ and 13/13′ were minor products (Scheme 3). Major products in these reactions were vinyl silanes 12 and 14 from 6 and 7, respectively. These isomeric vinyl silanes could be formed by a (formal) 1,7-H shift of 1-azatrienes 9 or 10. Although the competition of a 1,7-H shift with the desired annulation pathway has been documented,13 the isomerizations responsible for the formation of 12 and 14 have never been observed before. The E-configurations of vinyl silanes 12 and 14 were assigned using NOE experiments.
Scheme 3. An Unexpected Competing 1,7-Hydrogen Shift.
Annulations using 6-membered ring vinylogous amides were more successful. As shown in Scheme 4, although the reaction of vinylogous amide 15a still yielded the 1,7-H shift product (16a) as the major product, respective 1-azatrienes from 6-membered ring vinylogous amides 15b and 15c predominantly underwent ring-closure in high yields and diastereoselectivity. This is also true in cases of electrocyclizations that led to 18 and 20 with the respective vinyl silanes byproducts 19 and 21 being isolated only in small amounts.
Scheme 4.
Aza-[3+3] Annulations Using 6-Membered Vinylogous Amides as Annulation Partners
Using the single crystal X-ray structure of 16b, we were able to unambiguously assigned the stereochemistry of 16b and confirm the prediction of a complete reversal of selectivity for electrocyclizations of these silyl-substituted 1-azatrienes. The attempted aza-annulations of 1-azatrienes bearing large N-substituent (such as the N-CHPh2 group of 15a) would still yield products of 1,7-H shift. This is presumably due to enhanced steric repulsion between the larger N-substituent and the TBDPS group at the electrocyclization transition state. It is noteworthy that in direct contrast, aza-annulations of 15a with non-silylated chiral enals were feasible and most diastereoselective as demonstrated by 22.8a
Table 1 illustrates the generality of this stereoselective aza annulation; an array of different γ–silyl-substituted enals 25a-h, including one substituted with a TBS group, were successfully used as annulation partners. In all cases, the selectivity is very high while the competing 1,7-H shift is by and large mitigated. It is noteworthy that this is the first time a very high level of diastereoselectivity could be achieved in aza-[3 + 3] annulations using acyclic chiral enals.
Table 1.
| entry | chiral enals | electrocyclization products | 1,7-H shift products |
|---|---|---|---|
| 1 |
|
|
|
| 2 | 25b: R = i-Bu |
|
|
| 3 | 25c: R = n-hex |
|
|
| 4 | 25d: R = Ph(CH2)3 |
|
|
| 5 | 25e: R = allyl |
|
|
| 6 | 25f: R = Bn |
|
|
| 7 |
|
|
|
| 8 |
|
|
|
All reactions were carried with vinylogous amide 16c using piperidine and Ac2O, and reactions were heated at 130 °C for 24 h.
All are isolated yields and dr ratios are determined using 1H NMR analysis of the crude reaction mixture.
To better understand why 1-azatrienes annulated with 5-membered rings (9 and 10), undergo competitive formal 1,7-hydrogen shifts rather than the desired aza-electrocyclizations, we modeled the reaction of truncated 1-azatrienes 28 and 30 (see Figure 2) computationally.15 In Scheme 5, a summary of four possible mechanisms by which isomerization may occur is shown. All pathways assume the intermediacy of 1-azatriene I, and pathways 1, 2, and 4 feature key steps that are concerted in nature. Consequently, in addition to modeling the electrocyclizations of 28 and 30, we have also modeled steps of these three pathways. The intermediacy of 1-azatriene I in pathway 3, which involves base-mediated proton transfer, has not been modeled; however, such a mechanism is a plausible alternative.
Figure 2.
Energetics of the electrocyclic ring closures of model 1-azatrienes 28 and 30. Energies are Gibbs free energies in kcal mol−1 determined at the M062-X/def2-QZVPP//M06-2X/6-31+G(d,p) level.
Scheme 5. Potential Mechanism for the Competitive Isomerization.
The energetics of the electrocyclizations of 1-azatrienes 28 and 30 are shown in Figure 2. At 130 °C, the aza-electrocyclizations of 28 and 30 are facile reactions (ΔG‡ < 20 kcal mol−1) that under kinetic control stereoselectively yield dihydropiperidines 29a and 31a, respectively. Electrocyclization of 30 is, according to theory, only slightly more facile than that of 28; however, it is significantly more exergonic than the ring-closure of 28 (ca. 8 kcal mol−1).16
Based on computations, pathways 1 and 2 involving a direct 1,7-hydrogen shift or 1,5-hydrogen shift of 1-azatriene 28,6d respectively, are unlikely. These sigmatropic rearrangements feature ΔG‡ of at least 30 kcal mol−1. However, the free energy of activation for the 1,7-hydrogen shift involved in pathway 4 is 17 kcal mol−1. Thus, pathway 4 is a plausible mechanism, so long as the required isomerizations (presumably promoted by base) are facile processes.
Interestingly, the rate of 1,7-hydrogen shift is 100-fold slower (ΔΔG‡ = 2.7 kcal mol−1) than the ring closure of 1-azatriene 30. However, for 1-azatriene 28, these two processes are very similar in activation free energies (Figure 3). Theses difference in reactivity may be (partially) responsible for the distinct product outcomes observed for this pair of azatrienes.
Figure 3.
M06-2X/6-31+G(d,p) structures of lowest 1,7-hydrogen shift featured in pathway 4. Energies shown are M06-2X/def2-QZVPP//M06-2X/6-31+G(d,p).
The lowest energy transition structures of the 1,7-hydrogen shift of II derived from substrates 28 and 30 (TS29c and TS31c) are shown in Figure 2. TS31c is destabilized by A1,3 strain between N-methyl substituent and annulated cyclohexanone (see green lines in Figure 3). This destabilizing interaction is less severe in TS29c featuring the γ-lactone because this moiety, unlike the corresponding cyclohexanone in TS31c is planar.
We have described here a highly torquoselective electrocyclization of a series of novel chiral 1-azatrienes. These 1-azatrienes contain an allylic stereocenter substituted with a silyl group, and are generated in situ by condensing γ-silyl-substituted enals with vinylogous amides. Theoretical calculations have provided mechanistic insights into a previously unknown competing 1,7-hydrogen shift from the same 1-azatriene intermediate. Efforts to explore synthetic applications of this torquoselective electrocyclization are underway. Full details regarding the stereochemical model that rationalizes the observed torquoselectivities will be reported in due course.
Supplementary Material
Figure 1.
X-Ray Structure of 1,2-Dihydropyridine 16b.
ACKNOWLEDGMENT
Z-X.M. and A.P. have contributed equally to this work. We thank the NIH (GM-66055 to R.P.H) and NSF (CHE-1059084 and CHE-1361104 to K.N.H.). A.P. thanks the Chemistry-Biology Interface Training Program (NIH Grant T32 GM 008496) for its support and the University of California, Los Angeles (UCLA) for funding. UCLA’s Beowulf cluster, Hoffman2, and the Extreme Science and Engineering Discovery Environment’s (Grant TG CHE 040013N) Gordon and Trestles supercomputer at the San Diego Supercomputing Center were used to perform computations. We also thank Dr. Victor Young at University of Minnesota for X-ray structural analysis.
Footnotes
ASSOCIATED CONTENT
Experimental procedures as well as NMR spectra, characterizations, and X-ray structure file for all newly synthesized compounds, Cartesian coordinates for all computed structures, their electronic energies, zero point energies (ZPE), thermal, and free energy corrections for all QM-optimized structure, and the imaginary frequencies of transition structures are available free of charge via the Internet (http://pubs.acs.org).
REFERENCES
- (1).For reviews on hetero-[3 + 3] annulations, see: Harrity JPA, Provoost O. Org. Biomol. Chem. 2005;3:1349. doi: 10.1039/b502349c. Hsung RP, Kurdyumov AV, Sydorenko N. Eur. J. Org. Chem. 2005:23.
- (2).For a symposium-in-print devoted to aza-annulations, see: Harrity JPA. Tetrahedron. 2008;64 Symposium-in-Print No. 133.
- (3).Also see: Tang Y, Oppenheimer J, Song Z, You L, Zhang X, Hsung RP. Tetrahedron. 2006;62:10785. Laschat S, Dickner T. Synthesis. 2000:1781.
- (4).For reviews on applications of aza-[3 + 3] annulation in natural product synthesis, see: Deng J, Wang X-N, Hsung RP. In: Methods and Applications of Cycloaddition Reactions in Organic Syntheses. Nishiwaki N, editor. Wiley-VCH; 2014. Chapter 12. Buchanan GS, Feltenberger JB, Hsung RP. Curr. Org. Chem. 2010;7:363. doi: 10.2174/157017910791414490. Gademann K, Lawrence AK. Synthesis. 2008:331. Hsung RP, Cole KP. In: Strategies and Tactics in Organic Synthesis. Harmata M, editor. Vol. 4. Elsevier Science: Pergamon Press; Oxford, England: 2004. pp. 41–70.
- (5).For a review, see: Okamura WH, de Lera AR. In: Comprehensive Organic Synthesis. Trost BM, Fleming I, Paquette LA, editors. Vol. 5. Pergamon Press; New York: 1991. pp. 699–750.
- (6).For accounts on stereoselective ring-closure of 1-azatrienes, see: Tanaka K, Katsumura S. J. Am. Chem. Soc. 2002;124:9660. doi: 10.1021/ja026464+. Tanaka K, Mori H, Yamamoto M, Katsumura S. J. Org. Chem. 2001;66:3099. doi: 10.1021/jo005779+. Tanaka K, Kobayashi T, Mori H, Katsumura S. J. Org. Chem. 2004;69:5906. doi: 10.1021/jo049381f. Sakaguchi T, Okuno Y, Tsutsumi Y, Tsuchikawa H, Katsumura S. Org. Lett. 2011;13:4292. doi: 10.1021/ol2016302.
- (7).(a) Sklenicka HM, Hsung RP, Wei L-L, McLaughlin MJ, Gerasyuto AI, Degen SJ, Mulder JA. Org. Lett. 2000;2:1161. doi: 10.1021/ol000049+. [DOI] [PubMed] [Google Scholar]; (b) Sklenicka HM, Hsung RP, McLaughlin MJ, Wei L-L, Gerasyuto AI, Brennessel WW. J. Am. Chem. Soc. 2002;124:10435. doi: 10.1021/ja020698b. [DOI] [PubMed] [Google Scholar]
- (8).(a) Sydorenko N, Hsung RP, Vera EL. Org. Lett. 2006;8:2611. doi: 10.1021/ol060932t. [DOI] [PubMed] [Google Scholar]; (b) Ghosh SK, Buchanan GS, Long QA, Wei Y, Al-Rashid ZF, Sklenicka HM, Hsung RP. Tetrahedron. 2008;63:883. doi: 10.1016/j.tet.2007.09.089. [DOI] [PMC free article] [PubMed] [Google Scholar]
- (9).For the use of chiral cycloalkylidene aldehydes, see: McLaughlin MJ, Hsung RP, Cole KC, Hahn JM, Wang J. Org. Lett. 2002;4:2017. doi: 10.1021/ol020052o.
- (10).(a) Wang X-N, Krenske EH, Johnston RC, Houk KN, Hsung RP. J. Am. Chem. Soc. 2014;136:9802. doi: 10.1021/ja502252t. [DOI] [PMC free article] [PubMed] [Google Scholar]; (b) Du Y, Krenske EH, Antoline JE, Lohse AG, Houk KN, Hsung RP. J. Org. Chem. 2013;78:1753. doi: 10.1021/jo3011792. [DOI] [PMC free article] [PubMed] [Google Scholar]; (c) Krenske EK, He S-Z, Huang J, Du Y, Houk KN, Hsung RP. J. Am. Chem. Soc. 2013;135:5242. doi: 10.1021/ja312459b. [DOI] [PMC free article] [PubMed] [Google Scholar]; (d) Antoline JE, Krenske EH, Lohse AG, Houk KN, Hsung RP. J. Am. Chem. Soc. 2011;133:14443. doi: 10.1021/ja205700p. [DOI] [PMC free article] [PubMed] [Google Scholar]; (e) Krenske EH, Houk KN, Lohse AG, Antoline JE, Hsung RP. Chem. Sci. 2010;1:387. doi: 10.1039/C0SC00280A. [DOI] [PMC free article] [PubMed] [Google Scholar]; (f) Lohse AG, Krenske EH, Antoline JE, Houk KN, Hsung RP. Org. Lett. 2010;12:5506. doi: 10.1021/ol1023745. [DOI] [PMC free article] [PubMed] [Google Scholar]
- (11).See Supporting Information.
-
(12).We used TBDPS-substituted enals, because we failed in our initial attempts of using TMS-substituted enals. Peterson-like elimination of TMS group was observed instead of the desired ring closure. We attempted reactions of enals substituted with silyl groups of intermediate sizes (Ph3Si and Ph2MeSi). However, synthesis of these enals proved difficult
-
(13).A 1,7-H shift of 1-azatriene i was also observed, giving 2-azatriene ii. This shift is quite distinct from the rearrangement of 9/10 to 12/14. See:
Sydorenko N, Hsung RP, Darwish OS, Hahn JM, Liu J. J. Org. Chem. 2004;69:6732. doi: 10.1021/jo049108d.
- (14).Azatrienes 28 and 30 (see Figure 2) are truncated model substrates featuring a TMS substituent instead of the bulky TBDPS group.
- (15).All computations were performed using Gaussian09. The calculations reported herein were performed using the M06-2X/def2-QZVPP//M06-2X/6-31+G(d,p) model chemistry and all energies reported are Gibbs free energies. Additional details and relevant references can be found in the Supporting Information.
- (16).Further discussion of the thermodynamics of the ring closures of 28 and 30 has been related to the Supporting Information.
- (17).See Supporting Information for details regarding pathways 1 and 2.
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.