DOI: 10.1002/adsc.201100632

Intramolecular Carbonyl Nitroso Ene Reaction Catalyzed byIron ACHTUNGTRENNUNG(III) Chloride/Hydrogen Peroxide as an Efficient Tool forDirect Allylic Amination

Duncan Atkinson,a Mikhail A. Kabeshov,b Mark Edgar,a and Andrei V. Malkova,*a Department of Chemistry, Loughborough University, Loughborough, LE11 3TU, U.K.

Fax: (+44)-1509-22-3925; e-mail: a.malkov@lboro.ac.ukb Current address: Chemistry Research Laboratory, University of Oxford, 12 Mansfield Road, Oxford, OX1 3TA, U.K.

Received: August 8, 2011; Published online: December 8, 2011

Dedicated to Professor Pavel Kocovsky on the occasion of his 60th birthday.

Supporting information for this article is available on the WWW underhttp://dx.doi.org/10.1002/adcs.201100632.

Abstract: A mild, simple oxidation protocol em-ploying iron ACHTUNGTRENNUNG(III) chloride as a catalyst and hydro-gen peroxide as a stoichiometric oxidant was foundto be compatible with an intramolecular carbonylnitroso ene reaction and allowed us to efficientlyconvert hydroxamic acids into a diverse range of1,2- and 1,3-amino alcohol derivatives in a singleoperation.

Keywords: amination; amino alcohols; cyclization;ene reaction; oxidation

Amino alcohols, particularly 1,2- and 1,3-derivatives,are commonly featured in many pharmaceuticals andnatural products, and they also find wide use as syn-thetic building blocks.[1] In recent years, syntheticstrategies towards these structural motifs employingC�H functionalization as a key bond-forming eventare increasingly gaining momentum.[2]

Intramolecular allylic amination (1!2, Scheme 1)can be accomplished in the presence of Pd(II)[3,4] andRh(II)[5] catalysts, however, in the latter case allylicC�H insertion of nitrenes competes with aziridina-tion, furnishing a mixture of the respective products 2and 3 (Scheme 1).

The nitroso ene reaction is another type of allylicamination that can give rise to similar products with-out using noble metal catalysts.[6] An intramolecularvariant of the carbonyl nitroso ene reaction was re-ported by Kirby[7] more than two decades ago butsince then it has remained unexplored (Scheme 2, 4!5!6!7). Due to the instability of nitroso compounds5, they were prepared in situ by low temperature per-ACHTUNGTRENNUNGiodate oxidation of the corresponding hydroxamicacids 4 and intercepted with cyclopentadiene to formadducts 6, which then, after isolation, were thermallyconverted into 7, presumably through the transitionstate A. This method represents a complementary al-ternative to the Pd- and Rh-catalyzed aminations:here the new C�N bond is formed with a concomitantallylic shift of the double bond. However, the synthet-ic potential of this method is limited by the need toisolate the Diels–Alder adducts. Herein, we report onthe development of a catalytic oxidation protocol toeffect cyclization of hydroxamic acids 4 into 7 in asingle step. While this manuscript was in preparation,a similar single-pot intramolecular acyl nitroso reac-tion was reported by Read de Alanis.[8]

Scheme 1. Transition metal-catalyzed intramolecular amina-tion. Scheme 2. Intramolecular carbonyl nitroso ene reaction.

Adv. Synth. Catal. 2011, 353, 3347 – 3351 � 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim 3347

COMMUNICATIONS

In the past, the procedures for the oxidation of hy-droxamic acids were tailored for cycloaddition of ni-troso derivatives to dienes (e.g., 5!6, Scheme 2).[9]

The ene reaction is substantially slower than cycload-dition. Strong oxidants employed in the cycloadditionmethods at prolonged exposures appear to destroyboth the carbonyl nitroso intermediates and the eneadducts and therefore, for the successful ene reaction,milder oxidation protocols are required. Followingearlier reports on a single-pot intermolecular carbonylnitroso ene reaction, which usually requires excessalkene,[10] we set out to investigate a more challengingintramolecular variant with an inherent 1:1 stoichiom-etry. Preliminary experiments were carried out em-ploying model substrates 8a and 8b derived from (E)-crotyl alcohol and (E)-2-hexenol (Table 1). Cycliza-tion of hydroxamic acid 8a (R =Me) proceededsmoothly in the presence of catalytic Cu(II) triflate(2 mol%) and 50% aqueous H2O2 (1.2 equiv.) to fur-nish the respective product 9a in 70% yield (Table 1,entry 1). For hexenyl derivative 8b (R= n-Pr), thesame Cu(II) system proved inefficient resulting in de-composition of the starting material. Changing thestoichiometric oxidant to t-BuOOH essentially repro-duced the results obtained with H2O2 (Table 1,entry 2); raising the reaction temperature to 60 8Cwas not helpful (Table 1, entry 3). A brief screeningof catalytic systems[11] revealed that MoO5ACHTUNGTRENNUNG(HMPA)Py(2 mol%) with 6 equiv. of H2O2 at 60 8C did showsome activity providing 9b in 45% yield (Table 1,entry 4). However, the best results were obtainedwith Fe ACHTUNGTRENNUNG(III) as a catalyst. Crotyl derivative 8a wasreadily cyclized into 9a in the presence of FeCl3·6 H2O

(4 mol%) and 50% aqueous H2O2 (1.2 equiv.) inMeOH at room temperature. The less reactive 8bunder these conditions gave only 10% of 9b (Table 1,entry 5), however raising the temperature to 60 8Cproduced the desired 9b in 56% yield as a 4:1 E/Zmixture (Table 1, entry 6). Interestingly, a further in-crease of the reaction temperature to 100 8C (the sol-vent was changed to i-PrOH) improved the E/Z ratioof 9b to 6:1 (yield 64%, Table 1, entry 7).

Oxidation of the hydroxamic acids to the respectivenitroso intermediates is likely to be carried out by themetal. The role of the stoichiometric oxidant is to re-generate the catalytically active metal species, as with-out the oxidant no catalytic turnover was observed(Table 1, entry 8). It is worth noting that the catalyticsystem based on FeACHTUNGTRENNUNG(III)/H2O2, known to promote ep-oxidation of alkenes,[12] under the reaction conditionsemployed here did not produce epoxides.

The apparent difference in reactivity of 8a and 8bprompted us to have a closer look at the steric factorsinfluencing removal of the allylic hydrogen(Scheme 3). In the set of substrates 8a–d, the reactivi-ty drops dramatically in the order Me>n-Pr>Bn> i-Pr, where 9c was isolated in just 10%, whereas 9d wasnot formed at all,[13] which reflects increase of thesteric congestion around the allylic C�H bond. Evenformation of the thermodynamically stable fragments,such as styrene (9c) or trisubstituted alkene (9d), wasnot able to offset this trend.

Next, we investigated diastereoselectivity of the ni-troso ene cyclization. To circumvent the issues of re-activity, secondary crotyl analogues 10a–d were select-ed (Scheme 4). By employing the reaction conditions

Table 1. Optimization of the intramolecular nitroso ene cyclization.[a]

Entry Catalyst (mol%) Oxidant Solvent, T [oC] Yield[b] 9a [%] Yield[b] 9b [%]

1 CuACHTUNGTRENNUNG(OTf)2, (2) H2O2 THF, r.t. 70 0[c]

2 CuACHTUNGTRENNUNG(OTf)2, (2) t-BuOOH THF, r.t. 69 0[c]

3 CuACHTUNGTRENNUNG(OTf)2, (2) H2O2 THF, 60 n/a 0[c]

4 MoO5 ACHTUNGTRENNUNG(HMPA)·py (2) H2O2[d] MeOH, 60 n/a 45]

5 FeCl3 (4) H2O2 MeOH, r.t. 60 106 FeCl3 (4) H2O2

[d] MeOH, 60 n/a 56,e]

7 FeCl3 (4) H2O2[d] i-PrOH, 100 n/a 64[f]

8 FeCl3 (25) – i-PrOH, r.t. 30 n/a

[a] The reactions were carried out on a 0.5-mmol scale with 1.2 equiv. of oxidant for 18 h, unless stated otherwise.[b] Isolated yield.[c] Not formed, starting material decomposed.[d] 6 equiv. of oxidant were used.[e] E/Z 4:1.[f] E/Z 6:1.

3348 asc.wiley-vch.de � 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Adv. Synth. Catal. 2011, 353, 3347 – 3351

COMMUNICATIONS Duncan Atkinson et al.

developed for 8a (4 mol% FeCl3·6 H2O and 1.2 equiv.H2O2 in i-PrOH at room temperature), cyclizationproceeded uneventfully to furnish mixtures of syn andanti products 11a–d. Rather surprisingly, a 2:1 syn/antimixture was formed regardless of the steric size of thesubstituent. Relative configurations of the stereoiso-mers were established by analysis of their 1H NMRspectra (for details, see Supporting Information). Var-iation of solvents (CH2Cl2, CHCl3, THF, toluene) andcatalyst [Cu ACHTUNGTRENNUNG(OTf)2, ligand free and as a complex withphenanthroline] had no effect on the syn/anti ratio.However, the diastereoselectivity was improved bylowering the reaction temperature. The reactivity ofhydroxamic acids 10 proved to be sufficient to attaingood conversion at �20 8C in 72 h. Thus, in the caseof 11d diastereoselectivity increased to 5:1 (yield51%). It is worth noting that the diastereoisomers arereadily separable by chromatography.

The scope of the reaction was investigated with theaid of hydroxamic acids 12, 14, 15, 17, 19 and 21 rep-resenting diverse substitution patterns (Table 2).Prenyl derivative 12 with a relatively electron-richdouble bond proved to be one of the most reactivesubstrates. The reaction was complete in less than 3 hwith both Fe ACHTUNGTRENNUNG(III) and Cu(II) catalysts furnishing therespective oxazolidinone 13 in good yields (Table 2,entry 1). E-Geometry of the alkene appears to be an

important factor affecting the cyclization, as 14, a Z-isomer of 8a, failed to give any product. Formation ofthe 6-membered ring required more forcing condi-tions than the 5-membered counterpart. At roomtemperature, cyclization of hydroxamic acid 15 into16 took 72 h (58% yield), but at 100 8C it was com-plete overnight (yield 60%, Table 2, entry 3). Cycliza-tion of the more substituted derivatives 17, 19 and 21required elevated temperature and was accomplishedin good yields (Table 2, entries 4–6). Importantly, incontrast to the high sensitivity of the cyclization tothe surroundings of the allylic C�H bond, steric fac-tors do not play any significant role in the formationof C�N bond, which is manifested by a facile forma-tion of spiro derivative 20 and a highly substituted ox-azolidinone 22 (Table 2, entries 5 and 6, respectively).Furthermore, 22 was obtained in high diastereoselec-tivity (dr>25:1).

Preliminary DFT calculations[14] suggest a pericyclic6-membered transition state D (Scheme 5) as a possi-ble key point of the reaction mechanism.[15,16] Wehave also identified formation of aziridine N-oxide C,however the extent of its contribution to the overallreaction mechanism is not clear at the moment. Atthe same time, we were unable to find a polarized bir-adical intermediate that played an important role inthe case of aryl and alkyl nitroso compounds.[16] Ab-sence of biradical species is indirectly supported bythe different reactivities of E-8a and Z-14, since oth-erwise they would converge to the same biradical in-termediate due to a rapid rotation about single bonds.The high relative energy of TS D (25 kcal mol�1) mayexplain the observed relatively slow rate of cycliza-tion. In the presence of CuCl2, the activation energyof H-transfer was found to be 10 kcalmol�1 lower(15 kcal mol�1 vs. 25 kcal mol�1). This suggests that inour catalytic system transition metal species may playthe dual role of both oxidant and Lewis acid.

Scheme 3. Steric effects in the intramolecular carbonyl ni-troso ene cyclization.

Scheme 4. Diastereoselectivity in the intramolecular carbonyl nitroso ene reaction.

Adv. Synth. Catal. 2011, 353, 3347 – 3351 � 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim asc.wiley-vch.de 3349

Intramolecular Carbonyl Nitroso Ene Reaction Catalyzed by Iron ACHTUNGTRENNUNG(III) Chloride/Hydrogen Peroxide

Experimental Section

General Method for Cyclization of Hydroxamic Acidswith Methyl Hydrogen Abstraction (9a, 11a–d)

50% aqueous H2O2 (0.041 mL, 0.6 mmol) was added to amixture of the hydroxamic acid (0.5 mmol) and FeCl3·6 H2O(5.4 mg, 4 mol%), in i-PrOH (4 mL), at 0 8C. The reactionmixture was brought to room temperature and left for 18 h.The mixture was then diluted with 1 M HCl (20 mL) and ex-tracted with ethyl acetate (3� 30 mL). The combined organicfractions were evaporated at reduced pressure and the crudeproduct was purified by chromatography on a column ofsilica gel with a mixture of petroleum ether (40–60) andAcOEt (2:1!1:1).

General Procedure for Cyclization of HydroxamicAcids with Methylene Hydrogen Abstraction (9b, 9c,18, 20, 22)

FeCl3·6 H2O (5.4 mg, 4 mol%) was added to the appropriatehydroxycarbamate (0.5 mmol) in i-PrOH (6 mL), and, afterthe addition of 50% H2O2 (0.20 mL, 3 mmol), heated at100 8C in a sealed vessel for 18 h. The mixture was then di-luted with 1 M HCl (20 mL) and extracted with ethyl acetate(3 �30 mL). The combined organic layers were evaporatedat reduced pressure and the crude product was purified bychromatography on a column of silica gel with a mixture ofpetroleum ether (40–60) and AcOEt (2:1!1:2).

Acknowledgements

We thank Loughborough University for a studentship toD. A.

References

[1] For overviews, see a) D. J. Ager, I. Prakash, D. R.Schaad, Chem. Rev. 1996, 96, 835–876; b) C. H. Sena-nayake, Aldrichimica Acta 1998, 31, 3–15; c) S. M. Lait,D. A. Rankic, B. A. Keay, Chem. Rev. 2007, 107, 767–796.

[2] For recent overviews, see a) H. M. L. Davies, J. R.Manning, Nature 2008, 451, 417–424; b) F. Collet, R. H.Dodd, P. Dauban, Chem. Commun. 2009, 5061–5074.

Table 2. Scope of the intramolecular nitroso ene cycliz-ACHTUNGTRENNUNGation.[a]

Entry Substrate T [oC] Product Yield[b] [%]

1 r.t.[c] 60 (83)[d]

2 100[e] 0[f]

3 100[g] 60 (58)[h]

4 100[e] 60

5 100[e] 65

6 100[e] 90[i]

[a] The reactions were carried out on a 0.5-mmol scale withFeCl3 (4 mol%) and 50% aqueous H2O2, in i-PrOH for18 h at the temperature indicated.

[b] Isolated yield.[c] 1.2 equiv. of oxidant were used.[d] Yield in parenthesis is for the reaction catalyzed by

CuACHTUNGTRENNUNG(OTf)2 (2 mol%).[e] 6 equiv. of oxidant were used.[f] Not formed, starting material decomposed.[g] 3 equiv. of oxidant were used.[h] Yield in parenthesis is for the reaction at room tempera-

ture (72 h).[i] Major diastereoisomer (dr>25:1).

Scheme 5. Possible mechanism for the intramolecular nitroso ene cyclization.

3350 asc.wiley-vch.de � 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Adv. Synth. Catal. 2011, 353, 3347 – 3351

COMMUNICATIONS Duncan Atkinson et al.

[3] For Pd-catalyzed allylic C�H amination, see a) K. J.Fraunhoffer, M. C. White, J. Am. Chem. Soc. 2007, 129,7274–7276; b) G. T. Rice, M. C. White, J. Am. Chem.Soc. 2009, 131, 11707–11711.

[4] For the synthesis of 1,2-diamines by an alternative pro-cedure based on an aminopalladation/b-eliminationmechanism, accompanied by a 1,3-allylic shift, see:R. I. McDonald, S. S. Stahl, Angew. Chem. 2010, 122,5661–5664; Angew. Chem. Int. Ed. 2010, 49, 5529–5532.

[5] For Rh-catalyzed reactions, see a) H. Lebel, K. Huard,S. Lectard, J. Am. Chem. Soc. 2005, 127, 14198–14199;b) C. J. Hayes, P. W. Beavis, L. A. Humphries, Chem.Commun. 2006, 4501–4502. For Cu-catalyzed reactions,see: c) D. N. Barman, K. M. Nicholas, Eur. J. Org.Chem. 2011, 908–911.

[6] For overviews, see a) M. Johannsen, K. A. Jorgensen,Chem. Rev. 1998, 98, 1689–1708; b) W. Adam, O.Krebs, Chem. Rev. 2003, 103, 4131–4146; c) S. Iwasa, A.Fakhruddin, H. Nishiyama, Mini-Rev. Org. Chem. 2005,2, 157–175.

[7] a) G. W. Kirby, H. McGuigan, D. J. McLean, J. Chem.Soc. Perkin Trans. 1 1985, 1961–1966; for a related acyl-nitroso-ene reaction, see: b) G. E. Keck, R. R. Webb,Tetrahedron Lett. 1979, 20, 1185–1186; c) G. E. Keck,R. R. Webb, J. B. Yates, Tetrahedron 1981, 37, 4007–4016; for an acylaza-ene reaction, see: d) E. Vedejs,G. P. Meier, Tetrahedron Lett. 1979, 20, 4185–4188;e) M. Scartozzi, R. Grondin, Y. Leblanc, TetrahedronLett. 1992, 33, 5717–5720.

[8] C. P. Frazier, J. R. Engelking, J. Read deAlanis, J. Am.Chem. Soc. 2011, 133, 10430–10433.

[9] For overviews, see: a) G. W. Kirby, Chem. Soc. Rev.1977, 6, 1–24; b) H. Yamamoto, N. Momiyama, Chem.Commun. 2005, 3514–3525; c) Y. Yamamoto, H. Yama-

moto, Eur. J. Org. Chem. 2006, 2031–2043; d) B. S.Bodnar, M. J. Miller, Angew. Chem. 2011, 123, 5746–5764; Angew. Chem. Int. Ed. 2011, 50, 5630–5647.

[10] a) W. Adam, N. Bottke, O. Krebs, C. R. Saha-Mçller,Eur. J. Org. Chem. 1999, 1963–1965; b) A. Fakhruddin,S. Iwasa, H. Nishiyama, K. Tsutsumi, Tetrahedron Lett.2004, 45, 9323–9326; c) B. Kalita, K. M. Nicholas, Tetra-hedron Lett. 2005, 46, 1451–1453. For recent reports onmild protocols for the oxidation of hydroxamic acids inapplication to Diels–Alder reactions, see: d) M. F. A.Adamo, S. Bruschi, J. Org. Chem. 2007, 72, 2666–2669;e) D. Chaiyaveij, L. Cleary, A. Batsanov, T. B. Marder,K. J. Shea, A. Whiting, Org. Lett. 2011, 13, 3442–3442.

[11] Catalytic systems based on RuCl3/H2O2, RuCl3/PhIO,VOACHTUNGTRENNUNG(acac)2/t-BuOOH showed low-to-moderate conver-sions in the case of 8a but all failed in the case of 8b.

[12] For an overview, see a) A. C. Mayer, C. Bolm, in: IronCatalysis in Organic Chemistry (Ed. B. Plietker), Wiley-VCH, Weinheim, 2008, pp 73–124; for a recent asym-metric version, see: b) F. G. Gelalcha, B. Bitterlich, G.Anilkumar, M. K. Tse, M. Beller, Angew. Chem. 2007,119, 7431–7435; Angew. Chem. Int. Ed. 2007, 46, 7293–7296.

[13] Formation of the respective nitroso intermediate doestake place, as it can be intercepted with 2,3-dimethylcy-clobutadiene as a Diels–Alder adduct in 74% yield. Inthe absence of diene, due to a very slow nitroso ene re-action the nitroso compounds decompose by alterna-tive routes.

[14] For details, see the Supporting Information.[15] X. Lu, Org. Lett. 2004, 6, 2813–2815.[16] A. G. Leach, K. N. Houk, J. Am. Chem. Soc. 2002, 124,

14820–14821.

Adv. Synth. Catal. 2011, 353, 3347 – 3351 � 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim asc.wiley-vch.de 3351

Intramolecular Carbonyl Nitroso Ene Reaction Catalyzed by Iron ACHTUNGTRENNUNG(III) Chloride/Hydrogen Peroxide