Tracking gold acetylides in gold(I)-catalyzed cycloisomerization reactions of enynes

Abstract

The intermediacy of gold acetylides in the gold(I)-catalyzed cycloisomerization of enynes was questioned. While dinuclear gold complexes are observed under electrospray ionization conditions, the solution reactivity of gold acetylides also leads to the conclusion of their high affinity for the second coordination of a gold moiety, leading to dinuclear gold complexes. However, the involvement of gold acetylides and the corresponding diaurated species in the elementary steps of cycloisomerization mechanisms of enynes appears unlikely.

---

Gold-catalyzed reactions have lately emerged as a very powerful tool in organic synthesis.¹ Although they have been increasingly documented, the true nature of the intermediates involved in these reactions remains quite uncertain.² Mechanistic scenarios have been postulated, yet characterizations of putative intermediates that could support these hypotheses are still scarce³ and often lead to surprises. For instance, the formation of gem-diaurated species during gold(I)-catalyzed nucleophilic addition to alkynes and allenes has been recently experimentally demonstrated (Scheme 1, eqns (1) and (2)).⁴ Compelling evidence that gold-complexes of terminal alkynes can evolve into polynuclear species has also been provided.⁵ On the basis of deuterium-labeling experiments and DFT computations, it was also suggested that gold-catalyzed transformations involving terminal alkynes could involve, as elementary mechanistic steps, the formation of gold acetylides and the corresponding gem-diaurated species (eqn (3)).⁶

Scheme 1 The prototypical formation of gem-diaurated species.

Although in the case of the allenes depicted in eqn (2), the formation of the dinuclear complex was attributed to a catalyst resting state, diaurated species have also been considered as reactive intermediates.⁷ This was the case for the (Ph₃PAu)₃O⁺-catalyzed cycloisomerization of 1,5-allenynes, for which computations supported a dual activation mechanism, as shown in Scheme 2.⁶ Rather than a typical Au-triggered nucleophilic attack of the allene onto the complexed triple bond (pathway (i)), the 5-exo-dig cyclization is preceded by the formation of a gold acetylide and the corresponding diaurated species (pathway (ii)).

Scheme 2 Monaurated (i) vs. diaurated (ii) mechanistic proposals for the gold(I)-catalyzed cycloisomerization of 1,5-enynes involving diaurated species.

This type of reaction can be categorized within the vast area of 1,n-enyne cycloisomerizations, which has attracted considerable attention due to the high versatility of the process and its exceptional synthetic potential.^8,9 We decided to check the validity of the dual activation mechanism depicted in Scheme 2 for the cycloisomerization of widely used 1,6-enynes. Using a combined approach involving mass spectrometry/solution reactivity/theoretical treatment,¹⁰ we herein provide new mechanistic elements.

As a background for our study, we investigated the reactivity of enynes 1 and 2 with three cationic gold(I)-catalysts A,¹¹B¹² and C,¹³ which resulted in the very major formation of the 6-membered ring derivatives 3 and 5, accompanied by the formal metathesis products 4 and 6,¹⁴ respectively (Table 1).

Table 1 Background reactions

a 60% of starting material recovered, no reaction at RT. b 55% of starting material recovered. c 38% of starting material recovered, no reaction at RT.
Entry	Substrate	Cat.	Conditions	% Yield(endo:exo)
1	1	A	RT, 1 h	87(20 : 1)
2	1	B	reflux, 7 h	35(1 : 0)^a
3	1	C	RT, 1 h	quant. (5 : 1)
4	2	A	RT, 3 d	38(6 : 1)^b
5	2	B	reflux, 15 h	47(4 : 1)^c
6	2	C	RT,1 h	95(1.6 : 1)

---

Then, we engaged enyne 1 with A–C in mass spectrometry conditions (Scheme 3, see also Figs S1–3 in the supporting information†). A solution of 1 and catalysts A–C in CDCl₃ was subject to electrospray ionization (ESI) in positive mode and the resulting adduct ions were analyzed in a modified triple-quadrupole instrument.¹⁵ After mixing 1 with A or B, a peak corresponding to the formula [1–H + (Ph₃PAu)₂]⁺ (1-Au₂) was found at m/z 1099.¹⁶ With 1 and C, the same type of adduct of formula [1–H + ((biphenyl-2-yl)(t-Bu)₂PAu)₂]⁺ (1′-Au₂2) was recorded at m/z 1171. The observation of these ions suggests the formation of gold acetylides, on which a second gold entity coordinates, thus creating gold dinuclear species. This was further corroborated by submitting the independently synthesized gold acetylide 1-Au^17,18 to A–C or AgSbF₆, leading to the expected peaks of the corresponding dinuclear species.¹⁹ The latter were next fragmented by Collision Induced Dissociation (CID), giving rise to cations bearing all of the phosphines. Therefore, in all cases, the same neutral counterpart of formula [1–H + Au] (loss of 378) was formed during CID. Because structural pieces of information could not be gleaned from fragment analysis, we carried out DFT computations at the B3LYP/LANL2DZ(Au)/6-31G(d,p) level of theory using hept-1-en-6-yne and PH₃ as models (Scheme 4). From the dissociation free energies, one can notice that the naked gold acetylide or the related chelate require prohibitive energies to form under CID conditions. Thus, the neutral counterpart obtained after CID could be an already cyclized form of the enyne bound to gold.

Scheme 3 The identification of dinuclear species (biph = biphenyl-2-yl, CID = Collision Induced Dissociation).

Scheme 4 Calculated free energies of dissociation (kcal mol⁻¹).

This preliminary observation was intriguing. As mentioned above, gold acetylides and the corresponding diaurated species have been postulated as putative intermediates by Toste⁶ and by Gagosz,⁷ respectively, in the cycloisomerization of 1,5-allenynes and diynes. Although ESI spectrometry showed that such species could indeed be obtained, the transfer of this information to a solution phase catalytic event should be made with care. This drove us to the examination of the solution reactivity by NMR spectrometry. Monitoring the reaction of 1 with catalysts A or C could not be achieved due to their high activity, even at low temperatures. On the other hand, catalyst B requires heat to be active (see Table 1). At RT, using 0.33 equiv. of B, diagnostic ¹H NMR signals of the free enyne, such as the propargylic protons at 2.16 ppm, were still clearly present after 5 min (Scheme 5). A new compound was observed in the ³¹P spectra, displaying a peak at 38.4 ppm. In ¹H NMR, the propargylic protons were downfield shifted to 2.68 ppm, whereas the ethylenic ones showed virtually no changes. A broad peak at 1.64 ppm was also clearly visible. The addition of D₂O to the mixture caused the disappearance of this signal, suggesting complexed water (see the supporting information for details†). However, none of ¹H and ³¹P peaks correspond to those of the independently prepared dinuclear species 1-Au₂2BF₄4, the propargylic protons of this complex resonating at 2.80 ppm and the phosphorus atoms at 37.1 ppm. We envisaged the formation of the trinuclear dimer 7 at RT, in which two gold acetylide moieties share the same Ph₃PAu⁺ unit. This hypothesis could be supported by a new mass spectrometry experiment: using a low cone voltage (leading to a low internal energy), the fragile species 7 was clearly observed in the ESI spectra at m/z 1739 (see the supporting information for details, Figs S4–S5†). Presumably, aggregate 7 corresponds to a stabilized form of 1-Au₂2BF₄4 in solution and not a cycloisomerization intermediate. Species 7 can be observed in mass spectrometry only at extremely low energy conditions because of its instability in the gas phase in contrast to 1-Au₂2BF₄4.

Scheme 5 NMR monitoring of enynes 1 and 1-D with catalyst B.

Thus, upon the addition of 0.33 equiv. of B, if our assignment is correct, a 36 : 64 ratio between 7 and 1 can be estimated from ¹H NMR. It changed to 40 : 60 after 1 h. Increasing the amount of B to 0.66 equiv. gave rise to a 63 : 37 mixture of 7 and 1 after 5 min and only 7 could be observed after 1 h. With 1 equiv. of B, 1 was already no longer observed after 5 min. The expected kinetic isotope effect associated with the deprotonation of 1 could be evidenced by using 1-D, leading to a 5 : 95 mixture of 7 and 1 after 5 min using 0.33 equiv. of B. Lastly, the 37 : 63 mixture obtained after 1 h using 1 and 0.33 equiv. of B was heated to 50 °C for 24 h. Although the amount of 7 remained almost the same (34% vs. 37%), the NMR yield of 1 decreased to 29%, while 38% of the cycloisomerization product 3 was formed (see ESI†). Therefore, this experiment suggests that the gold acetylides derived from enynes are actually not cyclization intermediates. To gain further insights, some gold acetylides were prepared and their reactivity studied by NMR experiments. A first follow-up confirmed the formation of new complexes when the gold acetylides were mixed with a stoichiometric amount of A at RT in CDCl₃ (Scheme 6, eqn (1)).

Scheme 6 The reactivity of gold acetylide derivatives.

According to the ¹H, ¹³C and ³¹P NMR chemical shifts, the expected dinuclear complexes 1-Au₂2NTf₂2 and 2-Au₂2NTf₂2 were actually formed. However, when 1-Au and 2-Au were submitted to catalysts A or B (5 or 100 mol%) under refluxing conditions, no cycloisomerization product was observed, even after 50 h (eqn (2)). Thus, it seems unlikely that gold acetylides themselves may be involved in a catalytic cyclizing pathway. However, based on the mechanism proposed by Houk and Toste for 1,5-allenynes (Scheme 2),⁶ it can be argued that in order to have an effective catalytic cycle, a proton source is required, e.g. the enyne itself. Thus, when a mixture of 1 and 8-D (or 2 and 8-D) was engaged, no scrambling of the deuterium was observed (Scheme 7). So, it is most likely that deprotonation does not even take place in these reactions.

Scheme 7 The deuterium scrambling experiment.

Once we established that the starting enyne cannot act as a proton donor, we wanted to compare in solution the reactivity of the enyneversus its corresponding acetylide. Thus, we ran the following set of experiments with both substrates 1 and 2 (Table 2). Enyne 1 (or 2) was first mixed with 5 mol% of 1-Au (or 2-Au), and exposed to 5 mol% of A. Curiously, this resulted in the formation of 3 (or 5) as traces (entries 1 and 3), as if the catalysis process was inhibited. So, we decided to increase the quantity of catalyst and, very interestingly, when 5 mol% of A was further added, a clean and rapid formation of product 3 (or 5) was observed (entries 2 and 4). This series of experiments demonstrates the greater affinity of A for 1-Au (or 2-Au) than for the free enyne and further discredits the intermediacy of gold acetylides, at least at RT.

Table 2 Cross experiments

Entry	Substrate	Gold acetylide	Solvent	A (mol%)	% yield
1	1	1-Au	CH2Cl2	5	trace
2	1	1-Au	CH2Cl2	10	70
3	2	2-Au	CDCl3	5	trace
4	2	2-Au	CDCl3	10	70

---

Based on some findings from a related study²⁰ and recent literature reports by Nolan,²¹ who showed that diaurated hydroxide (IPrAu)₂OH⁺ can serve as a reservoir of IPrAu⁺ or as the active species itself in some gold-catalyzed reactions, we wondered whether (LAu)₂OH⁺ could be involved in the formation of diaurated enyne.

Preliminary confirmation came when enyne 1 was introduced in ESI with catalyst C. An ion at m/z 1189 [1 + ((biphenyl-2-yl)(t-Bu)₂PAu)₂OH]⁺ was also observed (see the supporting information, Fig. S1†). This ion was shifted to m/z 1190 using 1-D instead of 1, whereas the m/z 1171 ion [1–H(D) + ((biphenyl-2-yl)(t-Bu)₂PAu)₂]⁺ was unchanged. Although the formation of the (IPrAu)₂OH⁺ complex can be controlled by “proton-catalyzed” hydrolysis of IPrAu⁺, it could arise in our case from adventitious water and a proton source when catalyst A or C is used, or just with a proton source with catalyst B. Assuming that catalysts A and B contain traces of acid, a mixture of enyne 1 and catalyst B in CH₃CN (instead of CDCl₃ to allow water miscibility) was re-injected into the mass spectrometer. Using this inadequate solvent for cycloisomerizations,²² the peak at m/z 1099 ([1–H + (Ph₃PAu)₂]⁺) was now observed with a low relative abundance. Nevertheless, under these conditions, the addition of 5% of water increased the intensity of the m/z 1099 ion by a factor of 10 (see supporting information, Fig. S6†). Moreover, a new peak at m/z 1117, potentially corresponding to [1 + (Ph₃PAu)₂OH]⁺ was now weakly observed. Conversely, when enyne 1 was treated with 0.33 equiv. of B in strictly anhydrous CDCl₃, formation of 7 appeared to be slower. After 5 min, the 7 : 1 ratio was 23 : 77, compared to 34 : 66 (Scheme 5).

Calculations were carried out to shed more light on that matter (Scheme 8). While the formation of (H₃PAu)₂OH⁺ by hydration of H₃PAu⁺ would be appreciably exothermic (eqn (1)), it would be quite endothermic when starting from (H₃PAu)₃O⁺ (eqn (2)). On the other hand, under acidic conditions, a strongly exothermic process may take place (eqn (3)).

Scheme 8 Calculated free energies for the formation of (H₃PAu)₂OH⁺.

We next compared the various routes to the diaurated gold species, starting either from H₃PAu⁺, (H₃PAu)₃O⁺, or (H₃PAu)₂OH⁺ (Scheme 9). With the former, water was chosen as a base (eqn (1)). While the complexation of H₃PAu⁺ is strongly exothermic, the formation of the gold acetylide is hampered by a strong endothermicity. Nevertheless, the experimentally observed affinity of H₃PAu⁺ for the gold acetylide was well reproduced computationally and the overall process is exothermic by 35.1 kcal mol⁻¹. Starting from (H₃PAu)₃O⁺, although the liberation of H₃PAu⁺ to complex the triple bond seems difficult to achieve (endothermic by 42.3 kcal mol⁻¹), the formation of the gold acetylide is only slightly endothermic (eqn (2)). However, the acetylide is found to be more stable by 0.4 kcal mol⁻¹ than the corresponding diaurated species. These results are very close to those reported by Toste and Houk in the 1,5-allenyne series,⁶ including the fact that the dinuclear complex is not the ground state in the gas phase. From (H₃PAu)₂OH⁺, both the dissociation/triple bond complexation and acetylide formation are quite endothermic, yet the formation of the diaurated species is exothermic by 10.3 kcal mol⁻¹ (eqn (3)). These computations suggest that when an oxonium catalyst such as B is used, the formation of dinuclear species from enynes or allenynes could in fact involve (Ph₃PAu)₂OH⁺. This feature does not seem to be a requirement with a type Acatalyst.

Scheme 9 Computational prediction of the stability of the gold complexes of an enyne (free energies relative to 0.0 in kcal mol⁻¹).

In conclusion, our study confirms the very high affinity of gold acetylides for a second complexation of gold. Both types of species appear to be non-reactive in the 1,6-enyne cycloisomerization catalytic cycle. We are now working on allenyne substrates to gain more insight into the reactivity of the corresponding gold acetylide intermediates. In addition, we have observed the presence of complexes involving organic substrates with a diaurated gold hydroxide generated from the catalyst under certain conditions, which seems to confirm the role of this recently described active form of gold.

Acknowledgements

The authors thank UPMC, CNRS, ANR BLANC Allenes, Ministère de l'Enseignement Supérieur et de la Recherche (Ph. D. grant for A. S.) and IUF (MM and LF) for financial support. Calculations were performed at the CRIHAN, project 2006-13.

Notes and references

1. For some selected reviews, see: (a) Z. Li, C. Brouwer and C. He, Chem. Rev., 2008, 108, 3239; (b) A. Arcadi, Chem. Rev., 2008, 108, 3266; (c) D. J. Gorin, B. D. Sherry and F. D. Toste, Chem. Rev., 2008, 108, 3351; (d) A. S. K. Hashmi, Chem. Rev., 2007, 107, 3180; (e) D. J. Gorin and F. D. Toste, Nature, 2007, 446, 395; (f) A. Fürstner and P. W. Davies, Angew. Chem., Int. Ed., 2007, 46, 3410; (g) C. H. M. Amijs, C. Ferrer and A. M. Echavarren, Chem. Commun., 2007, 698; (h) E. Jiménez-Núñez and A. M. Echavarren, Chem. Commun., 2007, 333.
2. One of the most emblematic issues perhaps being the cation vs.carbene rendition of gold species: (a) G. Seidel, R. Mynott and A. Fürstner, Angew. Chem., Int. Ed., 2009, 48, 2510; (b) A. M. Echavarren, Nat. Chem., 2009, 1, 431; (c) D. Benitez, N. D. Shapiro, E. Tkatchouk, Y. Wang, W. A. Goddard III and F. D. Toste, Nat. Chem., 2009, 1, 482.
3. (a) A. S. K. Hashmi, Angew. Chem., Int. Ed., 2010, 49, 5232; (b) R. L. LaLonde, W. E. Brenzovich, Jr., D. Benitez, E. Tkatchouk, K. Kelley, W. A. Goddard III and F. D. Toste, Chem. Sci., 2010, 1, 226.
4. (a) M. R. Gagné and D. Weber, Org. Lett., 2009, 11, 4962; (b) G. Seidel, C. W. Lehmann and A. Fürstner, Angew. Chem., Int. Ed., 2010, 49, 8466 . For a review, see; (c) H. Schmidbaur and A. Schier, Organometallics, 2010, 29, 2.
5. (a) T. N. Hooper, M. Green and C. A. Russel, Chem. Commun., 2010, 46, 2313 . For a recent report on the reactivity of gold acetylides, see; (b) M. C. Blanco, J. Camara, M. C. Gimeno, P. G. Jones, A. Laguna, J. M. Lopez-de-Luzuriaga, M. E. Olmos, M. D. Villacampa, Organometallics DOI:10.1021/om200397t.
6. P. H.-Y. Cheong, P. Morganelli, M. R. Luzung, K. N. Houk and F. D. Toste, J. Am. Chem. Soc., 2008, 130, 4517.
7. Y. Odabachian, X.-F. Le Goff and F. Gagosz, Chem.–Eur. J., 2009, 15, 8966.
8. For reviews, see: (a) V. Michelet, P. Y. Toullec and J.-P. Genêt, Angew. Chem., Int. Ed., 2008, 47, 4268; (b) E. Jiménez-Núñez and A. M. Echavarren, Chem. Rev., 2008, 108, 3326 Allenynes substrates, see: C. Aubert, L. Fensterbank, P. Garcia, M. Malacria and A. Simonneau, Chem. Rev., 2011, 111, 1954.
9. For relevant reports, see: (a) C. Nieto-Oberhuber, S. Lopez, M. P. Muñoz, D. J. Cárdenas, E. Buñuel, C. Nevado and A. M. Echavarren, Angew. Chem., Int. Ed., 2005, 44, 6146; (b) C. Nieto-Oberhuber, M. P. Muñoz, S. López, E. Jiménez-Núñez, C. Nevado, E. Herrero-Gómez, M. Raducan and A. M. Echavarren, Chem.–Eur. J., 2006, 12, 1677.
10. S. Baumgarten, D. Lesage, V. Gandon, J.-P. Goddard, M. Malacria, J.-C. Tabet, Y. Gimbert and L. Fensterbank, ChemCatChem, 2009, 1, 138.
11. N. Mézailles, L. Ricard and F. Gagosz, Org. Lett., 2005, 7, 4133.
12. (a) A. N. Nesmeyanov, E. G. Perevalova, Y. T. Struchkov, M. Y. Antipin, K. I. Grandberg and V. P. Dyadchenko, J. Organomet. Chem., 1980, 201, 343; (b) B. D. Sherry and F. D. Toste, J. Am. Chem. Soc., 2004, 126, 15978.
13. (a) C. Nieto-Oberhuber, S. Lopez and A. M. Echavarren, J. Am. Chem. Soc., 2005, 127, 6178; (b) P. Perez-Galan, N. Delpont, E. Herrero-Gomez, F. Maseras and A. M. Echavarren, Chem.–Eur. J., 2010, 16, 5324.
14. N. Cabello, E. Jiménez-Núñez, E. Buñuel, D. J. Cárdenas and A. M. Echavarren, Eur. J. Org. Chem., 2007, 4217.
15. R. Thota, D. Lesage, Y. Gimbert, L. Giordano, S. Humbel, A. Milet, G. Buono and J.-C. Tabet, Organometallics, 2009, 28, 2735.
16. It is worthy of note that, when using 1–D, an analog of 1 bearing a deuterium atom at the triple bond, with B, m/z 1099 ion was also observed.
17. The structure of 1-Au was ascertained by means of a X-ray diffraction study (CCDC 794624†).
18. For a recent report on the preparation of gold acetylides, see: G. C. Fortman, A. Poater, J. W. Levell, S. Gaillard, A. M. Z. Slawin, I. D. W. Samuel, L. Cavallo and S. P. Nolan, Dalton Trans., 2010, 39, 10382.
19. Interestingly, a molecular ion is absent in the ESI mass spectrum of 1-Au alone, which rather exhibits m/z 1099. While this may question the fact that A or B actually transfer Ph₃PAu⁺ to 1-Au in the ESI source, the results obtained with C or AgSbF₆ leave no doubt about the proposed intermolecular process. When the latter two are used, m/z 1099 is still observed, but weakly.
20. A. Simonneau, F. Jaroschik, D. Lesage, M. Karanik, R. Guillot, M. Malacria, J.-C. Tabet, J.-P. Goddard, L. Fensterbank, V. Gandon, Y. Gimbert to be published.
21. For mono- and diaurated gold hydroxides, see: (a) S. Gaillard, A. M. Z. Slawin and S. P. Nolan, Chem. Commun., 2010, 46, 2742; (b) S. Gaillard, P. Nun, A. M. Z. Slawin and S. P. Nolan, Organometallics, 2010, 29, 5402; (c) S. Gaillard, R. S. Ramón, P. Nun, A. M. Z. Slawin and S. P. Nolan, Chem.–Eur. J., 2010, 16, 13729; (d) N. Marion, P. Carlqvist, R. Gealageas, P. de Frémont, F. Maseras and S. P. Nolan, Chem.–Eur. J., 2007, 13, 6437; (e) R. S. Ramón, S. Gaillard, A. Poater, L. Cavallo, A. M. Z. Slawin and S. P. Nolan, Chem.–Eur. J., 2011, 17, 1238.
22. The use of acetonitrile instead of CDCl₃ results in very low conversion of 1 or 2 into 3 or 5. This is presumably due to the coordination of the solvent to gold.

---

Footnote

† Electronic supplementary information (ESI) available: Procedure for the synthesis of gold acetylides, preparation of the diaurated complexes, NMR and mass studies, computational details. CCDC reference number 794624. For ESI and crystallographic data in CIF or other electronic format see DOI: 10.1039/c1sc00478f

---