Skip to main content
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2016 Oct 5.
Published in final edited form as: Angew Chem Int Ed Engl. 2015 Aug 19;54(41):12097–12101. doi: 10.1002/anie.201504723

Oxidative Allene Amination for the Synthesis of Azetidin-3-ones

Eileen G Burke 1, Jennifer M Schomaker 1,
PMCID: PMC4701577  NIHMSID: NIHMS741675  PMID: 26387687

Abstract

Regioselectivity in the aziridination of silyl-substituted homoallenic sulfamates is readily diverted to the distal double bond of the allene to yield endocyclic bicyclic methyleneaziridines with excellent stereocontrol. Subsequent reaction with electrophilic oxygen sources initiates facile rearrangement to densely functionalized, fused azetidin-3-ones in excellent dr, effectively transferring the axial chirality of the allene to central chirality in the products. The steric nature of the silyl group dictates which of the two rings of the fused azetidin-3-one will undergo further functionalization, providing an additional element of diversity for the preparation of enantioenriched azetidine scaffolds with potential biological activity.

Keywords: allenes, azetidin-3-ones, aziridination, rearrangement, heterocycles


Densely functionalized azetidine motifs exhibit a broad range of biological activities, including antibacterial activity, inhibitory activity against epidermal growth factor receptor kinases, antagonists of the P2Y12 receptor, components of dopamine antagonists, ATP-ase activators and inhibitors of various glucosidases for the treatment of diabetes (Figure 1).[1] While the syntheses of related azetidin-2-ones (β-lactams) are well-documented,[2] approaches to the analogous azetidin-3-ones are much less common.[3] Early strategies reported by the De Kimpe group (Scheme 1) required multiple steps to convert methyl 4-chloro-3-oxobutanoate or butane-2,3-dione to racemic 2,4-disubstituted azetidin-3-ones.[4,5] Cu- and Rh-catalyzed carbenoid insertions into the N-H bonds of α,α'-dialkyl-α-diazoketones offer a more streamlined approach to cis-2,4-disubstituted azetidin-3-ones from amino acid precursors, but the scope of the products is limited.[6,7] Complex 3-hydroxy-4-(hydroxymethyl)-3-methylazetidine-2-carboxylic acids can be prepared through the intermediacy of stable bicyclic azetidin-3-ones, but this approach also requires several steps from the pyranoside substrates.[8] In this communication, we report a simple protocol that permits rapid access to fully substituted azetidin-3-ols in good yields and excellent diastereoselectivities from simple homoallenic sulfamate precursors.

Figure 1.

Figure 1

Bioactive azetidines.

Scheme 1.

Scheme 1

Other approaches to azetidin-3-ones.

Our group has been engaged in developing an array of oxidative allene amination strategies that provide both heteroatom and stereochemical diversity for the synthesis of complex amine stereotriads that contain three contiguous and heteroatom-bearing chiral carbons (Scheme 2).[9] A key component of our strategy is a highly chemo-, regio- and stereoselective allene aziridination of the proximal double bond to yield the versatile exocyclic methylene aziridine intermediate A, which is typically opened in situ with a nucleophile. We hypothesized that if aziridination could be directed to the distal double bond of the allene, the resultant endocyclic bicyclic methyleneaziridine B could function as a useful scaffold for constructing heterocycles with potential biological activity, including the azetidin-3-ones (Scheme 2, bottom) described in this communication.[9d,10]

Scheme 2.

Scheme 2

Divergent regiocontrol in allene aziridination.

Our first challenge was to ascertain how the identity and the specific substitution pattern of the groups on the allene could be employed to reliably divert aziridination to the distal double bond. Placing a bulky t-butyl group at C1 of the homoallenic sulfamate 1 (Scheme 3) did not completely prevent aziridination yielding a 2:1 mixture of exo- and endocyclic bicyclic methyleneaziridines 2a and 2b.[11] Manipulation of electronic effects proved to be a better strategy, as placement of a t-butyldimethylsilyl (TBS) group at C1 of 3 resulted in exclusive aziridination of the distal double bond to 4 using a variety of dinuclear Rh(II) catalysts. The superior regioselectivity exhibited by silylated allene 3 was attributed to stabilizing electronic interactions between the C-Si bond and the π-bonds of the allene (Scheme 3, bottom). As the aziridine begins to form a new C-N bond at the distal double bond of the allene, the C-Si bond becomes co-planar with the π-orbitals, allowing hyperconjugation to stabilize the developing positive charge at the β carbon. However, when aziridination occurs at the proximal double bond, the π-bonds are parallel to the C—Si bond, negating the potential for this stabilization.[12]

Scheme 3.

Scheme 3

Initial studies to control regioselectivity.

In contrast to exocyclic sulfamate-containing bicyclic methyleneaziridines employed in previous studies, endocyclic methyleneaziridine 4 (Table 1) exhibited good stability and was readily purified by column chromatography. We expected this stability would enable unprecedented functionalization of the alkene prior to competing aziridine ring-opening, followed by rearrangement to the desired azetidin-3-one scaffold. Indeed, treatment of 4 with a series of electrophilic oxygen sources revealed that mCBPA provided 6 in excellent yield when the reaction was run under dilute conditions (Table 1, entry 5). Other electrophilic oxygen reagents, including magnesium monoperoxyphthalate hexa-hydrate (MMPP, entry 2) or dimethyldioxirane (DMDO, entry 1), showed low reactivity. The highly reactive 1,4-oxaza[2.2]-spiropentane 5 was presumed to be a reactive intermediate, but was never isolated due to spontaneous rearrangement to a single diastereomer of the product 6, as determined by 1H NMR.[13] The relative stereochemistry between the silyl group and the substituent at C1 of the azetidin-3-one 6 was determined to be anti through nOe studies (Table 1, bottom). Further corroboration of this initial assignment was provided through more definitive nOe studies on a related compound (see Scheme 5).

Table 1.

Rearrangement of endocyclic bicyclic methyleneaziridines to fused azetidin-3-ones.

graphic file with name nihms741675t1.jpg

entry conditions yield dr
1 2 equiv DMDO, 0.1 M CH2Cl2 27% ---
2 1 equiv MMPP in 4:1 H2O:MeOH, 0.1 M 0% ---
3 1 equiv Davis oxaziridine, 0.1 M CH2Cl2 0% ---
4 3 equiv mCPBA, 0.3 M CH2Cl2 5% ---
5 3 equiv mCPBA, 0.1 M CH2Cl2 75% > 19:1
6 3 equiv mCPBA, 0.05 M CH2Cl2 84% > 19:1

graphic file with name nihms741675t2.jpg

Scheme 5.

Scheme 5

Transformations of H-substituted azetidin-3-one 10.

A selected scope of oxidative allene amination/rearrangement is illustrated in Table 2. The identity of the silyl group could be changed to either a TMS or TES group (entries 3, 4) without affecting regioselectivity of the aziridination or the subsequent transformation to the azetidin-3-ones 10 and 13. In addition, treatment of the TMS-substituted azetidin-3-one resulting from methyleneaziridine 9 with silica gel promoted ready protodesilylation to yield 10. These results provide further support for our argument that electronic factors, rather than sterics, dictate regioselectivity (Scheme 2), as the TMS group is much smaller than the tBu group (A values 2.5 and 4.9, respectively).[14] Groups containing longer alkyl chains, as well as branching, were tolerated at the terminus of the allene (entries 5, 6) and gave products in good yields and excellent dr. Installing a Me group at the allenic carbon of the tether between the allene and the sulfamate decreased the yields of both steps (entry 7), presumably due to steric interactions. However, 52% of 21 was isolated which translates into a 77% yield of 22 based on recovered 21. The 1:1 dr of substrate 20 translated into a 1:1 mixture of the diastereomeric products of 22, where the dr of each isomer was dr > 19:1. The highly stereocontrolled nature of the rearrangement for each diastereomer provides additional flexibility in the substituted azetidin-3-ones that can be formed using this chemistry. Moving the alkyl substitution to the carbon adjacent to the oxygen of the sulfamate (entries 8, 9) did not overly impact the aziridination step, but did result in decreased yields in the rearrangement step. In the case of transforming 26 to 28, 4% of the endocyclic methylene aziridine 27 was recovered after the epoxidation. A protected oxygen at the allenic position of 29 was also tolerated, providing a convenient synthetic handle for further reactions (entry 10). The lower yield of 31 likely results from the steric bulk of the protecting group; in this case, 16% of unreacted 29 was also recovered from the reaction mixture. Despite our best efforts to install aromatic functionality at C1 or C3 of the allene, the azetidin-3-ones could only be obtained in low yields, as ring-opening of the bicyclic methyleneaziridine at the activated benzylic carbon was quite facile. However, aromatic groups could be introduced into the products in other ways (see Schemes 4 and 5).

Table 2.

Substrate scope of the azetidin-3-one formation.

graphic file with name nihms741675t3.jpg
a

1 mol% Rh2(OAc)4 was used in entry 1.

Scheme 4.

Scheme 4

Selected transformations of TBS-substituted azetidin-3-one 6.

A nice feature of our chemistry is the ability to achieve divergent functionalization of the azetidin-3-one products through judicious choice of the silyl group.[3] For example, when the bulky TBS group was employed to prepare the azetidin-3-one 6, the carbonyl group of the product was sufficiently shielded from reaction with nucleophiles that would typically engage such functionalities (Scheme 4). Small hydride nucleophiles were able to react, as treatment of 6 with excess NaBH4 in EtOH gave a 49% yield of 32 resulting from reduction and presumed 1,2-Brook rearrangement; the remainder of the mass balance consisted of unreacted starting material.[15] However, the TBS group was immune to removal by standard desilylation protocols, including TBAF, HF-pyridine and tris(dimethylamino)sulfonium difluorotrimethylsilicate (TASF). We were pleased to find that this unexpected stability enabled strong, bulky reducing agents, including L-selectride, to chemoselectivity open the sulfamate group in lieu of carbonyl reduction, providing products such as 33 in good yield with no erosion in the diastereoselectivity. In a similar vein, hydrolysis of the sulfamate protecting group gave the 1,3-aminoalcohol 34 in excellent yield. Finally, treatment of 6 with concentrated HCl in THF gave the ring-opened product 35 in excellent yield, containing a useful leaving group positioned for further reaction.

In contrast to the TBS-substituted azetidin-3-ones, utilizing the less bulky TMS group in the reaction sequence permitted orthogonal reactivity of the resulting heterocycle (Scheme 5). This ability to remove the TMS group unlocks the reactivity at the carbonyl group of the azetidin-3-one scaffold. For example, subsequent treatment of 10 with L-selectride enables a highly diastereoselective reduction to yield the 2° alcohol 36 in excellent dr. The relative stereochemistry amongst the three stereogenic carbons of the azetidine was established as all syn through definitive nOe studies (Scheme 5, see SI for additional details). Another advantageous feature of 10 is that Grignard reagents, as illustrated by vinyl MgBr in Scheme 4, react with the carbonyl in a highly diastereoselective manner to yield a 3° alcohol 37 in a dr >19:1. Aryl groups could be readily introduced into the azetidine products, as illustrated by the conversion of 10 to 39 in good yield and excellent dr. Finally, treatment of 10 with a stabilized phosphonium ylide gave 38 in >9:1 E:Z. Further diastereoselective transformations of the products illustrated in both Schemes 4 and 5 can be readily envisaged to provide remarkable flexibility in using azetidin-3-one scaffolds for the rapid extension of chemical space.

In our previous studies involving Rh-catalyzed aziridination of the proximal double bond of homoallenic sulfamates, the transfer of axial to central chirality occurred with excellent fidelity to give enantioenriched amines.[9] To ensure that this same transfer of axial chirality was operative in the synthesis of these heterocycles, the synthesis of the enantioenriched TBS-azetidinone (Scheme 6) was accomplished starting from (R)-(+)-3-butyn-2-ol, employing the same methodology used to synthesize racemic analogs. The enantioenriched azetidinone was derivatized by opening of the sulfamate ring with thiophenol to provide a chromophore for chiral HPLC analysis. The product 42 was isolated with 96% ee. The low yield was due to a side reaction initiated by deprotonation of 41, followed by electrocyclic ring-opening and hydrolysis to 43 (see SI for details). Nonetheless, this experiment clearly showed effective transfer of the axial chirality of the allene to the azetidin-3-one.

Scheme 6.

Scheme 6

Axial-to-central chirality transfer.

In conclusion, we have described a facile synthesis of azetidin-3-ones and azetidines that proceeds through a highly regioselective aziridination of the distal bond of homoallenic sulfamates. One key feature of our new methodology is the highly diastereocontrolled epoxidation of an endocyclic bicyclic methyleneaziridine to yield reactive oxazaspiropentanes that rapidly rearrange to the title azetidin-3-ones in excellent dr. Another key feature of this chemistry is the flexibility of subsequent transformations that yield densely functionalized azetidines with potentially useful biological activity. Future studies are directed towards increasing the reaction scope, exploring other electrophilic reagents that react with the endocyclic methyleneaziridines and expanding the chemical space that can be accessed via flexible manipulations of various azetidin-3-ones.

Supplementary Material

Supporting Information

Acknowledgements

The authors thank Dr. Charles Fry and Nicholas Dolan for assistance with NMR studies. Nels Gerstner is thanked for his efforts on aryl-substituted allene substrates. JMS thanks the NSF-CAREER Award 1254397 and the NIH 1R01GM111412-01 for funding of this research. JMS is an Alfred P. Sloan fellow. The NMR facilities at UW-Madison are funded by the NSF (CHE-9208463, CHE-9629688) and NIH (RR08389-01). The National Magnetic Resonance Facility at Madison is supported by the NIH (P41GM103399, S10RR08438, S10RR029220) and the NSF (BIR-0214394).

REFERENCES

  • 1.For selected references on bioactive azetidin-3-ones and related compounds, see: Wellington K, Scott LJ. Drugs. 2003;63:2613. doi: 10.2165/00003495-200363230-00004. Leangh-Snow GM, Araujo N, Jenkinson SF, Rutherford C, Nakagawa S, Kato A, Yu C-Y, Weymouth-Wilson AC, Fleet GWJ. Org. Lett. 2011;13:5834. doi: 10.1021/ol2024482. Frigoli J, Torrens A, Castrillo JA, Mas J, Vañó D, Berrocal JM, Calvet C, Salgado L, Redondo J, Garcia-Granda S, Valentí E, Quintana JR. J. Med. Chem. 1994;37:4195. doi: 10.1021/jm00050a016. Kobayashi J, Cheng J, Ishibashi M, Wälchli MR, Yamamura S, Ohizumi Y. J. Chem. Soc., Perk. Trans. I. 1991:1135. Martínez RF, Fleet GWJ. Tetrahedron: Asymm. 2014;25:373. and references therein.
  • 2.For a review of recent approaches to the syntheses of β-lactams, see: Pitts CR, Lectka T. Chem. Rev. 2014;16:7930. doi: 10.1021/cr4005549.
  • 3.Dejaegher Y, Kuz’menok NM, Zvonok AM, De Kimpe N. Chem. Rev. 2002;102:29. doi: 10.1021/cr990134z. [DOI] [PubMed] [Google Scholar]
  • 4.Salgado A, Boeykens M, Gauthier C, Declercq J-P, De Kimpe N. Tetrahedron. 2002;58:2763. [Google Scholar]
  • 5.Salgado A, Dejaegher Y, Verniest G, Boeykens M, Gauthier C, Lopin C, Tehrani KA, De Kimpe N. Tetrahedron. 2003;59:2231. [Google Scholar]
  • 6.Burtoloso ACB, Correira CRD. Tetrahedron Lett. 2004;45:3355. [Google Scholar]
  • 7.Burtoloso ACB, Correira CRD. J. Organomet. Chem. 2005;690:5636. [Google Scholar]
  • 8.Martinez RF, Fleet GWJ. Tetrahedron:Asymm. 2014;25:373. [Google Scholar]
  • 9. Adams CS, Boralsky LA, Guzei IA, Schomaker JM. J. Am. Chem. Soc. 2012;134:10807. doi: 10.1021/ja304859w. Adams CS, Grigg RD, Schomaker JM. Chem. Sci. 2014;5:3046. Boralsky LA, Marston D, Grigg RD, Hershberger JC, Schomaker JM. Org. Lett. 2011;13:1924. doi: 10.1021/ol2002418. Adams CS, Weatherly CD, Burke EG, Schomaker JM. Chem. Soc. Rev. 2014;43:3136. doi: 10.1039/c3cs60416k.
  • 10.Tehrani KA, De Kimpe N. Curr. Org. Chem. 2009;13:854. [Google Scholar]
  • 11.Feast GC, Page LW, Robertson J. Chem. Commun. 2010;46:2835. doi: 10.1039/b926179f. [DOI] [PubMed] [Google Scholar]
  • 12. Gregg TM. Tetrahedron Lett. 2010;51:6429. Gregg TM. Org. Lett. 2009;11:4434. doi: 10.1021/ol9017968.
  • 13.Sharma R, Williams LJ. Org. Lett. 2013;15:2202. doi: 10.1021/ol400749e. [DOI] [PubMed] [Google Scholar]
  • 14. Kitching W, Olszowy HA, Drew GM, Adcock W. J. Org. Chem. 1982;47:5153. Eliel EL, Wilen SH. Stereochemistry of Organic Compounds. Wiley; 1993. p. 696.
  • 15.For perspectives on Brook rearrangements in anion relay chemistry, see: Smith AB, III, Wuest WM. Chem. Commun. 2008:5883. doi: 10.1039/b810394a.

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supporting Information

RESOURCES