FULL PAPER DOI: 10.1002/asia.201402159
Synthetic and Computational Evaluation of Regiodivergent Epoxide Opening for Diol and Polyol Synthesis Andreas Gansuer,*[a] Peter Karbaum,[a] David Schmauch,[a] Martin Einig,[a] Lili Shi,[b] Anakuthil Anoop,[b] and Frank Neese*[b] Abstract: In a combined synthetic and computational study, the factors governing the selectivity of the titanoceneACHTUNGRE(III)-catalyzed regiodivergent epoxide opening (REO) with Kagans complex via electron transfer leading to derivatives of 1,2-, 1,3-, and 1,4-diols were investigated. In this manner, valuable building blocks for the synthesis of 1,3- and 1,4-diols were identified. The computational study provides crucial structural features and energies of the transition states of ring opening that are important for the design of more selective catalysts.
lyzed diastereoselective epoxidation[4] of enantiomerically pure hydroxy olefins. Moreover, epoxides are reactive functional groups and can be easily activated for ring opening by a suitable catalyst. Also, the products of ring opening, usually alcohols, are important synthetic intermediates or targets. We have contributed the regiodivergent epoxide opening (REO) to the field of divergent reactions.[5] The method generates radicals catalytically from epoxides by electron transfer from titanoceneACHTUNGRE(III) complexes[6] and is based on the stoichiometric reaction described by Nugent and RajanBabu (Scheme 1).[7]
Introduction The use of divergent reactions of enantiomerically pure catalysts on racemic mixtures is an attractive approach to the synthesis of enantioenriched compounds.[1] It relies on the generation of separable isomeric products from the enantiomers of the substrates at similar rates through reaction with an enantiomerically pure catalyst. The latter point may provide a distinct advantage of divergent reactions over kinetic resolutions that require large differences in the relative rates (selectivity > 100) for obtaining high yields and high enantioselectivities of the desired products. Regiodivergent reactions are also highly attractive in the manipulation of enantiomerically pure substrates. Employing either enantiomer of the catalyst ideally leads to the reagent-controlled, highly selective access to regioisomeric products from a single substrate. For a number of reasons, epoxide openings are attractive substrates for divergent reactions. Epoxides can be readily synthesized in a highly enantiomerically enriched form by a number of methods, such as metal-catalyzed enantioselective epoxidation of unfunctionalized olefins or suitably unsaturated alcohols,[2] the chromium-catalyzed hydrolytic kinetic resolution of racemic epoxides,[3] and the vanadium-cata-
Scheme 1. Typical examples of the REO with 1 as catalyst.
[a] Prof. Dr. A. Gansuer, Dr. P. Karbaum, D. Schmauch, M. Einig Kekul-Institut fr Organische Chemie und Biochemie Universitt Bonn Gerhard Domagk Str. 1, 53121 Bonn (Germany) Fax: (+ 49) 0228-73-5683 E-mail: [email protected]
Kagans complex (1) has,[8] as yet, resulted in the highest enantioselectivities. The radical intermediates have been used in epoxide reduction through hydrogen-atom transfer or in cyclization reactions that increase molecular complexity substantially. Our REO is complimentary to regiodivergent reactions of strained heterocycles with nucleophiles that yield 1,2-difuntionalized compounds.[9] Here, we report a combined synthetic and theoretical approach towards understanding what factors control the re-
[b] Dr. L. Shi, Dr. A. Anoop, Prof. Dr. F. Neese Max-Planck Institut for Chemical Energy Conversion Stiftstr. 34-36, 45470 Mlheim an der Ruhr Fax: (+ 49) 0208-306-3951 E-mail: [email protected] Supporting information for this article is available on the WWW under http://dx.doi.org/10.1002/asia.201402159.
Chem. Asian J. 2014, 00, 0 – 0
These are not the final page numbers! ÞÞ
Keywords: catalysis · density functional calculations · enantioselectivity · epoxides · radicals
1
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
www.chemasianj.org
Andreas Gansuer, Frank Neese et al.
gioselectivity of the REO. By comparing experimental results with a computational analysis of the REO we have identified value building blocks for the synthesis of 1,3- and 1,4-diols. Moreover, the structures and energies of transition structures of the stereodifferentiating ring opening step were studied.
Results and Discussion Synthetic Studies In order to employ the REO for diol and polyol synthesis, (protected) hydroxylated epoxides have to be employed as substrates. We chose the REO of racemic protected Sharpless epoxides[2b] as starting point for our investigation because the unprotected derivatives lead to low conversion. Racemic substrates result in parallel kinetic resolutions and, hence, the reactions of both enantiomers of the substrate can be studied in a single reaction. Of course, for target-oriented synthetic applications these substrates are also readily available in enantiomerically enriched form.[2b] A typical example with a silylated racemic epoxide (2) and Kagans complex (1) is shown in Scheme 2.
Scheme 3. Qualitative analysis of the REO of 2.
Scheme 4. Influence of steric bulk on the REO of protected Sharpless epoxides.
Scheme 2. REO of 2 with 1 as catalyst. Scheme 5. REO of epoxide 8 derived from a homoallylic alcohol.
The absolute configurations of the products were assigned by comparing the signs of the specific rotation of the diols obtained after deprotection of 3 and 4 to literature values of these compounds (see the Supporting Information for details). The relatively low overall regioselectivity (3:4 = 36:64) and the high enantioselectivity in the formation of 3 (e.r. = 98:2) allow a qualitative analysis of the REO of racemic 2 (Scheme 3). The (2R,3S) enantiomer of 2 is opened with very high regioselectivity, whereas the (2S,3R) enantiomer shows a low selectivity. This results in a highly selective formation of (R)-3, in agreement with the cyclization shown in Scheme 1. Next, we investigated the introduction of the bulkier iPr substituent in 5 (Scheme 4) to probe the influence of steric bulk of the alkyl substituent in such substrates. The regio- and enantioselectivities are essentially identical to the reaction of 2. Therefore, the absolute configurations of the products were assigned as for 2. Since the REO is usually rather sensitive to steric effects,[5] this suggests that the electron-withdrawing nature of the -OTBS group may be a selectivity-determining factor too. This point was further probed with substrate 8. The additional CH2 group reduces the inductive effect of the OTBS group and results in a highly enantioselective formation of the products together with a ratio of 9:10 that is close to the
&
&
Chem. Asian J. 2014, 00, 0 – 0
ideal value of 50:50 (Scheme) 5.[1] Thus, derivatives of 8 seem attractive substrates for the preparation of 1,3- or 1,4diols with the REO. The regioselectivity of the REO of 8 deserves comment. In an ideal reaction, the REO cannot produce the major regioisomer in higher enantiomeric ratio. In the experiment, this is not the case, however. Since the outcome of the REO of 8 is reproducible, experimental errors, for example, in the determination of the enantiomeric ratios, are rather unlikely. To explain the regioselectivity, we propose that the radicals leading to the generation of 10 are partially consumed by an undesired side reaction, such as deoxygenation. This issue will be discussed further in the context of the computational analysis of the REO of 8. In Scheme 6 the reaction of 11 containing a third stereocenter is summarized. Here, the enantiomer of 1 was used as catalyst. It should be noted that, in contrast to the Sharpless epoxides, the free hydroxy group is readily tolerated in the REO. The diols 12 and 13 contain two stereocenters and are therefore more relevant for synthetic applications than 3, 4, 6, and 7. In this reaction the regioselectivity of the REO is close to 50:50. The 1,3-diol 13 is obtained with high enantioselectivity and the 1,4-diol 12with somewhat lower selectivity. The
2
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!
www.chemasianj.org
Andreas Gansuer, Frank Neese et al.
Our theoretical calculations provide differences in enthalpies of activation and the structures of the transition states of ring opening. The very accurate high-level wave functionbased ab initio methods are unsuitable for this particular problem. Therefore, the less rigorous approach of density functional theory (DFT) is used. To determine which method is suited best for our purpose, the transition states of the opening of 2 a (Scheme 7) were studied and compared
Scheme 6. Influence of an additional stereocenter on the REO of epoxides from homoallylic alcohols.
absolute configuration of the products could not be assigned. Thus, our synthetic studies demonstrate that silylated Sharpless epoxides are not suitable substrates for the REO. It seems that this is due to electronic effects of the -OTBS substituent. However, epoxides derived from homoallylic epoxides constitute more attractive precursors for the preparation of 1,3- and 1,4-diols via the REO. The underlying mechanistic principles for these observations were investigated with the aid of computational methods.
Computational Studies of the REO Scheme 7. Mechanistic scheme for the computational analysis of the REO of 2 a.
Methods The ORCA[10] electronic structure program package was used for all calculations. Structures and transition states were optimized with the BP86 functional[11] and Ahlrichs def2-SVP[12] basis set, making use of the resolution-of-theidentity RI approximation (in the Split-RI-J variant[13] with the appropriate Coulomb fitting sets).[14] The transition state searcher of the ORCA package is based on eigenvector following and is very efficient due to the use of partial Hessians. Transition states were found by first performing a relaxed surface scan along the bonds to be broken. The maximum energy structures of these scans were taken as a guess for the transition state search. Transition states and minima were verified through frequency calculations that were performed by two-sided numerical differentiation of analytic gradients. Zero-point vibrational energy (ZPVE) and thermal corrections were computed from the calculated harmonic frequencies and are included in the activation energies throughout. All calculations on open-shell species employed the spin-unrestricted formalism. The final single point energies were computed with the hybrid B3LYP[15] density functional and the basis set of triple-z quality including high angular momentum polarization functions (def2-TZVPP).[16]The density fitting and chain of sphere (RIJCOSX) approximations[17] have been employed for the Coulomb and Hartree–Fock exchange terms. To take the effect of solvent into account, THF was simulated by a polarizable continuum model (COSMO).[18]
to the experimental results of the closely related 2. From the energy differences of the transition states the enantiomeric ratios of 3 a and 4 a can be directly calculated assuming a quick and irreversible radical reduction by hydrogen atom transfer. From these ratios the regioselectivity of the REO can be directly deduced. The results of the calculations employing BP86 and B3LYP as well as the values for the COSMO-B3LYP calculations are summarized in Table 1. Since we are interested in the selectivities, only differences of the enthalpy of activation are given (see the Supporting Information for a table of the absolute values). The BP functional overestimates the formation of 3 a (47:53 vs. 36:64 in the experiment with 2) and its enantiomeric ratio substantially and is therefore not suitable for the description of the REO. The B3LYP and COSMO-B3LYP calculations are in good agreement with the experimental data. We decided to employ COSMO-B3LYP to include the effect of bulk solvation in the following calculations. In agreement with the experiment, the computational REO of 8 suggests that epoxides derived from homoallylic alcohols are more attractive substrates than protected Sharpless epoxides (Table 2). The analysis was carried out in analogy to Scheme 7. The enantioselectivites are in very good agreement with the synthetic results that are only slightly underestimated for 10. Therefore, the absolute configurations of the synthetic compounds were assigned as shown above. The regioselectivity of the computational REO (9:10 = 45.5:54.5) correctly predicts a higher enantiomeric ratio for the minor
Computational investigation of REO selectivity Quality of the DFT calculations and computational examples of the REO:
Chem. Asian J. 2014, 00, 0 – 0
These are not the final page numbers! ÞÞ
3
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
Andreas Gansuer, Frank Neese et al.
www.chemasianj.org
Table 1. Experimental result of the REO with 2 and computational analysis of the REO of 2 a for different functionals (BP86, B3LYP, and COSMO-B3LYP) and the basis set def2-TZVPP at 293 K. The differences in transition states energies are given in kcal mol 1.
Method
TS3 a– TS4 a
BP86 B3LYP COSMOB3LYP
1.2 0.3 0.2
TS3 a*– TS4 a*
e.r. (3 a)[a]
e.r. (4 a)[a]
3 a:4 a
2.8 4.4 4.8
99:1 100:0 100:0
10.5:89.5 27:73 29.5:70.5
47:53 31:69 29:71
Table 3. Experimental result of the REO of 11 and computational analysis of the REO of 11 a at the COSMO-B3LYP/def2-TZVPP//BP86/def2SVP level at 293 K. The differences in transition states energies are in kcal mol 1.
TS12 a–TS13 a 2.5
TS12 a*–TS13 a*
e.r. (12 a)[a]
e.r. (13 a)[b]
12 a:13 a
1.2
89.5:10.5
1.5:98.5
55:45
[a] (2S,5R):ACHTUNGRE(2R,5S). [b] (2S,4R):ACHTUNGRE(2R,4S).
[a] (R):(S).
served experimentally and computationally. Both enantiomers of the substrate are opened with high and opposite selectivity concerning the absolute configuration of the respective products. Thus, the opening of 8 is occurring under reagent control. This is in line with the selectivities observed in the opening of meso-epoxides by 1. Thus, for 2 there must be an influence of substrate control that results in either a matching or mismatching with reagent control exerted by the catalyst. The reason for this can be deduced from the transition structures for the opening of 2 a (Figure 1). The matched case is the highly selective opening of (2R,3S)-2 and (2R,3S)-2 a and the mismatched case the essentially unselective opening of (2S,3R)-2 and (2S,3R)-2 a. Transition states TS-3 a and TS-4 a that describe the mismatched opening of (2S,3R)-2 a are analyzed first. Based on the proposed mechanism for epoxide opening,[19] TS-4 a is disfavored because breaking of the left C O bond results in an increased interaction between the right ligand and the CH2OTBS group. TS-3 a leads to a reduction of this interaction and is therefore favored for steric reasons. However, the steric preference is counterbalanced by the electronwithdrawing effect of the CH2OTBS group that destabilizes the emerging radical center in TS-3 a. The respective enthalpies of activation are almost identical and hence (2R,3S)-2 a is predicted to be opened unselectively. This is indeed observed for (2R,3S)-2. For the matched case (opening of (2R,3S)-2 a via TS-3 a* and TS-4 a*), the sterically disfavored TS-3 a* is also disfavored electronically. As a consequence, the difference between the enthalpies is large (4.8 kcal mol 1) and the opening of (2R,3S)-2 a highly selective. This is in agreement with the results for (2R,3S)-2. The transition states for the opening of 11 a [(2S,4S,5R)11 a: TS-12 a and TS-13 a; (2R,4R,5S)-11 a: TS-12 a* and TS13 a*] are shown in Figure 2. For both enantiomers, the differences in the enthalpies of activation are in the typical range for simple steric effects (about 2 kcal mol 1 as derived from the selectivities in the
Table 2. Results of the computational analysis of the REO of 8 and experimental result of the REO at the COSMO-B3LYP/def2-TZVPP// BP86/def2-SVP level at 293 K. The differences in transition states energies are in kcal mol 1.
TS9–TS10 1.6
TS9*–TS10*
e.r. (9)[a]
e.r. (10)[a]
9:10
1.0
94.5:6.5
8.5:91.5
45.5:54.5
[a] (R):(S).
product. This is a consequence of the assumption of a complete reduction of all intermediate radicals. This lends support to the notion that in the experimental REO of 8 incomplete radical reduction is an issue, albeit a minor one. Finally, we analyzed the REO of 11 a by means of COSMO-B3LYP (Table 3) to understand the influence of the second stereocenter on the outcome of the reaction. Compound 11 a is closely related to the synthetic substrate 11. The computational selectivities nicely match the experiment. Thus, the absolute configuration of 12 and 13 was assigned as shown above keeping in mind that in the experiment the ent-1 was used as catalyst. Analysis of the transition structures of the REO The consistently high agreement between experimental and computational selectivity allows the proposal of a mechanistic model for the origins of selectivity of the REO based on the structure of the transition states. This is outlined for the differences between the REO of 2 a and 8. For 2 a, the (2R,3S) enantiomer is opened with very high selectivity to radical 4 a*. The (2S,3R) enantiomer is opened with low selectivity (3 a:4 a = 58:42). For 11, this is not ob-
&
&
Chem. Asian J. 2014, 00, 0 – 0
4
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!
www.chemasianj.org
Andreas Gansuer, Frank Neese et al.
opening of meso-epoxides).[20] It should be noted that in TS13 a the steric destabilization is larger than in TS-12 a* since the group containing the secondary hydroxy group (TS-13 a) is bulkier than the n-alkyl chain (TS-12 a*). Thus, in the absence of electronic effects, the REO selectivity is indeed governed by steric effects in agreement with the postulated mechanism for the enantioselective opening of meso-epoxides by 1.[20]
Conclusions In summary, we have demonstrated that epoxides derived from homoallylic alcohols are very attractive substrates for the synthesis of 1,3- and 1,4- diols via the REO. Protected Sharpless epoxides are less suitable for the synthesis of 1,2and 1,3-diols because a combination of reagent and substrate control leads to matched and mismatched cases of epoxide opening. The reasons for the differences, that is, electronic effects influencing the stability of the emerging radical centers in the case of Sharpless epoxides, were identified with the aid of computational investigations. Moreover, the computational results allow an assignment of the absolute configuration of the experimentally obtained products. The transition structures provide valuable clues for the design of more selective catalysts. Investigations in this area are underway, and results will be reported in due course.
Figure 1. Transition states of the REO of the (2S,3R) enantiomer of 2 a (TS-3 a and TS-4 a) and (2R,3S) enantiomer of 2 a (TS-3 a* and TS-4 a*) and their respective energies (COSMO-B3LYP).
Experimental Section General procedure for the REO of substrates 2, 5, and 8 To a mixture of 2,4,6-collidine hydrochloride (1.5 equiv), 1 (0.1 equiv) and Mn (1.5 equiv) in THF (0.1 m with respect to epoxide) were added 1,4-cyclohexadiene (4.4 equiv) and the epoxide (1 equiv). After stirring at room temperature overnight, the reaction was quenched by the addition of H3PO4/KH2PO4 buffer (2:1, 3 mol L 1, pH = 1.9). The mixture was then extracted twice with EtOAc and twice with CH2Cl2. The organic layers were dried over MgSO4 and the volatiles removed in vacuo. The crude product was purified by chromatography on silica gel using the given eluents. Procedure for the REO of substrate 2 According to the general procedure, 2,4,6-collidine hydrochloride (470 mg, 3.0 mmol, 1.5 equiv), Mn (180 mg, 3.27 mmol, 1.7 equiv), (2S*,3R*)-2 (540 mg, 2.0 mmol, 1.0 equiv), and 1,4-cyclohexadiene (700 mg, 8.8 mmol, 4.4 equiv) in THF (5.5 mL) were treated with 1 (103 mg, 0.2 mmol, 0.1 equiv) overnight. Purification (cyclohexane/ EtOAc 90:10) yielded (3R)-3 (159 mg, 29 %) and (2S)-4 (282 mg, 52 %). Procedure for the REO of substrate 5 According to the general procedure, 2,4,6-collidine hydrochloride (472 mg, 3.0 mmol, 1.5 equiv), Mn (166 mg, 3.0 mmol, 1.5 equiv), (2S*,3R*)-5 (462 mg, 2.0 mmol, 1.0 equiv), and 1,4-cyclohexadiene (700 mg, 8.8 mmol, 4.4 equiv) in THF (6 mL) were treated with 1 (106 mg, 0.2 mmol, 0.1 equiv) overnight. Purification (cyclohexane/ EtOAc 96:4) yielded (3S)-6 (112 mg, 29 %) and (2S)-7 (297 mg, 59 %). Procedure for the REO of substrate 8 Figure 2. Transition states of the REO of the (2S,4S,5R) enantiomer of 11 a (TS-12 a and TS-13 a) and (2R,4R,5S) enantiomer of 11 a (TS-12 a* and TS-13 a*) and their respective energies (COSMO-B3LYP).
Chem. Asian J. 2014, 00, 0 – 0
These are not the final page numbers! ÞÞ
According to the general procedure, 2,4,6-collidine hydrochloride (236 mg, 1.5 mmol, 1.5 equiv), Mn (83 mg, 1.5 mmol, 1.5 equiv), (3S*,4R*)-8 (237 mg, 1.03 mmol, 1 equiv), and 1,4-cyclohexadiene
5
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
www.chemasianj.org
(350 mg, 4.4 mmol, 4.4 equiv) in THF (3 mL) were treated with 1 (53 mg, 0.1 mmol, 0.1 equiv) overnight. Purification (cyclohexane/EtOAc 90:10) gave 9 (111 mg, 47 %) and 10 (100 mg, 42 %). Procedure for the REO of substrate 11 To a mixture of 2,6-lutidine hydrochloride (280 mg, 2.0 mmol, 2.0 equiv), ent-1 (106 mg, 0.2 mmol, 0.2 equiv), and Mn (120 mg, 2.2 mmol, 2.2 equiv) in THF (3 mL) was added Bu3SnH (580 mg, 2.0 mmol, 2.0 equiv) and (2S*,4S*,5R*)-11 (166 mg, 1.0 mmol, 1.0 equiv). After stirring at room temperature for 72 h, the reaction was quenched by the addition of phosphate buffer (1 mL) and Et2O (5 mL). The mixture was then filtered through a short column (Et2O, SiO2/K2CO3). Purification (cyclohexane/EtOAc 80:20, SiO2) yielded 12 (80 mg, 48 %) and 13 (68 mg, 40 %).
[7]
[8]
Further details of compound preparation and characterization as well as transition state coordinates and energies are provided in the Supporting Information. [9]
Acknowledgements [10] [11]
We are indebted to the Deutsche Forschungsgemeinschaft (SFB 813 “Chemistry at Spin Centers”) for generous financial support.
[12] [1] Reviews: a) H. B. Kagan, Croat. Chem. Acta 1996, 69, 669 – 680; b) J. Eames, Angew. Chem. 2000, 112, 913 – 916; Angew. Chem. Int. Ed. 2000, 39, 885 – 888; c) H. B. Kagan, Tetrahedron 2001, 57, 2449 – 2468; d) J. R. Dehli, V. Gotor, Chem. Soc. Rev. 2002, 31, 365 – 370; e) E. Vedejs, M. Jure, Angew. Chem. 2005, 117, 4040 – 4069; Angew. Chem. Int. Ed. 2005, 44, 3974 – 4001; f) R. R. Kumar, H. B. Kagan, Adv. Synth. Catal. 2010, 352, 231 – 242; g) L. C. Miller, R. Sarpong, Chem. Soc. Rev. 2011, 40, 4550 – 4562; h) M. Pineschi, V. di Bussolo, P. Crotti, Chirality 2011, 23, 703 – 710. [2] Hydroxylated olefins: a) T. Katsuki, K. B. Sharpless, J. Am. Chem. Soc. 1980, 102, 5974 – 5976; b) R. A. Johnson, K. B. Sharpless in Catalytic Asymmetric Synthesis (Ed.: I. Ojima), Wiley-VCH, New York, 2000, pp. 231 – 280; c) W. Zhang, H. Yamamoto, J. Am. Chem. Soc. 2007, 129, 286 – 287; Unfunctionalized olefins: d) W. Zhang, J. L. Loebach, S. R. Wilson, E. N. Jacobsen, J. Am. Chem. Soc. 1990, 112, 2801 – 2803; e) R. Irie, K. Noda, N. Matsumoto, T. Katsuki, Tetrahedron Lett. 1990, 31, 7345 – 7348; f) Y. Sawada, K. Matsumoto, S. Kondo, H. Watanabe, T. Ozawa, K. Suzuki, B. Saito, T. Katsuki, Angew. Chem. 2006, 118, 3558 – 3560; Angew. Chem. Int. Ed. 2006, 45, 3478 – 3480. [3] a) E. N. Jacobsen, M. Tokunaga, J. F. Larrow, F. Kakiuchi, Science 1997, 277, 936 – 938; b) H. Lebel, E. N. Jacobsen, Tetrahedron Lett. 1999, 40, 7303 – 7306; c) E. N. Jacobsen, Acc. Chem. Res. 2000, 33, 421 – 431. [4] a) K. B. Sharpless, R. C. Michaelson, J. Am. Chem. Soc. 1973, 95, 6136 – 6137; substrate directed reactions: b) A. H. Hoveyda, D. A. Evans, G. C. Fu, Chem. Rev. 1993, 93, 1307 – 1370. [5] a) A. Gansuer, C.-A. Fan, F. Keller, J. Keil, J. Am. Chem. Soc. 2007, 129, 3484 – 3485; b) A. Gansuer, C.-A. Fan, F. Keller, P. Karbaum, Chem. Eur. J. 2007, 13, 8084 – 8090; c) A. Gansuer, S. Lei, M. Otte, J. Am. Chem. Soc. 2010, 132, 11858 – 11859; d) A. Gansuer, L. Shi, F. Keller, P. Karbaum, C.-A. Fan, Tetrahedron: Asymmetry 2010, 21, 1361 – 1369. [6] a) A. Gansuer, M. Pierobon, H. Bluhm, Angew. Chem. Int. Ed. 1998, 37, 101 – 103; Angew. Chem. 1998, 110, 107 – 109; b) A. Gansuer, H. Bluhm, M. Pierobon, J. Am. Chem. Soc. 1998, 120, 12849 – 12859; c) A. F. Barrero, A. Rosales, J. M. Cuerva, J. E. Oltra, Org.
&
&
Chem. Asian J. 2014, 00, 0 – 0
[13]
[14]
[15] [16] [17]
[18] [19]
[20]
6
Andreas Gansuer, Frank Neese et al.
Lett. 2003, 5, 1935 – 1938; d) J. Justicia, A. Rosales, E. BuÇuel, J. L. Oller-Lpez, M. Valdivia, A. Haidour, J. E. Oltra, A. F. Barrero, D. J. C rdenas, J. M. Cuerva, Chem. Eur. J. 2004, 10, 1778 – 1788; Reviews: e) A. Gansuer, T. Lauterbach, S. Narayan, Angew. Chem. Int. Ed. 2003, 42, 5556 – 5573; Angew. Chem. 2003, 115, 5714 – 5731; f) J. M. Cuerva, J. Justicia, J. L. Oller-Lpez, J. E. Oltra, Top. Curr. Chem. 2006, 264, 63 – 92; g) A. Gansuer, J. Justicia, C.-A. Fan, D. Worgull, F. Piestert, Top. Curr. Chem. 2007, 279, 25 – 52. a) W. A. Nugent, T. V. RajanBabu, J. Am. Chem. Soc. 1988, 110, 8561 – 8562; b) T. V. RajanBabu, W. A. Nugent, J. Am. Chem. Soc. 1989, 111, 4525 – 4527; c) T. V. RajanBabu, W. A. Nugent, M. S. Beattie, J. Am. Chem. Soc. 1990, 112, 6408 – 6409; d) T. V. RajanBabu, W. A. Nugent, J. Am. Chem. Soc. 1994, 116, 986 – 997. a) E. Cesarotti, H. B. Kagan, R. Goddard, C. Krger, J. Organomet. Chem. 1978, 162, 297 – 309; b) Structure: A. Gansuer, H. Bluhm, M. Pierobon, M. Keller, Organometallics 2001, 20, 914 – 919; c) M. Klahn, P. Arndt, A. Spannenberg, A. Gansuer, U. Rosenthal, Organometallics 2008, 27, 5846 – 5851. a) B. Wu, J. R. Parquette, T. V. RajanBabu, Science 2009, 326, 1662 – 1662; b) R. Webster, C. Bçing, M. Lautens, J. Am. Chem. Soc. 2009, 131, 444 – 445; c) B. Wu, J. C. Gallucci, J. R. Parquette, T. V. RajanBabu, Chem. Sci. 2014, 5, 1102 – 1117. F. Neese, WIREs Comput. Mol. Sci. 2012, 2, 73 – 78. a) A. D. Becke, Phys. Rev. A 1988, 38, 3098 – 3100; b) J. P. Perdew, Phys. Rev. B 1986, 33, 8822 – 8824. a) F. Weigend, R. Ahlrichs, Phys. Chem. Chem. Phys. 2005, 7, 3297 – 3305; b) A. Schfer, H. Horn, R. Ahlrichs, J. Chem. Phys. 1992, 97, 2571 – 2577. a) B. I. Dunlap, J. W. D. Connolly, J. R. Sabin, J. Chem. Phys. 1979, 71, 3396 – 3402; b) E. J. Baerends, D. E. Ellis, P. Ros, Chem. Phys. 1973, 2, 41 – 51; c) K. Eichkorn, O. Treutler, H. hm, M. Hser, R. Ahlrichs, Chem. Phys. Lett. 1995, 242, 652 – 660. a) K. Eichkorn, O. Treutler, H. hm, M. Hser, R. Ahlrichs, Chem. Phys. Lett. 1995, 240, 283 – 289; b) K. Eichkorn, F. Weigend, O. Treutler, R. Ahlrichs, Theor. Chem. Acc. 1997, 97, 119 – 124. a) A. D. Becke, J. Chem. Phys. 1993, 98, 5648 – 5652; b) C. T. Lee, W. T. Yang, R. G. Parr, Phys. Rev. B 1988, 37, 785 – 789. F. Weigend, M. Haser, H. Patzelt, R. Ahlrichs, Chem. Phys. Lett. 1998, 294, 143 – 152. a) F. Neese, F. Wennmohs, A. Hansen, U. Becker, Chem. Phys. 2009, 356, 98 – 109; b) S. Kossmann, F. Neese, Chem. Phys. Lett. 2009, 481, 240 – 243. A. Klamt, G. Schuurmann, J. Chem. Soc. Perkin Trans. 2 1993, 799 – 805. a) K. Daasbjerg, H. Svith, S. Grimme, M. Gerenkamp, C. MckLichtenfeld, A. Gansuer, A. Barchuk, F. Keller, Angew. Chem. Int. Ed. 2006, 45, 2041 – 2044; Angew. Chem. 2006, 118, 2095 – 2098; b) A. Gansuer, A. Barchuk, F. Keller, M. Schmitt, S. Grimme, M. Gerenkamp, C. Mck-Lichtenfeld, K. Daasbjerg, H. Svith, J. Am. Chem. Soc. 2007, 129, 1359 – 1371; c) A. Gansuer, A. Fleckhaus, M. Alejandre Lafont, A. Okkel, K. Kotsis, A. Anoop, F. Neese, J. Am. Chem. Soc. 2009, 131, 16989 – 16999. a) A. Gansuer, T. Lauterbach, H. Bluhm, M. Noltemeyer, Angew. Chem. Int. Ed. 1999, 38, 2909 – 2910; Angew. Chem. 1999, 111, 3112 – 3114; b) A. Gansuer, M. Pierobon, Synlett 2000, 1357 – 1359; c) A. Gansuer, H. Bluhm, T. Lauterbach, Adv. Synth. Catal. 2001, 343, 785 – 787; d) A. Gansuer, H. Bluhm, B. Rinker, S. Narayan, M. Schick, T. Lauterbach, M. Pierobon, Chem. Eur. J. 2003, 9, 531 – 542; e) J. Friedrich, K. Walczak, M. Dolg, F. Piestert, T. Lauterbach, D. Worgull, A. Gansuer, J. Am. Chem. Soc. 2008, 130, 1788 – 1796; f) A. Gansuer, C.-A. Fan, F. Piestert, J. Am. Chem. Soc. 2008, 130, 6916 – 6917. Received: March 3, 2014 Published online: && &&, 0000
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!
FULL PAPER Catalysis
Reasons for the selectivity of the REO unraveled: Through a combination of synthesis and theory, the factors governing the selectivity of the regiodiver-
gent epoxide opening (REO) were resolved. Our study suggests useful new building blocks for the synthesis of 1,3- and 1,4-diol units.
Chem. Asian J. 2014, 00, 0 – 0
These are not the final page numbers! ÞÞ
7
Andreas Gansuer,* Peter Karbaum, David Schmauch, Martin Einig, Lili Shi, Anakuthil Anoop, &&&&—&&&& Frank Neese* Synthetic and Computational Evaluation of Regiodivergent Epoxide Opening for Diol and Polyol Synthesis
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
Synthetic and Computational Evaluation of Regiodivergent Epoxide Opening for Diol and Polyol Synthesis Andreas Gansuer,*[a] Peter Karbaum,[a] David Schmauch,[a] Martin Einig,[a] Lili Shi,[b] Anakuthil Anoop,[b] and Frank Neese*[b] Abstract: In a combined synthetic and computational study, the factors governing the selectivity of the titanoceneACHTUNGRE(III)-catalyzed regiodivergent epoxide opening (REO) with Kagans complex via electron transfer leading to derivatives of 1,2-, 1,3-, and 1,4-diols were investigated. In this manner, valuable building blocks for the synthesis of 1,3- and 1,4-diols were identified. The computational study provides crucial structural features and energies of the transition states of ring opening that are important for the design of more selective catalysts.
lyzed diastereoselective epoxidation[4] of enantiomerically pure hydroxy olefins. Moreover, epoxides are reactive functional groups and can be easily activated for ring opening by a suitable catalyst. Also, the products of ring opening, usually alcohols, are important synthetic intermediates or targets. We have contributed the regiodivergent epoxide opening (REO) to the field of divergent reactions.[5] The method generates radicals catalytically from epoxides by electron transfer from titanoceneACHTUNGRE(III) complexes[6] and is based on the stoichiometric reaction described by Nugent and RajanBabu (Scheme 1).[7]
Introduction The use of divergent reactions of enantiomerically pure catalysts on racemic mixtures is an attractive approach to the synthesis of enantioenriched compounds.[1] It relies on the generation of separable isomeric products from the enantiomers of the substrates at similar rates through reaction with an enantiomerically pure catalyst. The latter point may provide a distinct advantage of divergent reactions over kinetic resolutions that require large differences in the relative rates (selectivity > 100) for obtaining high yields and high enantioselectivities of the desired products. Regiodivergent reactions are also highly attractive in the manipulation of enantiomerically pure substrates. Employing either enantiomer of the catalyst ideally leads to the reagent-controlled, highly selective access to regioisomeric products from a single substrate. For a number of reasons, epoxide openings are attractive substrates for divergent reactions. Epoxides can be readily synthesized in a highly enantiomerically enriched form by a number of methods, such as metal-catalyzed enantioselective epoxidation of unfunctionalized olefins or suitably unsaturated alcohols,[2] the chromium-catalyzed hydrolytic kinetic resolution of racemic epoxides,[3] and the vanadium-cata-
Scheme 1. Typical examples of the REO with 1 as catalyst.
[a] Prof. Dr. A. Gansuer, Dr. P. Karbaum, D. Schmauch, M. Einig Kekul-Institut fr Organische Chemie und Biochemie Universitt Bonn Gerhard Domagk Str. 1, 53121 Bonn (Germany) Fax: (+ 49) 0228-73-5683 E-mail: [email protected]
Kagans complex (1) has,[8] as yet, resulted in the highest enantioselectivities. The radical intermediates have been used in epoxide reduction through hydrogen-atom transfer or in cyclization reactions that increase molecular complexity substantially. Our REO is complimentary to regiodivergent reactions of strained heterocycles with nucleophiles that yield 1,2-difuntionalized compounds.[9] Here, we report a combined synthetic and theoretical approach towards understanding what factors control the re-
[b] Dr. L. Shi, Dr. A. Anoop, Prof. Dr. F. Neese Max-Planck Institut for Chemical Energy Conversion Stiftstr. 34-36, 45470 Mlheim an der Ruhr Fax: (+ 49) 0208-306-3951 E-mail: [email protected] Supporting information for this article is available on the WWW under http://dx.doi.org/10.1002/asia.201402159.
Chem. Asian J. 2014, 00, 0 – 0
These are not the final page numbers! ÞÞ
Keywords: catalysis · density functional calculations · enantioselectivity · epoxides · radicals
1
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
www.chemasianj.org
Andreas Gansuer, Frank Neese et al.
gioselectivity of the REO. By comparing experimental results with a computational analysis of the REO we have identified value building blocks for the synthesis of 1,3- and 1,4-diols. Moreover, the structures and energies of transition structures of the stereodifferentiating ring opening step were studied.
Results and Discussion Synthetic Studies In order to employ the REO for diol and polyol synthesis, (protected) hydroxylated epoxides have to be employed as substrates. We chose the REO of racemic protected Sharpless epoxides[2b] as starting point for our investigation because the unprotected derivatives lead to low conversion. Racemic substrates result in parallel kinetic resolutions and, hence, the reactions of both enantiomers of the substrate can be studied in a single reaction. Of course, for target-oriented synthetic applications these substrates are also readily available in enantiomerically enriched form.[2b] A typical example with a silylated racemic epoxide (2) and Kagans complex (1) is shown in Scheme 2.
Scheme 3. Qualitative analysis of the REO of 2.
Scheme 4. Influence of steric bulk on the REO of protected Sharpless epoxides.
Scheme 2. REO of 2 with 1 as catalyst. Scheme 5. REO of epoxide 8 derived from a homoallylic alcohol.
The absolute configurations of the products were assigned by comparing the signs of the specific rotation of the diols obtained after deprotection of 3 and 4 to literature values of these compounds (see the Supporting Information for details). The relatively low overall regioselectivity (3:4 = 36:64) and the high enantioselectivity in the formation of 3 (e.r. = 98:2) allow a qualitative analysis of the REO of racemic 2 (Scheme 3). The (2R,3S) enantiomer of 2 is opened with very high regioselectivity, whereas the (2S,3R) enantiomer shows a low selectivity. This results in a highly selective formation of (R)-3, in agreement with the cyclization shown in Scheme 1. Next, we investigated the introduction of the bulkier iPr substituent in 5 (Scheme 4) to probe the influence of steric bulk of the alkyl substituent in such substrates. The regio- and enantioselectivities are essentially identical to the reaction of 2. Therefore, the absolute configurations of the products were assigned as for 2. Since the REO is usually rather sensitive to steric effects,[5] this suggests that the electron-withdrawing nature of the -OTBS group may be a selectivity-determining factor too. This point was further probed with substrate 8. The additional CH2 group reduces the inductive effect of the OTBS group and results in a highly enantioselective formation of the products together with a ratio of 9:10 that is close to the
&
&
Chem. Asian J. 2014, 00, 0 – 0
ideal value of 50:50 (Scheme) 5.[1] Thus, derivatives of 8 seem attractive substrates for the preparation of 1,3- or 1,4diols with the REO. The regioselectivity of the REO of 8 deserves comment. In an ideal reaction, the REO cannot produce the major regioisomer in higher enantiomeric ratio. In the experiment, this is not the case, however. Since the outcome of the REO of 8 is reproducible, experimental errors, for example, in the determination of the enantiomeric ratios, are rather unlikely. To explain the regioselectivity, we propose that the radicals leading to the generation of 10 are partially consumed by an undesired side reaction, such as deoxygenation. This issue will be discussed further in the context of the computational analysis of the REO of 8. In Scheme 6 the reaction of 11 containing a third stereocenter is summarized. Here, the enantiomer of 1 was used as catalyst. It should be noted that, in contrast to the Sharpless epoxides, the free hydroxy group is readily tolerated in the REO. The diols 12 and 13 contain two stereocenters and are therefore more relevant for synthetic applications than 3, 4, 6, and 7. In this reaction the regioselectivity of the REO is close to 50:50. The 1,3-diol 13 is obtained with high enantioselectivity and the 1,4-diol 12with somewhat lower selectivity. The
2
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!
www.chemasianj.org
Andreas Gansuer, Frank Neese et al.
Our theoretical calculations provide differences in enthalpies of activation and the structures of the transition states of ring opening. The very accurate high-level wave functionbased ab initio methods are unsuitable for this particular problem. Therefore, the less rigorous approach of density functional theory (DFT) is used. To determine which method is suited best for our purpose, the transition states of the opening of 2 a (Scheme 7) were studied and compared
Scheme 6. Influence of an additional stereocenter on the REO of epoxides from homoallylic alcohols.
absolute configuration of the products could not be assigned. Thus, our synthetic studies demonstrate that silylated Sharpless epoxides are not suitable substrates for the REO. It seems that this is due to electronic effects of the -OTBS substituent. However, epoxides derived from homoallylic epoxides constitute more attractive precursors for the preparation of 1,3- and 1,4-diols via the REO. The underlying mechanistic principles for these observations were investigated with the aid of computational methods.
Computational Studies of the REO Scheme 7. Mechanistic scheme for the computational analysis of the REO of 2 a.
Methods The ORCA[10] electronic structure program package was used for all calculations. Structures and transition states were optimized with the BP86 functional[11] and Ahlrichs def2-SVP[12] basis set, making use of the resolution-of-theidentity RI approximation (in the Split-RI-J variant[13] with the appropriate Coulomb fitting sets).[14] The transition state searcher of the ORCA package is based on eigenvector following and is very efficient due to the use of partial Hessians. Transition states were found by first performing a relaxed surface scan along the bonds to be broken. The maximum energy structures of these scans were taken as a guess for the transition state search. Transition states and minima were verified through frequency calculations that were performed by two-sided numerical differentiation of analytic gradients. Zero-point vibrational energy (ZPVE) and thermal corrections were computed from the calculated harmonic frequencies and are included in the activation energies throughout. All calculations on open-shell species employed the spin-unrestricted formalism. The final single point energies were computed with the hybrid B3LYP[15] density functional and the basis set of triple-z quality including high angular momentum polarization functions (def2-TZVPP).[16]The density fitting and chain of sphere (RIJCOSX) approximations[17] have been employed for the Coulomb and Hartree–Fock exchange terms. To take the effect of solvent into account, THF was simulated by a polarizable continuum model (COSMO).[18]
to the experimental results of the closely related 2. From the energy differences of the transition states the enantiomeric ratios of 3 a and 4 a can be directly calculated assuming a quick and irreversible radical reduction by hydrogen atom transfer. From these ratios the regioselectivity of the REO can be directly deduced. The results of the calculations employing BP86 and B3LYP as well as the values for the COSMO-B3LYP calculations are summarized in Table 1. Since we are interested in the selectivities, only differences of the enthalpy of activation are given (see the Supporting Information for a table of the absolute values). The BP functional overestimates the formation of 3 a (47:53 vs. 36:64 in the experiment with 2) and its enantiomeric ratio substantially and is therefore not suitable for the description of the REO. The B3LYP and COSMO-B3LYP calculations are in good agreement with the experimental data. We decided to employ COSMO-B3LYP to include the effect of bulk solvation in the following calculations. In agreement with the experiment, the computational REO of 8 suggests that epoxides derived from homoallylic alcohols are more attractive substrates than protected Sharpless epoxides (Table 2). The analysis was carried out in analogy to Scheme 7. The enantioselectivites are in very good agreement with the synthetic results that are only slightly underestimated for 10. Therefore, the absolute configurations of the synthetic compounds were assigned as shown above. The regioselectivity of the computational REO (9:10 = 45.5:54.5) correctly predicts a higher enantiomeric ratio for the minor
Computational investigation of REO selectivity Quality of the DFT calculations and computational examples of the REO:
Chem. Asian J. 2014, 00, 0 – 0
These are not the final page numbers! ÞÞ
3
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
Andreas Gansuer, Frank Neese et al.
www.chemasianj.org
Table 1. Experimental result of the REO with 2 and computational analysis of the REO of 2 a for different functionals (BP86, B3LYP, and COSMO-B3LYP) and the basis set def2-TZVPP at 293 K. The differences in transition states energies are given in kcal mol 1.
Method
TS3 a– TS4 a
BP86 B3LYP COSMOB3LYP
1.2 0.3 0.2
TS3 a*– TS4 a*
e.r. (3 a)[a]
e.r. (4 a)[a]
3 a:4 a
2.8 4.4 4.8
99:1 100:0 100:0
10.5:89.5 27:73 29.5:70.5
47:53 31:69 29:71
Table 3. Experimental result of the REO of 11 and computational analysis of the REO of 11 a at the COSMO-B3LYP/def2-TZVPP//BP86/def2SVP level at 293 K. The differences in transition states energies are in kcal mol 1.
TS12 a–TS13 a 2.5
TS12 a*–TS13 a*
e.r. (12 a)[a]
e.r. (13 a)[b]
12 a:13 a
1.2
89.5:10.5
1.5:98.5
55:45
[a] (2S,5R):ACHTUNGRE(2R,5S). [b] (2S,4R):ACHTUNGRE(2R,4S).
[a] (R):(S).
served experimentally and computationally. Both enantiomers of the substrate are opened with high and opposite selectivity concerning the absolute configuration of the respective products. Thus, the opening of 8 is occurring under reagent control. This is in line with the selectivities observed in the opening of meso-epoxides by 1. Thus, for 2 there must be an influence of substrate control that results in either a matching or mismatching with reagent control exerted by the catalyst. The reason for this can be deduced from the transition structures for the opening of 2 a (Figure 1). The matched case is the highly selective opening of (2R,3S)-2 and (2R,3S)-2 a and the mismatched case the essentially unselective opening of (2S,3R)-2 and (2S,3R)-2 a. Transition states TS-3 a and TS-4 a that describe the mismatched opening of (2S,3R)-2 a are analyzed first. Based on the proposed mechanism for epoxide opening,[19] TS-4 a is disfavored because breaking of the left C O bond results in an increased interaction between the right ligand and the CH2OTBS group. TS-3 a leads to a reduction of this interaction and is therefore favored for steric reasons. However, the steric preference is counterbalanced by the electronwithdrawing effect of the CH2OTBS group that destabilizes the emerging radical center in TS-3 a. The respective enthalpies of activation are almost identical and hence (2R,3S)-2 a is predicted to be opened unselectively. This is indeed observed for (2R,3S)-2. For the matched case (opening of (2R,3S)-2 a via TS-3 a* and TS-4 a*), the sterically disfavored TS-3 a* is also disfavored electronically. As a consequence, the difference between the enthalpies is large (4.8 kcal mol 1) and the opening of (2R,3S)-2 a highly selective. This is in agreement with the results for (2R,3S)-2. The transition states for the opening of 11 a [(2S,4S,5R)11 a: TS-12 a and TS-13 a; (2R,4R,5S)-11 a: TS-12 a* and TS13 a*] are shown in Figure 2. For both enantiomers, the differences in the enthalpies of activation are in the typical range for simple steric effects (about 2 kcal mol 1 as derived from the selectivities in the
Table 2. Results of the computational analysis of the REO of 8 and experimental result of the REO at the COSMO-B3LYP/def2-TZVPP// BP86/def2-SVP level at 293 K. The differences in transition states energies are in kcal mol 1.
TS9–TS10 1.6
TS9*–TS10*
e.r. (9)[a]
e.r. (10)[a]
9:10
1.0
94.5:6.5
8.5:91.5
45.5:54.5
[a] (R):(S).
product. This is a consequence of the assumption of a complete reduction of all intermediate radicals. This lends support to the notion that in the experimental REO of 8 incomplete radical reduction is an issue, albeit a minor one. Finally, we analyzed the REO of 11 a by means of COSMO-B3LYP (Table 3) to understand the influence of the second stereocenter on the outcome of the reaction. Compound 11 a is closely related to the synthetic substrate 11. The computational selectivities nicely match the experiment. Thus, the absolute configuration of 12 and 13 was assigned as shown above keeping in mind that in the experiment the ent-1 was used as catalyst. Analysis of the transition structures of the REO The consistently high agreement between experimental and computational selectivity allows the proposal of a mechanistic model for the origins of selectivity of the REO based on the structure of the transition states. This is outlined for the differences between the REO of 2 a and 8. For 2 a, the (2R,3S) enantiomer is opened with very high selectivity to radical 4 a*. The (2S,3R) enantiomer is opened with low selectivity (3 a:4 a = 58:42). For 11, this is not ob-
&
&
Chem. Asian J. 2014, 00, 0 – 0
4
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!
www.chemasianj.org
Andreas Gansuer, Frank Neese et al.
opening of meso-epoxides).[20] It should be noted that in TS13 a the steric destabilization is larger than in TS-12 a* since the group containing the secondary hydroxy group (TS-13 a) is bulkier than the n-alkyl chain (TS-12 a*). Thus, in the absence of electronic effects, the REO selectivity is indeed governed by steric effects in agreement with the postulated mechanism for the enantioselective opening of meso-epoxides by 1.[20]
Conclusions In summary, we have demonstrated that epoxides derived from homoallylic alcohols are very attractive substrates for the synthesis of 1,3- and 1,4- diols via the REO. Protected Sharpless epoxides are less suitable for the synthesis of 1,2and 1,3-diols because a combination of reagent and substrate control leads to matched and mismatched cases of epoxide opening. The reasons for the differences, that is, electronic effects influencing the stability of the emerging radical centers in the case of Sharpless epoxides, were identified with the aid of computational investigations. Moreover, the computational results allow an assignment of the absolute configuration of the experimentally obtained products. The transition structures provide valuable clues for the design of more selective catalysts. Investigations in this area are underway, and results will be reported in due course.
Figure 1. Transition states of the REO of the (2S,3R) enantiomer of 2 a (TS-3 a and TS-4 a) and (2R,3S) enantiomer of 2 a (TS-3 a* and TS-4 a*) and their respective energies (COSMO-B3LYP).
Experimental Section General procedure for the REO of substrates 2, 5, and 8 To a mixture of 2,4,6-collidine hydrochloride (1.5 equiv), 1 (0.1 equiv) and Mn (1.5 equiv) in THF (0.1 m with respect to epoxide) were added 1,4-cyclohexadiene (4.4 equiv) and the epoxide (1 equiv). After stirring at room temperature overnight, the reaction was quenched by the addition of H3PO4/KH2PO4 buffer (2:1, 3 mol L 1, pH = 1.9). The mixture was then extracted twice with EtOAc and twice with CH2Cl2. The organic layers were dried over MgSO4 and the volatiles removed in vacuo. The crude product was purified by chromatography on silica gel using the given eluents. Procedure for the REO of substrate 2 According to the general procedure, 2,4,6-collidine hydrochloride (470 mg, 3.0 mmol, 1.5 equiv), Mn (180 mg, 3.27 mmol, 1.7 equiv), (2S*,3R*)-2 (540 mg, 2.0 mmol, 1.0 equiv), and 1,4-cyclohexadiene (700 mg, 8.8 mmol, 4.4 equiv) in THF (5.5 mL) were treated with 1 (103 mg, 0.2 mmol, 0.1 equiv) overnight. Purification (cyclohexane/ EtOAc 90:10) yielded (3R)-3 (159 mg, 29 %) and (2S)-4 (282 mg, 52 %). Procedure for the REO of substrate 5 According to the general procedure, 2,4,6-collidine hydrochloride (472 mg, 3.0 mmol, 1.5 equiv), Mn (166 mg, 3.0 mmol, 1.5 equiv), (2S*,3R*)-5 (462 mg, 2.0 mmol, 1.0 equiv), and 1,4-cyclohexadiene (700 mg, 8.8 mmol, 4.4 equiv) in THF (6 mL) were treated with 1 (106 mg, 0.2 mmol, 0.1 equiv) overnight. Purification (cyclohexane/ EtOAc 96:4) yielded (3S)-6 (112 mg, 29 %) and (2S)-7 (297 mg, 59 %). Procedure for the REO of substrate 8 Figure 2. Transition states of the REO of the (2S,4S,5R) enantiomer of 11 a (TS-12 a and TS-13 a) and (2R,4R,5S) enantiomer of 11 a (TS-12 a* and TS-13 a*) and their respective energies (COSMO-B3LYP).
Chem. Asian J. 2014, 00, 0 – 0
These are not the final page numbers! ÞÞ
According to the general procedure, 2,4,6-collidine hydrochloride (236 mg, 1.5 mmol, 1.5 equiv), Mn (83 mg, 1.5 mmol, 1.5 equiv), (3S*,4R*)-8 (237 mg, 1.03 mmol, 1 equiv), and 1,4-cyclohexadiene
5
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
www.chemasianj.org
(350 mg, 4.4 mmol, 4.4 equiv) in THF (3 mL) were treated with 1 (53 mg, 0.1 mmol, 0.1 equiv) overnight. Purification (cyclohexane/EtOAc 90:10) gave 9 (111 mg, 47 %) and 10 (100 mg, 42 %). Procedure for the REO of substrate 11 To a mixture of 2,6-lutidine hydrochloride (280 mg, 2.0 mmol, 2.0 equiv), ent-1 (106 mg, 0.2 mmol, 0.2 equiv), and Mn (120 mg, 2.2 mmol, 2.2 equiv) in THF (3 mL) was added Bu3SnH (580 mg, 2.0 mmol, 2.0 equiv) and (2S*,4S*,5R*)-11 (166 mg, 1.0 mmol, 1.0 equiv). After stirring at room temperature for 72 h, the reaction was quenched by the addition of phosphate buffer (1 mL) and Et2O (5 mL). The mixture was then filtered through a short column (Et2O, SiO2/K2CO3). Purification (cyclohexane/EtOAc 80:20, SiO2) yielded 12 (80 mg, 48 %) and 13 (68 mg, 40 %).
[7]
[8]
Further details of compound preparation and characterization as well as transition state coordinates and energies are provided in the Supporting Information. [9]
Acknowledgements [10] [11]
We are indebted to the Deutsche Forschungsgemeinschaft (SFB 813 “Chemistry at Spin Centers”) for generous financial support.
[12] [1] Reviews: a) H. B. Kagan, Croat. Chem. Acta 1996, 69, 669 – 680; b) J. Eames, Angew. Chem. 2000, 112, 913 – 916; Angew. Chem. Int. Ed. 2000, 39, 885 – 888; c) H. B. Kagan, Tetrahedron 2001, 57, 2449 – 2468; d) J. R. Dehli, V. Gotor, Chem. Soc. Rev. 2002, 31, 365 – 370; e) E. Vedejs, M. Jure, Angew. Chem. 2005, 117, 4040 – 4069; Angew. Chem. Int. Ed. 2005, 44, 3974 – 4001; f) R. R. Kumar, H. B. Kagan, Adv. Synth. Catal. 2010, 352, 231 – 242; g) L. C. Miller, R. Sarpong, Chem. Soc. Rev. 2011, 40, 4550 – 4562; h) M. Pineschi, V. di Bussolo, P. Crotti, Chirality 2011, 23, 703 – 710. [2] Hydroxylated olefins: a) T. Katsuki, K. B. Sharpless, J. Am. Chem. Soc. 1980, 102, 5974 – 5976; b) R. A. Johnson, K. B. Sharpless in Catalytic Asymmetric Synthesis (Ed.: I. Ojima), Wiley-VCH, New York, 2000, pp. 231 – 280; c) W. Zhang, H. Yamamoto, J. Am. Chem. Soc. 2007, 129, 286 – 287; Unfunctionalized olefins: d) W. Zhang, J. L. Loebach, S. R. Wilson, E. N. Jacobsen, J. Am. Chem. Soc. 1990, 112, 2801 – 2803; e) R. Irie, K. Noda, N. Matsumoto, T. Katsuki, Tetrahedron Lett. 1990, 31, 7345 – 7348; f) Y. Sawada, K. Matsumoto, S. Kondo, H. Watanabe, T. Ozawa, K. Suzuki, B. Saito, T. Katsuki, Angew. Chem. 2006, 118, 3558 – 3560; Angew. Chem. Int. Ed. 2006, 45, 3478 – 3480. [3] a) E. N. Jacobsen, M. Tokunaga, J. F. Larrow, F. Kakiuchi, Science 1997, 277, 936 – 938; b) H. Lebel, E. N. Jacobsen, Tetrahedron Lett. 1999, 40, 7303 – 7306; c) E. N. Jacobsen, Acc. Chem. Res. 2000, 33, 421 – 431. [4] a) K. B. Sharpless, R. C. Michaelson, J. Am. Chem. Soc. 1973, 95, 6136 – 6137; substrate directed reactions: b) A. H. Hoveyda, D. A. Evans, G. C. Fu, Chem. Rev. 1993, 93, 1307 – 1370. [5] a) A. Gansuer, C.-A. Fan, F. Keller, J. Keil, J. Am. Chem. Soc. 2007, 129, 3484 – 3485; b) A. Gansuer, C.-A. Fan, F. Keller, P. Karbaum, Chem. Eur. J. 2007, 13, 8084 – 8090; c) A. Gansuer, S. Lei, M. Otte, J. Am. Chem. Soc. 2010, 132, 11858 – 11859; d) A. Gansuer, L. Shi, F. Keller, P. Karbaum, C.-A. Fan, Tetrahedron: Asymmetry 2010, 21, 1361 – 1369. [6] a) A. Gansuer, M. Pierobon, H. Bluhm, Angew. Chem. Int. Ed. 1998, 37, 101 – 103; Angew. Chem. 1998, 110, 107 – 109; b) A. Gansuer, H. Bluhm, M. Pierobon, J. Am. Chem. Soc. 1998, 120, 12849 – 12859; c) A. F. Barrero, A. Rosales, J. M. Cuerva, J. E. Oltra, Org.
&
&
Chem. Asian J. 2014, 00, 0 – 0
[13]
[14]
[15] [16] [17]
[18] [19]
[20]
6
Andreas Gansuer, Frank Neese et al.
Lett. 2003, 5, 1935 – 1938; d) J. Justicia, A. Rosales, E. BuÇuel, J. L. Oller-Lpez, M. Valdivia, A. Haidour, J. E. Oltra, A. F. Barrero, D. J. C rdenas, J. M. Cuerva, Chem. Eur. J. 2004, 10, 1778 – 1788; Reviews: e) A. Gansuer, T. Lauterbach, S. Narayan, Angew. Chem. Int. Ed. 2003, 42, 5556 – 5573; Angew. Chem. 2003, 115, 5714 – 5731; f) J. M. Cuerva, J. Justicia, J. L. Oller-Lpez, J. E. Oltra, Top. Curr. Chem. 2006, 264, 63 – 92; g) A. Gansuer, J. Justicia, C.-A. Fan, D. Worgull, F. Piestert, Top. Curr. Chem. 2007, 279, 25 – 52. a) W. A. Nugent, T. V. RajanBabu, J. Am. Chem. Soc. 1988, 110, 8561 – 8562; b) T. V. RajanBabu, W. A. Nugent, J. Am. Chem. Soc. 1989, 111, 4525 – 4527; c) T. V. RajanBabu, W. A. Nugent, M. S. Beattie, J. Am. Chem. Soc. 1990, 112, 6408 – 6409; d) T. V. RajanBabu, W. A. Nugent, J. Am. Chem. Soc. 1994, 116, 986 – 997. a) E. Cesarotti, H. B. Kagan, R. Goddard, C. Krger, J. Organomet. Chem. 1978, 162, 297 – 309; b) Structure: A. Gansuer, H. Bluhm, M. Pierobon, M. Keller, Organometallics 2001, 20, 914 – 919; c) M. Klahn, P. Arndt, A. Spannenberg, A. Gansuer, U. Rosenthal, Organometallics 2008, 27, 5846 – 5851. a) B. Wu, J. R. Parquette, T. V. RajanBabu, Science 2009, 326, 1662 – 1662; b) R. Webster, C. Bçing, M. Lautens, J. Am. Chem. Soc. 2009, 131, 444 – 445; c) B. Wu, J. C. Gallucci, J. R. Parquette, T. V. RajanBabu, Chem. Sci. 2014, 5, 1102 – 1117. F. Neese, WIREs Comput. Mol. Sci. 2012, 2, 73 – 78. a) A. D. Becke, Phys. Rev. A 1988, 38, 3098 – 3100; b) J. P. Perdew, Phys. Rev. B 1986, 33, 8822 – 8824. a) F. Weigend, R. Ahlrichs, Phys. Chem. Chem. Phys. 2005, 7, 3297 – 3305; b) A. Schfer, H. Horn, R. Ahlrichs, J. Chem. Phys. 1992, 97, 2571 – 2577. a) B. I. Dunlap, J. W. D. Connolly, J. R. Sabin, J. Chem. Phys. 1979, 71, 3396 – 3402; b) E. J. Baerends, D. E. Ellis, P. Ros, Chem. Phys. 1973, 2, 41 – 51; c) K. Eichkorn, O. Treutler, H. hm, M. Hser, R. Ahlrichs, Chem. Phys. Lett. 1995, 242, 652 – 660. a) K. Eichkorn, O. Treutler, H. hm, M. Hser, R. Ahlrichs, Chem. Phys. Lett. 1995, 240, 283 – 289; b) K. Eichkorn, F. Weigend, O. Treutler, R. Ahlrichs, Theor. Chem. Acc. 1997, 97, 119 – 124. a) A. D. Becke, J. Chem. Phys. 1993, 98, 5648 – 5652; b) C. T. Lee, W. T. Yang, R. G. Parr, Phys. Rev. B 1988, 37, 785 – 789. F. Weigend, M. Haser, H. Patzelt, R. Ahlrichs, Chem. Phys. Lett. 1998, 294, 143 – 152. a) F. Neese, F. Wennmohs, A. Hansen, U. Becker, Chem. Phys. 2009, 356, 98 – 109; b) S. Kossmann, F. Neese, Chem. Phys. Lett. 2009, 481, 240 – 243. A. Klamt, G. Schuurmann, J. Chem. Soc. Perkin Trans. 2 1993, 799 – 805. a) K. Daasbjerg, H. Svith, S. Grimme, M. Gerenkamp, C. MckLichtenfeld, A. Gansuer, A. Barchuk, F. Keller, Angew. Chem. Int. Ed. 2006, 45, 2041 – 2044; Angew. Chem. 2006, 118, 2095 – 2098; b) A. Gansuer, A. Barchuk, F. Keller, M. Schmitt, S. Grimme, M. Gerenkamp, C. Mck-Lichtenfeld, K. Daasbjerg, H. Svith, J. Am. Chem. Soc. 2007, 129, 1359 – 1371; c) A. Gansuer, A. Fleckhaus, M. Alejandre Lafont, A. Okkel, K. Kotsis, A. Anoop, F. Neese, J. Am. Chem. Soc. 2009, 131, 16989 – 16999. a) A. Gansuer, T. Lauterbach, H. Bluhm, M. Noltemeyer, Angew. Chem. Int. Ed. 1999, 38, 2909 – 2910; Angew. Chem. 1999, 111, 3112 – 3114; b) A. Gansuer, M. Pierobon, Synlett 2000, 1357 – 1359; c) A. Gansuer, H. Bluhm, T. Lauterbach, Adv. Synth. Catal. 2001, 343, 785 – 787; d) A. Gansuer, H. Bluhm, B. Rinker, S. Narayan, M. Schick, T. Lauterbach, M. Pierobon, Chem. Eur. J. 2003, 9, 531 – 542; e) J. Friedrich, K. Walczak, M. Dolg, F. Piestert, T. Lauterbach, D. Worgull, A. Gansuer, J. Am. Chem. Soc. 2008, 130, 1788 – 1796; f) A. Gansuer, C.-A. Fan, F. Piestert, J. Am. Chem. Soc. 2008, 130, 6916 – 6917. Received: March 3, 2014 Published online: && &&, 0000
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!
FULL PAPER Catalysis
Reasons for the selectivity of the REO unraveled: Through a combination of synthesis and theory, the factors governing the selectivity of the regiodiver-
gent epoxide opening (REO) were resolved. Our study suggests useful new building blocks for the synthesis of 1,3- and 1,4-diol units.
Chem. Asian J. 2014, 00, 0 – 0
These are not the final page numbers! ÞÞ
7
Andreas Gansuer,* Peter Karbaum, David Schmauch, Martin Einig, Lili Shi, Anakuthil Anoop, &&&&—&&&& Frank Neese* Synthetic and Computational Evaluation of Regiodivergent Epoxide Opening for Diol and Polyol Synthesis
2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
&
![Direct synthesis and identification of benzo[a]pyrene diol epoxide-deoxyguanosine binding sites in modified oligodeoxynucleotides.](/assets/img/hold.png)
