Gold-Catalyzed Synthesis of N,O-Heterocycles*

A. Stephen K. Hashmi

Institute of Organic Chemistry, Ruprecht-Karls-University Heidelberg, Im Neuenheimer Feld 270, 69120 Heidelberg, Germany

Abstract: The selective synthesis of oxazoles, alkylidene oxazoles, and 1,3-oxazines from N-propargyl carboxamides by choosing gold(III) or gold(I) catalysts and selecting terminal or internal alkynes as substrates is discussed. The mechanistic studies based on labeling ex- periments and intensive in situ NMR studies indicate an anti-oxyauration step and a proto- deauration with retention of the sterical arrangement at the double bond. The synthetic scope for gold(I) catalysts is very broad, allowing selective conversions to products with properties of chelate ligands, reactions of bromoarenes and the formation of acceptor-substituted oxazoles. With N-heterocyclic carbene (NHC) ligands in the presence of a base, even the vinylgold intermediates of the reactions leading to 1,3-oxazines could be isolated, a reaction with a great potential for future mechanistic studies and future applications of gold catalysis.

Keywords: alkynes; amides; gold; oxazoles; transition-metal-catalyzed reactions.

INTRODUCTION Homogeneous gold catalysis has developed to be a hot spot in catalysis research. After a pre-peak of activity initiated in 1986 by Ito, Sawarmura, and Hayashi as well as by Utimoto et al. [1], our first two papers of the year 2000 [2] have triggered an enormous activity with still continuing exponential growth. Reviews nicely summarize the many creative innovations of almost 100 groups world-wide [3]. In comparison to other transition-metal catalysts, most of the gold-catalyzed reactions are both atom- economic and remarkably mild with regard to the reaction conditions. Among these transformations, especially the intramolecular addition of heteroatom-nucleophiles to carbon–carbon triple bonds has become a powerful tool for the synthesis of various heterocycles. A first example for this reactivity pattern was provided by Utimoto et al in 1987 [1b], the intramolcecular hydroamination reactions of alkynes. These reactions usually led to tetrahydropyridines after an isomerization of the initial enamine product. The gold-catalyzed synthesis of furans by a 5-endo-trig cyclization of allenyl ketones have for the first time shown that even weak nucleophiles like the carbonyl oxygen atoms can participate in gold- catalyzed rearrangements [2a]. Following these initial findings, a large family of gold-catalyzed hetero- cycle syntheses basing on nucleophilic cyclizations involving carbonyl groups were very successfully developed by a number of groups [4]: furans were obtained by a 5-endo-dig cyclization [4a,4j], fura- nones by 5-endo-trig cyclization [4b], γ-lactones by 5-exo-dig cyclization [4e], 4-alkylidene-1,3-diox- olan-2-ones by 5-exo-dig cyclization [4g], 2-oxazolinones by 5-endo-dig cyclization [4i], 4-alkylidene- 2-oxazolidinones [4f] and 5-methylene-1,3-oxazolidin-2-ones [4h] by 5-exo-dig cyclization, even 6-membered heterocycles as 5,6-dihydro-1,3-oxazines by 6-exo-dig cyclization [4c], and oxazolinones

*Paper based on a presentation at the 9th International Conference on Heteroatom Chemistry (ICHAC-9), 30 June–4 July 2009, Oviedo, Spain. Other presentations are published in this issue, pp. 505–677.

657 658 A. S. K. HASHMI by 6-exo-dig cyclization [4d]. In addition to the initial allenyl ketone cyclizations [2a], all these very successful conversions were also preceded by the synthesis of oxazole derivatives by 5-exo-dig cyclization. The detailed synthetic and mechanistic study of this cyclization reaction will be discussed in the following sections. These details are of fundamental importance to the other cyclizations cited above and many more related reactions.

GOLD(III) PRECATALYSTS DELIVER AROMATIC OXAZOLES The work of Arcadi, Cacchi et al., who published a Pd(0)-catalyzed coupling-cyclization sequence for the synthesis of disubstituted oxazoles [5], inspired us for the gold-catalyzed synthesis of oxazoles 2 by a 5-exo-dig cyclizations of N-propargyl carboxamides 1 (Scheme 1) [6].

Scheme 1 Gold(III) chloride-catalyzed cycloisomerization of N-propargyl carboxamides 1 delivers oxazoles 2.

This reaction has a high potential for organic synthesis, as disubstituted oxazoles are often found in natural products and often exhibit interesting pharmaceutical properties. For example, such disubstituted oxazoles were reported to be anti-tumor agents, antifungal agents, herpes simplex virus type 1 (HSV-1) inhibitors, serine-threonine phosphate inhibitors, and antibacterials [7]. Most of the target compounds mentioned above are highly functionalized, thus the mild conditions [8] of the gold-catalyzed oxazole synthesis are very attractive if compared to the classical routes like condensation reactions, which often demand harsh reaction conditions [7]. Recently, the gold-catalyzed oxazole synthesis has even attracted industrial chemists and patents use the gold-catalyzed transformation as a key step [9]. The previous work of Hacksell et al. [10] has already shown that the propar- gylamide/oxazole transformation is possible with potassium tert-butoxide, in direct comparison the gold catalysts are not only convincing by their much higher functional group tolerance but also by the unprecedented mild conditions, which allow the detection and isolation of a previously assumed intermediate of the isomerization, the alkylidene oxazoline 3 (Scheme 2).

Scheme 2 Alkylidene oxazolines 3 are intermediates in the gold-catalyzed cycloisomerization of 1 to 2 and could be detected by in situ NMR spectroscopy.

The detection of the intermediate 3 by in situ NMR spectroscopy enabled us to investigate the reaction mechanism. We prepared the deuterated derivative D-1a, and by in situ NMR spectroscopy we could detect only one diastereomer of the intermediate D-3a (Scheme 3). Initially, it was not obvious, which diastereomer was formed, and due to the direct allylic coupling in the system, nuclear Overhauser effect (NOE) spectra could not be used for the determination of the double-bond geometry

Scheme 3 The deuterated alkyne D-1a in a stereoselective reaction delivered only one diastereomer of the product D-3a. of the 1,1-disubstituted olefin. Thus, we measured fully coupled 13C NMR spectra. An inset on the signal of the allylic carbon attached to nitrogen is shown in Fig. 1. The second spectrum (b) shows the signal of the undeuterated product, which shows a triplet with a large coupling constant (induced by the two hydrogen atoms directly on that carbon), a doublet with a medium coupling constant (induced by the trans-hydrogen atom, according to the Karplus equation this coupling constant is larger than the cis- coupling) and a doublet with a small coupling constant (induced by the cis-hydrogen atom). In the first spectrum (a), only the large triplet coupling and the small cis-coupling remain, a clear proof for the trans-position of the deuterium label in D-3a (the trans-allylic D-C coupling is not resolved but only visible as a broadening of the signals as the gyromagnetic ratio of H/D is 6.5).

Fig. 1 Section of the carbon NMR of (a) D-3a and (b) H-3a showing the signal of C-4 without proton decoupling.

Basing on these experimental findings, the following mechanism can be assumed (Scheme 4):

Scheme 4 Proposed elementary reactions of the isomerization of 1 to 3.

The gold catalyst coordinates the carbon–carbon triple bond of the alkyne (intermediate A) [11] and by this activates the alkynyl group for the intramolecular nucleophilic attack of the oxygen atom of the carboxamide subunit. This nucleophilic attack proceeds as a back-side attack, thus only the (E)-diastereomer of the vinylgold intermediate B is formed by the anti-addition. Now the catalytically active gold species must be liberated, this occurs by a proto-deauration step (the proton needed was liberated from the amide nitrogen during the conversion of A to B). The product 3 finally aromatizes to 2. Only if both the oxyauration of the triple bond and the protodesauration of the vinyl gold species are diastereoselective, the overall process can be diastereoselective (as demanded by the exclusive experimental observation of D-3a from D-1a). After this investigation, we had to return to the mechanism once more. In a manuscript which ap- peared at the same time, Uemura et al. reported a combination of a ruthenium-catalyzed formation of the N-propargyl carboxamides 6 from carboxamides 5 and propargyl alcohols 4 and the subsequent gold-catalyzed formation of oxazoles (Scheme 5) [12]. For the gold-catalyzed part, quite different from

Scheme 5 For the cycloisomerization of 6, an alternative pathway with the allenylamine C as an intermediate was suggested by Uemura and co-workers.

© 2010, IUPAC Pure Appl. Chem., Vol. 82, No. 3, pp. 657–668, 2010 Gold-catalyzed synthesis of N,O-heterocycles 661 our findings, they suggested the initial isomerization of 6 to a more reactive allene C prior to the cyclization. They also reported that for 6a they could detect the allene intermediate C by in situ 1H NMR spectroscopy. Such an isomerization is conceivable, the conjugation to the phenyl group might influ- ence the reaction pathway [13]. In order to verify this, we repeated the reaction of pre-formed 6a in a concentration range where the reaction was slow enough for investigation by in situ 13C NMR (Scheme 6). A combination of 13C NMR, DEPT 135 and DEPT 90 spectra clearly proved the presence of a methine group with a chemical shift of 72 ppm in the intermediate, which excluded the allenic intermediate as the corresponding alkylidene oxazole but not intermediate C possesses such a methine group (Fig. 2).

Scheme 6 The control experiment with 6a should verify or disprove the intermediacy of allene C.

Fig. 2 In situ 13C NMR spectra and assignment of the signals taken during the gold-catalyzed conversion of 6a: (a) normal proton decoupled 13C NMR; (b) DEPT 135 spectrum; (c) DEPT 90 spectrum.

The synthetic scope of the gold(III)-catalyzed oxazole synthesis is quite broad (Scheme 7), even very bulky substituents, olefinic and heterocyclic substitutents are tolerated. Only strong acceptors in conjugation to the carbonyl group reduce the nucleophilicity of the latter and the reaction fails. Several of the products have been unambiguously been characterized by crystal structure analyses.

Scheme 7 Scope and limitations of the AuCl3-catalyzed formation of substituted oxazoles from N-propargyl carboxamides.

GOLD(I) PRECATALYSTS DELIVER UNIQUE ALKYLIDENE OXAZOLINES While the gold(III) catalyst always delivered the aromatic oxazoles, with gold(I) catalysts the reaction cleanly stopped at the stage of the alkylidene oxazolines 3 (in this case, 3 is no longer an intermediate, it is a stable product), no further isomerization to the aromatic oxazole 2 was observed [14]. This was most remarkable, as the substrates 3 previously were not accessible in organic synthesis [15]. With gold(I) pre-catalysts, the reaction still followed the same mechanism as with gold(III) pre- catalysts, for example, a control experiment with the undeuterated 1b in the presence of some D2O gave 3b with 30 % deuteration trans to oxygen (Scheme 8), which was in full accordance with the anti- oxyauration and subsequent diastereoselective deutero-deauration as indicated for gold(III) pre-catalysts in Scheme 4. As the deuterim was already in the starting material D-1a, the (Z)-diastereomer of D-3a was formed, while the incorporation of deuterium from an external source during the reaction of 1b delivered the (E)-diastereomer of D-3b. Different from the example D-3a from Scheme 4, due to the

Scheme 8 Diastereoselective incorporation of deuterium during the gold(I)-catalyzed reaction of 1b confirmed the mechanistic proposal from Scheme 4.

© 2010, IUPAC Pure Appl. Chem., Vol. 82, No. 3, pp. 657–668, 2010 Gold-catalyzed synthesis of N,O-heterocycles 663 primary kinetic isotope effect (protons from the NH-group in the starting material 1b were also pres- ent), the degree of deuteration could not be quantitative in the case of 3b). The synthetic scope of the reactions with the gold(I) pre-catalysts is even wider. Just to provide one out of many examples, the substrate 1c, which failed to react with gold(III) pre-catalysts (see Scheme 7), was successfully converted by the gold(I) pre-catalysts to 3c (Scheme 9).

Scheme 9 Even acceptor-substituted substrates like 1c, which failed to convert with the gold(III) catalyst, react with the gold(I) pre-catalysts.

Scheme 10 shows selected products obtained with the gold(I) pre-catalysts. Since gold(I) clearly prefers a linear coordination geometry, potential chelate ligands like 3d do not show the phenomenon of self-poisoning, the inhibition of the catalyst by the product. The selective conversion of the furan 1e

Scheme 10 Selected products from the gold(I)-catalyzed isomerization of the N-propargyl carboxamides.

© 2010, IUPAC Pure Appl. Chem., Vol. 82, No. 3, pp. 657–668, 2010 664 A. S. K. HASHMI demonstrates that ortho-metallation by gold in neither the starting material nor the product 3e is a prob- lem. 1f shows that bromoarenes are tolerated. 1g selectively reacts at the amide unit which is more dis- tant from the electron-withdrawing nitrogen atom of the pyridine ring. The connectivity of 3g was proved by two-dimensional NMR spectra. The chemoselectivity and potential applications were nicely represented by the reaction of 1h, which delivered the fully conjugated 2h with AuCl3 and the not fully conjugated 3h with Ph3PAuNTf2 (Scheme 11) [16]. 3h does not show any fluorescence, while 2h is strongly fluorescent, which illustrates the synthetic options in the field of materials science—a carboxylic acid derivative on a π-system can be converted into an oxazole ring, which possesses interesting photochemical properties, under mild conditions in only two steps.

Scheme 11 Chemoselectivity depending on the oxidation state of the catalyst—the structural assignment was unambiguously proofed by X-ray crystal structures of both products, the ORTEP plots of the solid-state molecular structures of 2h and 3h are shown [16].

In the case of substrates with propargylic disubstitution like 1i, even gold(III) catalysts gave the alkylidene oxazolines (3i, Scheme 12). The disubstitution made a subsequent isomerization to the aromatic oxazole impossible.

Scheme 12 No aromatization is possible with substrates bearing two substituents in propargylic position, then even with gold(III) chloride as the catalyst the alkylidene oxazolines 3 are the final products, the formation of the aromatic oxazoles 2 is not possible.

A second substituent at the nitrogen atom as in substrate 1j, on the other hand, delivered an allylic amine, which best was converted to the ammonium salt 9 for isolation. The reactivity pattern was still the same, but now the proton for the protodeauration step has to be delivered from an external acid and in intermediate 8 the iminium-substructure was reactive and was hydrolyzed via intermediate D (Scheme 13).

Scheme 13 With N,N-disubstituted substrate 1j, a new product type, the allylic amine 9, is formed.

INTERNAL ALKYNES LEAD TO 1,3-OXAZINES AND ISOLABLE ORGANOGOLD INTERMEDIATES With internal alkynes 10, a different reaction mode was observed. Now a 6-endo-dig cyclization was preferred, delivering the six-membered 1,3-oxazine heterocycles. With a N-heterocyclic carbene (NHC) ligand, the intermediate E could be deprotonated by a base, delivering isolable vinylgold species 11 in good yields (Scheme 14); all attempts to achieve the same with a phosphane ligand failed [17]. These are true intermediates of the reactions, in the presence of acid they served as catalysts, too. Figure 3 shows 11a as one example which also could be characterized by a crystal structure analysis, with 2.027(5) for Au1–C1 and 2.038(6) for Au1–C11 showing almost equidistant Au–C bond lengths for both gold–carbon bond.

Scheme 14 6-endo-dig ring closure for internal alkynes 10 delivers oxazines 12.

Fig. 3 In the reactions of the internal alkynes 10, the stable intermediates of gold(I)-catalyzed reactions can be isolated in the presence of simple bases like triethylamine, here the organometallic compound 11a, derived from (IPr)Au+ and 10a (R1 = Ph, R2 = Me), is shown.

The ability of isolating such organogold intermediates [18] opens entirely new options for organic synthesis, as other electrophiles than the simple proton can be incorporated into the product in the electro philic deauration step. Apart from halogens, the most recent efforts cover the combination of a gold-catalyzed reaction with a palladium-catalyzed cross-coupling, which would allow to form new carbon–carbon bond. The crucial transmetalation step has recently been investigated in depth [19].

SUMMARY AND OUTLOOK

The selectivity of the reaction is controlled by the oxidation state of the pre-catalyst, AuCl3 delivers the aromatic oxazoles, gold(I) catalysts always provide the five-membered alkylidene oxazolines. The intermediacy of the latter was clearly proven. The reaction mechanism can conveniently be studied by in situ NMR spectroscopy, indicating an anti-addition for the oxyauration step and a proto-deauration which does not change the sterical arrangement of the substituents on the double bond. For the gold(I) catalysts, the substrate scope is very broad, products which potentially serve as a chelating ligand can be obtained and bromoarenes as well as acceptor-substituted substrates are tolerated. In the case of internal alkynes, the six-membered 1,3-oxazines are obtained, in the presence of base and an NHC lig- and, even the intermediate vinylgold(I) species, can be isolated. These examples show that gold catalysis can be of great benefit for the synthesis of heterocyclic organic compounds. For the reaction type presented here, several recent reports in the literature, covering the use of different ligand systems on the gold center [20], the use of other transition metals as catalysts [21] as well as a Brønsted acid-catalyzed variant (under reflux conditions) [22] show the topicality of the subject and the superiority of the gold catalysts. The results described here will probably allow the isolation of numerous vinylgold intermediates in the future, and the option to transmetalate from these vinylgold species might become a new important topic in catalysis research.

REFERENCES 1. (a) Y. Ito, M. Sawamura, T. Hayashi. J. Am. Chem. Soc. 108, 6405 (1986); (b) Y. Fukuda, K. Utimoto, H. Nozaki. Heterocycles 25, 297 (1987). 2. (a) A. S. K. Hashmi, L. Schwarz, J.-H. Choi, T. M. Frost. Angew. Chem. 112, 2382 (2000); (b) A. S. K. Hashmi, L. Schwarz, J.-H. Choi, T. M. Frost. Angew. Chem., Int. Ed. 39, 2285 (2000); (c) A. S. K. Hashmi, T. M. Frost, J. W. Bats. J. Am. Chem. Soc. 122, 11553 (2000); see also the discussion in: (d) A. S. K. Hashmi. Angew. Chem. 112, 3737 (2000); (e) A. S. K. Hashmi. Angew. Chem., Int. Ed. 39, 3590 (2000). 3. For some representative reviews on gold catalysis, see: (a) A. S. K. Hashmi. Gold Bull. 36, 3 (2003); (b) A. S. K. Hashmi. Gold Bull. 37, 515 (2004); (c) A. S. K. Hashmi, G. J. Hutchings. Angew. Chem. 118, 8064 (2006); (d) A. S. K. Hashmi, G. J. Hutchings. Angew. Chem., Int. Ed. 45, 7896 (2006); (e) A. Fürstner, P. W. Davis. Angew. Chem. 119, 3478 (2007); (f) A. Fürstner, P. W. Davis. Angew. Chem., Int. Ed. 46, 3410 (2007); (g) D. J. Gorin, F. D. Toste. Nature 446, 395 (2007); (h) A. S. K. Hashmi. Chem. Rev. 107, 3180 (2007); (i) A. Arcadi. Chem. Rev. 108, 3266 (2008); (j) E. Jiménez-Núñez, A. M. Echavarren. Chem. Rev. 108, 3326 (2008); (k) Z. G. Li, C. Brouwer, C. He. Chem. Rev. 108, 3239 (2008). 4. (a) T. Yao, X. Zhang, R. C. Larock. J. Am. Chem. Soc. 126, 11164 (2004); (b) S. Shin. Bull. Korean Chem. Soc. 26, 1925 (2005); (c) J.-E. Kang, H.-B. Kim, J.-W. Lee, S. Shin. Org. Lett. 8, 3537 (2006); (d) R. Robles-Machin, J. Adrio, J. C. Carretero. J. Org. Chem. 71, 5023 (2006); (e) E. Genin, P. Y. Tullec, S. Antoniotti, C. Brancour, J.-P. Genêt, V. Michelet. J. Am. Chem. Soc. 128, 3112 (2006); (f) S. Ritter, Y. Horino, J. Lex, H.-G. Schmalz. Synlett 3309 (2006); (g) A. Buzas, F. Gagosz. Org. Lett. 8, 515 (2006); (h) A. Buzas, F. Gagosz. Synlett 2727 (2006); (i) A. S. K. Hashmi, R. Salathé, W. Frey. Synlett 1763 (2007); (j) V. Belting, N. Krause. Org. Biomol. Chem. 7, 1221 (2009). 5. A. Arcadi, S. Cacchi, L. Cascia, G. Fabrizi, F. Marinelli. Org. Lett. 3, 2501 (2001). 6. A. S. K. Hashmi, J. P. Weyrauch, W. Frey, J. W. Bats. Org. Lett. 6, 4391 (2004). 7. (a) D. C. Palmer, S. Venkatraman. In The Chemistry of Heterocyclic Compounds, A Series of Monographs, Oxazoles: Synthesis, Reactions, and Spectroscopy, Part A, D. C. Palmer (Ed.), John Wiley, Hoboken (2003); (b) S. Lang-Fugmann. In Methoden der Organischen Chemie (Houben- Weyl), Band E8a Hetarene III/Teil 1, E. Schaumann (Ed.), Georg Thieme Verlag, Stuttgart (1993); (c) R. Lakhan, B. Ternai. Adv. Heterocycl. Chem. 17, 99 (1974); (d) G. J. Pattenden. Heterocycl. Chem. 29, 607 (1992). 8. A recent review summarized the application of gold catalysts in total synthesis: A. S. K. Hashmi, M. Rudolph. Chem. Soc. Rev. 37, 1766 (2008). 9. (a) D. Dorsch, L. T. Burgdorf, R. Gericke, N. Beier, W. Mederski. WO 2005123688 A2, (2005); (b) D. Dorsch, L. T. Burgdorf, R. Gericke, N. Beier, W. Mederski. Chem. Abstr. 144, 69837 (2005); (c) D. Dorsch, O. Schadt, A. Blaukat, F. Stieber. WO 2008/017361 A2; (d) D. Dorsch, O. Schadt, A. Blaukat, F. Stieber. Chem. Abstr. 148, 190536 (2008). 10. B. M. Nilsson, U. Hacksell. J. Heterocycl. Chem. 26, 269 (1989). 11. For the first example of a crystal structure analysis of such a π-complex with gold(I), see: P. Schulte, U. Behrens. Chem. Commun. 1633 (1998). 12. M. D. Milton, Y. Inada, Y. Nishibayashi, S. Uemura. Chem. Commun. 2712 (2004). 13. (a) A. S. K. Hashmi. In Modern Allene Chemistry, A. S. K. Hashmi, N. Krause (Eds.), pp. 3–50, Wiley-VCH, Weinheim (2004); (b) A. S. K. Hashmi. In Science of Synthesis: Vol. 44 Compounds with All-Carbon Functions, N. Krause (Ed.), pp. 287–299; chap. 44.2.4, Thieme, Stuttgart (2007). 14. (a) A. S. K. Hashmi, M. Rudolph, S. Schymura, J. Visus, W. Frey. Eur. J. Org. Chem. 4905 (2006); compare also one single example observed in: (b) C. Ferrer, A. M. Echavarren. Angew. Chem. 118, 1123 (2006); (c) C. Ferrer, A. M. Echavarren. Angew. Chem., Int. Ed. 45, 1105 (2006).

15. Only three examples with a disubstituted carbon atom next to nitrogen, which prevent aromatization, were registered in the chemical databases. 16. J. P. Weyrauch, A. S. K. Hashmi, A. Schuster, T. Hengst, S. Schetter, A. Littmann, M. Rudolph, M. Hamzic, J. Visus, F. Rominger, W. Frey, J. W. Bats. Chem.—Eur. J. 16, 956 (2009). 17. (a) A. S. K. Hashmi, A. Schuster, F. Rominger. Angew. Chem. 121, 8396 (2009); (b) A. S. K. Hashmi, A. Schuster, F. Rominger. Angew. Chem., Int. Ed. 48, 8247 (2009). 18. A. S. K. Hashmi, T. D. Ramamurthi, F. Rominger. J. Organomet. Chem. 694, 592 (2009). 19. (a) A. S. K. Hashmi, C. Lothschütz, R. Döpp, M. Rudolph, T. D. Ramamurthi, F. Rominger. Angew. Chem. 121, 8392 (2009); (b) A. S. K. Hashmi, C. Lothschütz, R. Döpp, M. Rudolph, T. D. Ramamurthi, F. Rominger. Angew. Chem., Int. Ed. 48, 8243 (2009). 20. (a) D. Aguilar, M. Contel, R. Navarro, T. Soler, P. E. Urriolabeitia. J. Organomet. Chem. 694, 486 (2009); (b) G. Verniest, A. Padwa. Org. Lett. 10, 4379 (2008); (c) G. Verniest, A. Padwa. Org. Lett. 10, 4379 (2008). 21. (a) E. Merkul, T. J. J. Müller. Chem. Commun. 4817 (2006); (b) E. Merkul, O. Grotkopp, T. J. J. Müller. Synthesis 3, 502 (2009); (c) E. M. Beccalli, E. Borsini, G. Broggini, G. Palmisano, S. Sottocornola. J. Org. Chem. 73, 4746 (2008); (d) M. Harmata, C. Huang. Synlett 1399 (2008). 22. Y.-M. Pan, F.-J. Zheng, H.-X. Lin, Z.-P. Zhan. J. Org. Chem. 74, 3148 (2009).