When cobalt-mediated [2 + 2 + 2] cycloaddition reaction dares go astray: synthesis of unprecedented cobalt(III)-complexes

Marion Delormea, Alexander Puntera, Raquel Oliveirab, Corinne Aubertc, Yannick Carissana, Jean-Luc Parraina, Muriel Amatore*a, Paola Nava*a and Laurent Commeiras*a
aAix Marseille Univ, CNRS, Centrale Marseille, iSm2, Marseille, France. E-mail: laurent.commeiras@univ-amu.f; paola.nava@univ-amu.fr; muriel.amatore@univ-amu.fr
bPASTEUR, Département de chimie, Ecole Normale Supérieure, PSL University, Sorbonne Université, CNRS, 75005 Paris, France
cSorbonne Universités, UPMC Univ Paris 06, Institut Parisien de Chimie Moléculaire, UMR CNRS 8232, Case 229, 4 Place Jussieu, 75252 Paris Cedex 05, France

Received 14th August 2019 , Accepted 16th September 2019

First published on 17th September 2019

In sharp contrast with the standard [2 + 2 + 2] cycloaddition reaction of diyne/ene, cobalt-mediated cycloadditions with γ-alkylidenebutenolide led to unprecedented cobalt(III) polycyclic complexes. A plausible mechanism supported by a computational study based on an unusual fragmentation of the butenolide moiety was postulated to account for this original reaction.

The transition metal-catalyzed [2 + 2 + 2] cycloaddition reaction is well established as one of the most elegant methods for the rapid construction of functionalised aromatic and heteroaromatic polycyclic compounds, as well as 1,3-cyclohexadiene derivatives. Since the original report of a metal-mediated [2 + 2 + 2] cycloaddition reaction of acetylene derivatives,1 many groups have demonstrated the significance of this reaction through the use of various transition metal complexes, mainly from group 8–10 elements (Ir, Ru, Ni, Fe, Co, Rh, Nb the most employed).2 Among all the available catalysts, CpCoL2 complexes (L = CO, PR3, alkenes) have been frequently employed as an active source of “CpCo”. Such systems have proved their efficiency (i) for the co-cyclisation of alkynes with alkenes to give cyclohexadiene products 1 in their complexed form,2a (ii) for the linear 2[thin space (1/6-em)]:[thin space (1/6-em)]1 co-oligomerisation3 of alkynes with acyclic enol ethers (R3 = alkoxyl) to give 1-alkoxy-1,3,5-triene derivatives 2 (Scheme 1a). It is worth noting that the use of cyclic enol ethers such as 2,3-dihydrofuran produced unconjugated 1,3,6-trienes3 whereas benzofuran derivatives led to the formation of the corresponding CpCo-complexed cyclohexadienes.4 Herein, we report the development of a cobalt-mediated polycyclisation reaction with the original enediynes 3 bearing an alkylidenebutenolide moiety. Compared to “classic” alkenes as well as “classic” enol ethers, alkylidenebutenolides display a unique exo-cyclic double bond.
image file: c9dt03311d-s1.tif
Scheme 1 Outcomes of cobalt(I)-mediated process.

This double bond, which was never been used as 2π partner in intra- or intermolecular [2 + 2 + 2] cycloaddition reactions, could have two distinct electronic properties due to the cyclic enol acetate and γ,δ-unsaturated lactone moieties. This specificity would account for the unusual reactivity (Scheme 1b and c). The main features of the present work include (i) formal (2 + 2 + 3) cycloaddition reaction, (ii) unprecedented formation of cobalt(III) complexes 4 featuring both π-allyl and carbanion ligands (highlighted in green in Scheme 1c) and (iii) high levels of regio- and diastereoselectivities. These cobalt(III) complexes rearranged in acidic conditions to give bicyclo[3.3.1]non-3-en-2-one scaffolds found in polyprenylated acylphloroglucinol natural products. Furthermore, a mechanism pathway, supported by a computational approach, based on unusual fragmentation of the butenolide moiety was also postulated.

The study set out with the ene-1,6-diyne model 3a (Scheme 2). This benchmark substrate, featuring two gem-bis-ethylester tethers, was prepared from the known triyne 5a,5 via a palladium-free Sonogashira coupling/oxacyclisation sequence developed in our group.6 When diyne 3a was exposed to a stoichiometric amount of CpCo(CO)2 in toluene at 110 °C under irradiation, we did not observe the formation of the expected γ-spiro-γ-lactone 1a′ or 1,3,5-triene 2a′. Interestingly however, an unforeseen tetracyclic cobalt complex 4a has been isolated in 71% yield, as the sole diastereomer. Satisfyingly, the structure of 4a was established without ambiguity thanks to an X-ray diffraction analysis.7 This cobalt(III) complex exhibits a bicyclo[3.2.2]nonane-6,8-dione core, in which the cobalt center is bound to the cyclopentadienyl anion, to the π-allyl moiety and to the carbanion in alpha position of a ketone. NMR (1H, 13C and 59Co) and UV-Vis spectroscopies as well as electrochemical studies complete the characterisation of complex 4a (see the ESI). Although the formation of the two five-membered rings could be expected in such a process, the conversion of the γ-alkylidenebutenolide backbone to a cyclohexa-1,3-dione pattern is unusual. This surprising result encouraged us to evaluate the potential influence of other cobalt sources, solvent or temperature. One other air-stable cobalt complex, activated only by heating (without irradiation) was tested. In these conditions, the use of CpCo(CO)dmfu8 (1 equiv. of complex in the toluene at 110 °C for 16 hours) gave rise to 4a in 42% yield and with a conversion rate of 43%. An additional irradiation allowed an increase in the conversion rate (80%) as well the yield (54%). Because the CpCo(CO)2 complex with a thermal (110 °C in toluene) and photochemical activation (CO dissociation) gave better results for obtaining the polycyclic complex 4a, two other attempts in which we varied the nature of solvent were next investigated. Firstly, in o-xylene at 140 °C under irradiation, the product 4a was isolated in 28% yield. Many colored residues were obtained allowing us to evoke possible and important product degradation under these conditions. In addition, the activation of the CpCo(CO)2 complex without irradiation is made possible by heating the mixture to 170 °C in n-decane. Unfortunately, only traces of the expected product 4a were observed, despite the total disappearance of the starting material. These two results further support that the starting material and/or the product could be degraded at high temperature. Finally, the best conditions for the synthesis of 4a request a stoichiometric amount of CpCo(CO)2 in toluene at 110 °C under irradiation. With optimal conditions in hand, we further explored the versatility of this reaction for the synthesis of some tetracyclic cobalt(III) complexes by varying the nature and length of the two tethers (X and Y) along with the nature of the R1 and R2 substituents on the alkyne moiety and the γ-alkylidenebutenolide backbone, respectively (Scheme 3).

image file: c9dt03311d-s2.tif
Scheme 2 Cobalt(I)-mediated synthesis and crystallographically-derived molecular structure of polycyclic cobalt(III) complex 4a.

image file: c9dt03311d-s3.tif
Scheme 3 Cobalt(I)-mediated synthesis of polycyclic cobalt(III) complexes.

The nature of the tether Y was investigated first. In general, the desired products were obtained with similar efficiency and satisfactory yields. When Y = NTs, O or CH2, the complexes 4b,9 4c and 4d[thin space (1/6-em)]10 were obtained at a yield of 55, 66 and 55%, respectively. Not surprisingly, when the reaction was conducted with an enediyne 5e possessing an all carbon six-membered tether (Y = CH2CH2), the conversion rate was found to be lower (40%) even after a prolonged reaction time (24 h) and complex 4e was isolated in a 29% yield (72% based on recovery starting material, brsm). We then turned our attention to the γ-alkylidenebutenolide pattern. Several substituents, such as hydrogen, n-butyl and CH2OMe, are well tolerated, giving the desired cobalt(III) complexes 4f–h in satisfactory yields. When R2 is a phenyl group, and despite the complete consumption of the starting material, only products resulting from degradation were detected in the crude reaction mixture. In addition, changing the methyl group to a hydrogen atom on the alkyne part (R1 = H) did not affect the efficiency of the reaction. To complete our study, we altered both the nature of the Y tether and R2 group. Remarkably, the reaction proceeded smoothly giving the corresponding diketones 4j–k.

We then investigated the potential of cobalt(III) complexes towards demetallation (Scheme 4). While reductive demetallation (Zn in DCE) was unsuccessful, protodemetallation reaction of 4a (HCl 4 M in dioxane) gave a single product 6a arising from a rearrangement of the bicyclo[3.2.2]nonane-6,8-dione core to 5-hydroxybicyclo[3.3.1]non-3-en-2-one in 55% yield. Similar results were also obtained when X = NTs (complex 4b), giving polycyclic enone 6b that, to our delight, led to crystals suitable for X-ray diffraction studies.11 This rearrangement opens new opportunities to access bicyclo[3.3.1]non-3-en-2-one core of natural products from the family of polyprenylated acylphloroglucinol PPAP.12

image file: c9dt03311d-s4.tif
Scheme 4 Demetallation and rearrangement of the cobalt(III) complexes 4a,b.

A preliminary computational study allows us to propose a possible mechanism that accounts for the formation of the polycyclic Co(III) complexes from a simplified model of 4d, where X = C(CO2Me)2 (Fig. 1). It was previously shown that the two carbonyl ligands’ substitution by two acetylenic units on CpCo is dissociative, and leads to I.3 The initial dissociation of the CO ligand leads to a complex in its triplet state. This was experimentally13 and theoretically14 studied by Bengali et al. and by Siegbahn respectively. From here, an inter-system-crossing is necessary to reach I, which is in a singlet state, as extensively studied.3,15,16

image file: c9dt03311d-f1.tif
Fig. 1 Computed ΔG profile (T = 110 °C, kcal mol−1). Computational level: BP86-D3/def2-TZVP, solvent effects by the COSMO model.

Reported energies are Gibbs energies (T = 110 °C) relative to I. Structures were optimized and frequencies were computed at the BP86-D3/def2-TZVP level,17 following Kozlowski et al.18 Solvent effects were included by performing single-point energy calculations with the COSMO continuum solvent model.19 The first step, which requires an activation energy of 6.0 kcal mol−1 is the oxidative coupling of I to obtain II. The exo-cyclic double bond (C12–C13) binds to the cobalt metal center in II with a −24.2 kcal mol−1 stabilisation energy. Similar reactivities were already reported in an intermolecular fashion.3,15,16 In these reports, once the cobaltacyclopentadiene is formed, the triplet potential energy surface is reached. The following coordination of the 2π partner leads to a singlet intermediate, which would be analogous to II in our intramolecular approach. In both these inter- and intramolecular pathways, the mechanism proceeds from the coordinated singlet intermediate.20 The presence of the five-membered link that joins C8 to C12 imposes a constraint in the approach of the double bond to the metal, resulting in a low-energy demanding closure of the C8–C12 σ-bond to give III with a barrier of 4.2 kcal mol−1. From intermediate III, the next step is the formation of the C2–C14 bond, with a computed barrier of 19.4 kcal mol−1 to give the β-oxo-ketene intermediate IV, which is 10 kcal mol−1 more stable than III. The formation of the C2–C14 bond triggers a concomitant reorganization of the electronic system: the Co–C13 bond breaks and, more interestingly, so does the lactone moiety in III. It is worth noting that the formation of the C2–C13 bond, which would lead to the normally expected spiro-like product P1′, is kinetically slightly disfavored with a barrier computed at 20.8 kcal mol−1. Once IV is reached, the reaction proceeds to P4 with no further energetic obstacles, via a two-step mechanism. Firstly, a conformational change around the ketone is necessary, leading to V. This latter is slightly endothermic (by 5.2 kcal mol−1) and requires an activation Gibbs energy of 8.3 kcal mol−1. Secondly, a final transition state (with a barrier of 8.9 kcal mol−1) leads to the formation of both Co–C15 and C8–C16 bonds, giving the final product P4 in an irreversible manner. A concerted mechanism was also explored to connect directly II to IV, nevertheless the transition state was located at +1.4 kcal mol−1 with respect to I, which corresponds to a 25.6 kcal mol−1 barrier, higher than the ones associated to TsII–III and TsIII–IV.

In summary, the Co-mediated cycloaddition reactions of enediynes bearing a γ-alkylidenebutenolide provided unexpected outcomes. A Co(I)-mediated formal (2 + 2 + 3) cycloaddition was observed leading to original cobalt(III) complexes possessing a bicyclo[3.2.2]nonane-6,8-dione backbone. A plausible mechanism supported by a computational study based on an unusual fragmentation of the butenolide moiety was postulated to account for this original reaction.

Furthermore the cobalt(III) complexes rearranged in acidic conditions to give bicyclo[3.3.1]non-3-en-2-one scaffolds found in polyprenylated acylphloroglucinol natural products. The increasing complexity associated with a perfect diastereoselectivity makes this cobalt-mediated process particularly attractive and opens new possibilities. Experimental and computational studies on related Co-mediated cycloadditions are ongoing in our laboratories.

Conflicts of interest

There are no conflicts to declare.


M. D. thanks the French Research Ministry, Aix-Marseille University and CNRS for financial support. The authors wish to thank Dr Christian Amatore warmly for fruitful discussion about electrochemistry, as well as Drs Michel Giorgi and Gaëtan Herbette for X-Ray analysis and NMR studies, respectively.

Notes and references

  1. W. Reppe, O. Schichting, K. Klager and T. Toepel, Justus Liebigs Ann. Chem., 1948, 560, 1 CrossRef CAS.
  2. Selected references: (a) G. Domínguez and J. Pérez-Castells, Chem. – Eur. J., 2016, 22, 6720–6739 CrossRef PubMed; (b) A. Lledo, A. Pla-Quintana and A. Roglans, Chem. Soc. Rev., 2016, 45, 2010–2023 RSC; (c) Y. Sato and Y. Obora, Eur. J. Org. Chem., 2015, 5041–5054 CrossRef; (d) S. Okamoto and Y. Sugiyama, Synlett, 2013, 1044 CrossRef CAS; (e) Transition-Metal-Mediated Aromatic Ring Construction, ed. K. Tanaka, John Wiley & Sons, Hoboken, 2013 Search PubMed; (f) Y. Shibata and K. Tanaka, Synthesis, 2012, 323–350 CAS; (g) K. Tanaka, Heterocycles, 2012, 85, 1017–1043 CrossRef CAS; (h) D. L. J. Broere and E. Ruijter, Synthesis, 2012, 2639–2672 CAS; (i) G. Domínguez and J. Pérez-Castells, Chem. Soc. Rev., 2011, 40, 3430–3444 RSC; (j) M. R. Shaaban, R. El-Sayed and A. H. M. Elwahy, Tetrahedron, 2011, 67, 6095–6130 CrossRef CAS; (k) N. Wedding and M. Hapke, Chem. Soc. Rev., 2011, 40, 4525–4538 RSC; (l) D. Leboeuf, V. Gandon and M. Malacria, in Handbook of Cyclization Reactions, ed. S. Ma, Wiley-VCH, Weinheim, 2009 Search PubMed; (m) N. Agenet, V. Gandon, O. Buisine, F. Slowinski and M. Malacria, in Organic Reactions, ed. T. V. RajanBabu, Wiley, Hoboken, NJ, 2007 Search PubMed; (n) V. Gandon, C. Aubert and M. Malacria, Chem. Commun., 2006, 2209–2217 RSC.
  3. (a) D. Leboeuf, L. Iannazzo, A. Geny, M. Malacria, K. P. C. Vollhardt, C. Aubert and V. Gandon, Chem. – Eur. J., 2010, 16, 8904–8913 CrossRef CAS PubMed; (b) V. Gandon, N. Agenet, K. P. C. Vollhardt, M. Malacria and C. Aubert, J. Am. Chem. Soc., 2006, 128, 8509–8520 CrossRef CAS PubMed.
  4. (a) D. Pérez, B. A. Siesel, M. J. Malaska, E. David and K. P. C. Vollhardt, Synlett, 2000, 306–310 Search PubMed; (b) T. Aechtner, D. A. Barry, E. David, C. Ghellamallah, D. F. Harvey, A. de la Houpliere, M. Knopp, M. J. Malaska, D. Pérez, K. A. S. Chärer, B. A. Siesel, K. P. C. Vollhardt and R. Zitterbart, Synthesis, 2018, 50, 1053–1089 CrossRef CAS.
  5. P. Jungk, F. Fischer, I. Thiel and M. Hapke, J. Org. Chem., 2015, 80, 9781–9793 CrossRef CAS PubMed.
  6. (a) S. Inack-Ngi, R. Rahmani, L. Commeiras, G. Chouraqui, J. Thibonnet, A. Duchêne, M. Abarbri and J.-L. Parrain, Adv. Synth. Catal., 2009, 351, 779–788 CrossRef CAS; (b) S. Inack-Ngi, C. Khalil, V. Héran, L. Commeiras, J.-L. Parrain, A. Duchêne, M. Abarbri and J. Thibonnet, Chem. – Eur. J., 2011, 17, 13692–13696 CrossRef PubMed.
  7. CCDC 1853760 contains the supplementary crystallographic data for this paper. Even if one “ethyl” moiety was found to be disordered, the structure of 4a was established without ambiguity.
  8. A. Geny, N. Agenet, L. Iannazzo, M. Malacria, C. Aubert and V. Gandon, Angew. Chem., Int. Ed., 2009, 48, 1810–1813 CrossRef CAS PubMed.
  9. CCDC 1853762 contains the supplementary crystallographic data for this paper.
  10. CCDC 1853766 contains the supplementary crystallographic data for this paper.
  11. CCDC 1853778 contains the supplementary crystallographic data for this paper. Even if one “ethyl” moiety was found to be disordered, the structure of 6b was established without ambiguity.
  12. (a) C. Guttroff, A. Baykal, H. Wang, P. Popella, F. Kraus, N. Biber, S. Krauss, F. Götz and B. Plietker, Angew. Chem., Int. Ed., 2017, 56, 15852–15856 CrossRef CAS PubMed; (b) A. J. Grenning, J. H. Boyce and J. A. Porco, J. Am. Chem. Soc., 2014, 136, 1179–1804 CrossRef PubMed.
  13. A. A. Bengali, R. G. Bergman and C. B. Moore, J. Am. Chem. Soc., 1995, 117, 3879–3880 CrossRef CAS.
  14. P. E. M. Siegbahn, J. Am. Chem. Soc., 1996, 118, 1487–1496 CrossRef CAS.
  15. (a) A. A. Dahy and N. Koga, Bull. Chem. Soc. Jpn., 2005, 78, 781–791 CrossRef CAS; (b) A. A. Dahy, C. H. Suresh and N. Koga, Bull. Chem. Soc. Jpn., 2005, 78, 792–803 CrossRef CAS.
  16. (a) N. Agenet, V. Gandon, K. P. C. Vollhardt, M. Malacria and C. Aubert, J. Am. Chem. Soc., 2007, 129, 8860–8871 CrossRef CAS PubMed; (b) C. Aubert, V. Gandon, A. Geny, T. J. Heckrodt, M. Malacria, E. Paredes and K. P. C. Vollhardt, Chem. – Eur. J., 2007, 13, 7466–7478 CrossRef CAS PubMed.
  17. (a) P. A. M. Dirac, Proc. R. Soc. London, Ser. A, 1929, 123, 714–733 CrossRef CAS; (b) J. C. Slater, Phys. Rev., 1951, 81, 385–390 CrossRef CAS; (c) S. H. Vosko, L. Wilk and M. Nusair, Can. J. Phys., 1980, 58, 1200–1211 CrossRef CAS; (d) A. D. Becke, Phys. Rev. A, 1988, 38, 3098–3100 CrossRef CAS PubMed; (e) J. P. Perdew, Phys. Rev. B: Condens. Matter Mater. Phys., 1986, 33, 8822–8824 CrossRef PubMed. Erratum: Phys. Rev. B, 1986, 34, 7406. (f) S. Grimme, J. Antony, S. Ehrlich and H. Krieg, J. Chem. Phys., 2010, 132, 154104 CrossRef PubMed; (g) F. Weigend and R. Ahlrichs, Phys. Chem. Chem. Phys., 2005, 7, 3297–3305 RSC; (h) The RI-J technique was employed: K. Eichkorn, F. Weigend, O. Treutler and R. Ahlrichs, Theor. Chem. Acc., 1997, 97, 119–124 Search PubMed; (i) All Density Functional Theory calculations were performed using the TURBOMOLE program package. TURBOMOLE V7.2 2017, a development of University of Karlsruhe and Forschungszentrum Karlsruhe GmbH, 1989–2007, TURBOMOLE GmbH, since 2007; available from http://www.turbomole.com.turbo.
  18. P. M. Kozlowski, M. Kumar, P. Piecuch, W. Li, N. P. Bauman, J. A. Hansen, P. Lodowski and M. Jaworska, J. Chem. Theory Comput., 2012, 8, 1870–1894 CrossRef CAS PubMed.
  19. A. Klamt and G. Schüürmann, J. Chem. Soc., Perkin Trans. 2, 1992, 799–805 Search PubMed.
  20. As the novelty of this study occurs from III, we have not investigated yet the inter-system-crossings from I to II.


Electronic supplementary information (ESI) available: Synthesis procedures, 1H and 13C NMR spectra of each compounds and computational studies. CCDC 1853760, 1853762, 1853766 and 1853778. For ESI and crystallographic data in CIF or other electronic format see DOI: 10.1039/c9dt03311d

This journal is © The Royal Society of Chemistry 2019