Conformational control allows for [3,3]-sigmatropic rearrangements to proceed with torquoselectivity

Abstract

The Woodward–Hoffmann rules determine a first level of selectivity in pericyclic reactions. In this framework torqueselectivity can be defined as an additional degree of selectivity in electrocyclizations favouring one of the two allowed pathways. In this communication we extend this second level of selectivity beyond electrocyclizations. We have designed substrates that undergo a [3,3]-sigmatropic rearrangement in a torqueselective fashion. This behaviour is conformation dependent and is only observed when the reaction proceeds via a boat transition state. Structural design aimed at conformational control at the transition state, ensuring a boat-like structure, provides the first example of a [3,3]-sigmatropic shift that occurs with torqueselectivity.

---

Torquoselectivity is the term coined by Houk^1,2 to describe a second level of selectivity in the framework of electrocyclic ring opening and ring closing reactions where the first level would be the Woodward–Hoffmann rules.^3,4 This additional selection rule may be defined as follows: the selectivity for outward or inward rotation of the substituents of a breaking/forming C–C bond.

In electrocyclizations this second level of selectivity occurs both in disrotatory and conrotatory cyclizations, although torquoselectivity is usually stronger in the latter.^5–9 In the paradigmatic example of this effect Houk et al. demonstrated that, in the 4-π electron conrotatory ring-opening of substituted cyclobutenes, electron donating groups (EDG) at carbon 3 rotate away from the breaking bond (outwards) whereas electron withdrawing groups (EWG) rotate in the opposite direction (inwards, see Fig. 1).^10–14 This selection is explained as an electronic effect arising from the interaction of molecular orbitals localized at the substituent with the sigma orbitals of the bond being broken in the transition states.¹⁵

Fig. 1 Two conrotatory paths for the ring-opening of 3-substitued cyclobutenes. The energetic preference for electron donating groups (EDG) at C3 to rotate outward, rather than inward, increases with the π-donor nature of the substituent. In the same way the preference for electron withdrawing groups to rotate inward is related to their π-acceptor capability.

In our group we have widely studied the torquoselectivity in cationic, anionic and neutral electrocyclizations.^16–18 Furthermore, in a recent work, we have suggested it can be extended to diradical closing processes and therefore its scope can be larger than the classical pericyclic reactivity.¹⁹

Exploring cases of selectivity resembling torquoselectivity we found an example in a [1,5]-sigmatropic hydrogen shift of vinylallenes reported by Okamura.^20,21 and we performed a thorough computational study on its mechanism and the source of such selectivity.²² We concluded that this rearrangement resulted in stereodefined trienes only when the substituent was a sulfoxide or sulfone group through a preferred anti H migration. Taking into account these precedents we decided to further explore the realm of sigmatropic reactions looking for substrates that could feature a torquoselective transformation. For that purpose we firstly studied the 1,2,6-heptatriene Cope reaction, a classic pericyclic [3,3]-sigmatropic rearrangement. The reason for using 1,2,6-heptatriene as a substrate lies on the allene system. This fragment allows for an out-of-plane substituent that rotates in-plane along the Cope rearrangement, hence providing a rotating substituent that may impart the desired torquoselectivity (see Fig. 2). According to this idea we decided to replace a terminal hydrogen of the allene moiety in 1 by different groups. Although in [3,3]-sigmatropic rearrangements the chair transition state is energetically favored over the boat conformation some exceptions have been reported.^23,24 Since the allene moiety in the substrate could affect the preferred conformation for this reaction, both chair and boat transition states were considered in this study. Activation free energies for the Cope rearrangement of 1 featuring strong π-electron donor groups (OMe, OH, NH₂), soft electron donors with different steric demands (Me, ^tBu), a π donor and σ acceptor (F) and, finally, electron withdrawing groups (BH₂, CHO, NO₂, CN) with different electronic and steric demands have been computed.

Fig. 2 (a) 1,2,6-Heptatriene Cope rearrangement (b) substituted systems included in this study.

Table 1 summarizes the results obtained for the Cope rearrangement of 1_R. The activation energy values confirmed the energetic preference of the chair transition state over the boat one, in agreement with results for the classic [3,3]-sigmatropic rearrangement of 1,5-hexadiene. In terms of electronic effects of the substituents on the E/Z preference, no clear trend can be assigned to chair transition states, see column four where EDGs and EWGs can be found to favour either Z or E transition states in a seemingly uncorrelated way. These findings are consistent with the general view that selectivity in classic [3,3]-sigmatropic rearrangements, when occurring via a chair transition state, is dominated by steric effects, with bulky substituents occupying equatorial sites around the chair structure. Such clear steric effects are displayed for the H, Me, ^tBu series of entries (Table 1). Surprisingly, however, for the rearrangement operating through a boat-like transition state a clear effect can be observed when changes are applied to the electronic demand of the substituents. It seems unambiguous that EWG groups stabilize more TS2Z than TS2E, whereas the opposite effect is observed with strong electron donors. If the ^tBu group is also considered in the series (although its effects are steric in nature) a trend is clearly observed with a decreasing ΔΔG^‡_E–Z values in the following substituents ^tBu > NO₂ > CHO > BH₂ > Me > CN > NH₂ > H > F > OH > OMe. It should be noted that the strongest π-donor groups can even revert the steric bias and for OMe, OH and even F TS2Z is favoured. These results therefore suggest that torquoselectivity can be achieved in a [3,3]-sigmatropic reaction but it is precluded by the preferred chair-like conformation at the transition state. If a boat-like transition state can be imposed, a torquoselective process could be expected. Encouraged by these results we decided to design a substrate with potential torquoselectivity and whose [3,3]-sigmatropic rearrangement is constrained to occur via a boat conformation.

Table 1 M06-2X/Def2TZVPP Gibbs activation free energies (kcal mol⁻¹, 298.15 K) in solvent (acetonitrile) for the rearrangement of 1 into 3E and 3Z through a chair and a boat TS. The third column in each block represents the difference of activation energies for the formation of 3E and 3Z

R	Chair			Boat
	ΔG^‡E	ΔG^‡Z	ΔΔG^‡E–Z	ΔG^‡E	ΔG^‡Z	ΔΔG^‡E–Z
NO2	22.14	22.94	−0.8	43.99	39.77	4.2
CHO	29.83	29.23	0.6	45.88	42.14	3.8
BH2	29.36	31.72	−2.3	47.40	44.33	3.1
CN	27.53	27.41	0.1	41.79	41.45	0.3
H	34.48	34.48	0.0	44.10	44.10	0.0
Me	36.00	34.11	1.9	45.15	43.71	1.4
^tBu	41.11	35.70	5.4	51.26	45.00	6.3
NH2	33.45	32.28	1.2	42.28	42.11	0.2
F	30.27	30.69	−0.4	40.07	40.87	−0.8
OH	31.12	31.77	−0.6	40.74	41.64	−0.9
OMe	31.16	31.67	−0.5	40.28	41.39	−1.1

---

A simple example of a substrate that would undergo a [3,3]-sigmatropic rearrangement via a boat transition state is 1,4-dimethyliden cyclohexane. In this system the boat conformation is accesible with only 4 kcal mol⁻¹ energy difference with respect to the chair conformation. However, this is a substrate that features mirror plane symmetry and, to achieve torquoselectivity, the branches have to be discernible. We therefore decided to perform this study with an oxo derivative trying to incorporate two favourable effects in a single modification: the discernible rotating directions required for torquoselectivity and reduced activation barriers due to the formation of a carbonyl group in the course of the sigmatropic reaction (hence a Claisen type of rearrangement).²⁵ Substrate 4 was chosen as a candidate to undergo a sigmatropic rearrangement in a torquoselective fashion (see Fig. 3).

Fig. 3 (a) Torquoselective model for a Claisen rearrangement (b) substituted systems included in this study.

The results of the Claisen rearrangement for 4_R are collected in Table 2. The activation energy for the unsubstituted system is 36.2 kcal mol⁻¹ and the corresponding values for all the related reactions cluster around this value (the activation energies range between 36 and 41 kcal mol⁻¹, approximately). Again, the highest barrier 41.25 kcal mol⁻¹ is found for the tert-butyl group. Comparing the entries for tert-butyl and methyl it seems obvious that this is due to steric effects, and that these effects are more intense in the E transition state than in the Z one (41.2 vs. 36.7 and 36.6 vs. 36.3 kcal mol⁻¹, respectively). A more detailed analysis reveals an interplay between steric and electronic factors. This causes different activation energies depending on the substituent characteristics. As in the initial substrate 1_R, when the reaction occurs through a boat transition state, it seems unambiguous that electron withdrawing groups stabilize more TS5_RZ than TS5_RE, whereas the effect is reversed with strong electron donors. A series of decreasing ΔΔG^‡_E–Z values can be established as ^tBu > BH₂ ∼ CHO > NO₂ > Me > H > CN ∼ NH₂ > OH ∼ F > OMe. This trend is similar to that found for the Cope rearrangement of 1_R via a boat transition state, although the effects are somewhat damped on the electron withdrawing group side (more than 3 kcal mol⁻¹ between torquodivergent pathways for strong acceptors in 1_R versus differences between 1 and 2 kcal mol⁻¹ for the Claisen rearrangement of 4_R). The nitrile group has an unexpected behaviour in this reaction, yielding energy differences for the E and Z pathways that are compatible with mild election donors. We explored the orbitals involved in this rearrangement for an explanation to this anomalous behaviour of the CN derivative but our efforts were unsuccessful. We finally pondered that the energy difference is very small (−0.4 kcal mol⁻¹) and that it may be related with its very low steric demand and linear geometry.

Table 2 M06-2X/Def2TZVPP Gibbs free activation energies (kcal mol⁻¹, 298.15 K) in solvent (acetonitrile) for the two sigmatropic rearrangement alternatives of substrate 4_R

R	ΔG^‡E	ΔG^‡Z	ΔΔG^‡E–Z
NO2	38.73	37.80	0.9
CHO	38.00	36.20	1.7
BH2	37.10	35.32	1.7
CN	36.34	36.71	−0.4
H	36.16	36.16	0.0
Me	36.70	36.30	0.4
^tBu	41.25	36.64	4.6
NH2	37.00	37.36	−0.4
F	37.33	38.30	−1.0
OH	36.70	37.70	−1.0
OMe	36.70	38.40	−1.6

---

In order to provide an explanation for electronic origins of the observed selectivity we performed Natural Bond Orbital (NBO) analysis on the transition states of the Claisen rearrangement.^26–28 A clear orbital interaction as the one reported for 4-π-electron electrocyclizations¹⁵ could not be identified. The small energy difference observed between the two possible processes is consistent with the lack of a clear orbital effect. The substituent exerting the torqueselective effect in these substrates is not directly attached to the carbon atom whose bond is being broken as in electrocyclizations. This substituent is one bond away and therefore its interactions with the forming bond are not as strong. As a result there is no single interaction responsible for the selectivity, but a combined effect of different smaller interactions. Among other interactions found within the π orbital array a small set of orbital interactions seemed to correlate well with the observed selectivity. This set is composed of the π orbitals involved in the breaking/forming σ C–C bond: πC₂–C₈ → π*C₅–C₇ and πC₅–C₇ → π*C₂–C₈ (see Fig. 3 and 4). The expected strong charge donation between the previous π orbitals was found to be dependent on the electronic nature of the substituents. With π-electron donating groups the combined effect of these interactions is approximately 2 kcal mol⁻¹ (reaching up to 5 kcal mol⁻¹ for the fluoro derivative) stabilizing the E path with respect to the Z alternative. On the contrary, for electron withdrawing groups the energies associated to these interactions favor the Z process (with values between 3 and 7 kcal mol⁻¹, see the ESI†).

Fig. 4 Natural bond orbitals involved in the stabilizing interactions selected for the E isomer with an electron donor substituent (TS5_OMeE): πC₂–C₈ → π*C₅–C₇ (left) and πC₅–C₇ → π*C₂–C₈ (right).

In summary, these findings provide strong evidence for torquoselectivity in [3,3]-sigmatropic rearrangements, an effect that has not been reported up to date for this kind of reactions. This selectivity has been explained on the basis of a superposition of steric and electronic effects (the second ones dominate for strong donors and acceptors of moderate size whereas the first dominates in bulky substituents). Therefore beyond the Woodward–Hoffmann rules, this second level of selectivity can be extended to other pericyclic reactions, namely sigmatropic rearrangements. Exploration of other systems is being undertaken with the target of reducing the activation energies so that experimental proof of torquoselectivity can be collected.

During this work the Kohn–Sham formulation of the density functional theory was employed.^29,30 The meta-hybrid exchange–correlation functional, M06-2X, by Zhao and Truhlar³¹ was used with the triple-ζ quality Def2TZVPP basis set for the all the calculations. All geometry optimizations have been carried out using tight convergence criteria in order to obtain accurate stationary points. Such accuracy in the geometries also required a pruned grid for numerical integration with 99 radial shells and 590 angular points per shell. Analysis of the normal modes obtained via diagonalization of the Hessian matrix was used to confirm the topological nature of each stationary point. The wavefunction stability for each optimized structure has also been checked³² for all the computations included in this work. Solvation effects have been taken into account variationally throughout the optimization procedures via the polarizable continuum model (PCM) and the smooth switching function by York and Karplus.^33,34 All the calculations were performed with the Gaussian09 package.³⁵

Acknowledgements

The authors thank the Centro de Supercomputación de Galicia (CESGA) for time on HPC infraestructures. Ministerio de Economía y Competitividad (MINECO, CTQ2013-48937-C2-2P) and the Xunta de Galicia (EM2014/040, GPC2014/066) are also acknowledged for financial support.

References

1. K. N. Houk, Stereoselective electrocyclizations and sigmatropic shifts of strained rings: Torquoelectronics, in Strain and its implications in organic chemistry, Kluwer Academic Publishers, Dordrecht, 1989, pp. 25–37.
2. A. Buda, Y. Wang and K. Houk, J. Org. Chem., 1989, 54, 2264–2266.
3. R. B. Woodward and R. Hoffmann, J. Am. Chem. Soc., 1965, 87, 395–397.
4. R. B. Woodward and R. Hoffmann, Angew. Chem., Int. Ed., 1969, 8, 781–853.
5. S. J. Cristol, R. M. Sequeira and C. H. DePuy, J. Am. Chem. Soc., 1965, 87, 4007–4008.
6. C. H. DePuy, Acc. Chem. Res., 1968, 1, 33–41.
7. K. N. Houk, D. C. Spellmeyer, C. W. Jefford, C. G. Rimbault, Y. Wang and R. D. Miller, J. Org. Chem., 1988, 53, 2125–2127.
8. O. N. Faza, C. S. López, R. Álvarez and A. R. de Lera, Org. Lett., 2004, 6, 905–908.
9. O. Nieto Faza, C. Silva López, R. Álvarez and A. R. de Lera, J. Org. Chem., 2004, 69, 9002–9010.
10. W. R. Dolbier, H. Koroniak, D. J. Burton, A. R. Bailey, G. S. Shaw and S. W. Hansen, J. Am. Chem. Soc., 1984, 106, 1871–1872.
11. W. Kirmse, N. G. Rondan and K. N. Houk, J. Am. Chem. Soc., 1984, 106, 7989–7991.
12. N. G. Rondan and K. N. Houk, J. Am. Chem. Soc., 1985, 107, 2099–2111.
13. S. Niwayama, E. A. Kallel, D. C. Spellmeyer, C. Sheu and K. N. Houk, J. Org. Chem., 1996, 61, 2813–2825.
14. W. R. Dolbier Jr, H. Koroniak, K. N. Houk and C. Sheu, Acc. Chem. Res., 1996, 29, 471–477.
15. C. W. Jefford, G. Bernardinelli, Y. Wang, D. C. Spellmeyer, A. Buda and K. N. Houk, J. Am. Chem. Soc., 1992, 114, 1157–1165.
16. C. S. López, O. N. Faza, D. M. York and A. de Lera, J. Org. Chem., 2004, 69, 3635–3644.
17. O. N. Faza, C. S. Lopez, A. B. Gonzalez-Perez, M. Perez-Rodriguez and A. R. de Lera, J. Phys. Org. Chem., 2009, 22, 378–385.
18. O. Nieto Faza, C. Silva López, R. Álvarez and A. R. de Lera, Chem.–Eur. J., 2009, 15, 1944–1956.
19. R. Villar López, O. Nieto Faza and C. Silva López, RSC Adv., 2015, 5, 30405–30408.
20. W. Okamura, G.-Y. Shen and R. Tapia, J. Am. Chem. Soc., 1986, 108, 5018–5019.
21. G.-Y. Shen, R. Tapia and W. Okamura, J. Am. Chem. Soc., 1987, 109, 7499–7506.
22. O. Faza, C. López and A. De Lera, J. Org. Chem., 2007, 72, 2617–2624.
23. C. Chen, K. Namba and Y. Kishi, Org. Lett., 2009, 11, 409–412.
24. S. Gül, F. Schoenebeck, V. Aviyente and K. Houk, J. Org. Chem., 2010, 75, 2115–2118.
25. O. Wiest, K. Black and K. Houk, J. Am. Chem. Soc., 1994, 116, 10336–10337.
26. A. E. Reed, L. A. Curtiss and F. Weinhold, Chem. Rev., 1988, 88, 899–926.
27. E. D. Glendening and F. Weinhold, J. Comput. Chem., 1998, 19, 593–609.
28. F. Weinhold, J. Comput. Chem., 2012, 33, 2363–2379.
29. P. Hohenberg and W. Kohn, Phys. Rev., 1964, 136, B864–B871.
30. W. Kohn and L. Sham, Phys. Rev. A, 1965, 140, A1133–A1138.
31. Y. Zhao and D. G. Truhlar, Theor. Chem. Acc., 2008, 120, 215–241.
32. R. Bauernschmitt and R. Ahlrichs, J. Chem. Phys., 1996, 104, 9047–9052.
33. J. Tomasi, B. Mennucci and R. Cammi, Chem. Rev., 2005, 105, 2999–3093.
34. D. M. York and M. Karplus, J. Phys. Chem. A, 1999, 103, 11060–11079.
35. M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, G. Scalmani, V. Barone, B. Mennucci, G. A. Petersson, H. Nakatsuji, M. Caricato, X. Li, H. P. Hratchian, A. F. Izmaylov, J. Bloino, G. Zheng, J. L. Sonnenberg, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, T. Vreven, J. A. Montgomery Jr, J. E. Peralta, F. Ogliaro, M. Bearpark, J. J. Heyd, E. Brothers, K. N. Kudin, V. N. Staroverov, R. Kobayashi, J. Normand, K. Raghavachari, A. Rendell, J. C. Burant, S. S. Iyengar, J. Tomasi, M. Cossi, N. Rega, J. M. Millam, M. Klene, J. E. Knox, J. B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R. E. Stratmann, O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli, J. W. Ochterski, R. L. Martin, K. Morokuma, V. G. Zakrzewski, G. A. Voth, P. Salvador, J. J. Dannenberg, S. Dapprich, A. D. Daniels, Ã. Farkas, J. B. Foresman, J. V. Ortiz, J. Cioslowski and D. J. Fox, Gaussian09 Revision D.01, Gaussian Inc., Wallingford CT 2013.

---

Footnote

† Electronic supplementary information (ESI) available: Cartesian coordinates of all the optimized structures, SCF energies and number of imaginary frequencies. See DOI: 10.1039/c6ra10789c

---