Catalytic enantioselective arylative cyclizations of alkynyl 1,3-diketones by 1,4-rhodium(I) migration

Abstract

The enantioselective synthesis of densely functionalized polycarbocycles by the rhodium(I)-catalyzed reaction of arylboronic acids with 1,3-diketones is described. The key step in these desymmetrizing domino addition–cyclization reactions is an alkenyl-to-aryl 1,4-Rh(I) migration, which enables arylboronic acids to function effectively as 1,2-dimetalloarene surrogates.

---

The functionalization of remote C–H bonds offers a powerful method to develop new synthetic methods and achieve transformations that would otherwise be highly challenging.¹ Within this field, 1,4-migration of rhodium(I) between two carbon centers^2–4 has proven to be highly effective for the catalytic functionalization of remote C–H bonds, which has been used to impressive effect in a range of valuable synthetic methods.⁴

We have described catalytic arylative cyclizations from the reaction of alkynyl ketones with arylboronic acids, which produce densely functionalized polycarbocycles through a key step involving an alkenyl-to-aryl 1,4-metal migration (Scheme 1A).⁵ This desymmetrization reaction forms two new carbon–carbon bonds with complete diastereocontrol over two new stereocenters and a trisubstituted alkene. In the non-enantioselective variant of this process, rhodium(I) catalysis was only moderately successful because of the formation of significant quantities of side-products, and the highest yields were obtained using iridium(I) catalysis.^5,6 Although preliminary attempts towards an enantioselective variant using chiral bisphosphine–iridium complexes successfully gave products in high enantioselectivities, only modest catalytic activities were observed.⁵ Furthermore, only cyclic ketones were employed in that study.⁵ Yan and Yoshikai have reported related cobalt-catalyzed arylative cyclizations of acyclic 1,3-diketones with diarylzinc reagents; however, enantioselective reactions were not described (Scheme 1B).⁷ Therefore, to increase synthetic utility, there remains a need to discover more effective chiral catalysts that address these limitations by promoting high-yielding and highly enantioselective arylative cyclizations of a wider range substrates,⁸ including acyclic 1,3-diketones. Here, we report that a chiral bisphosphine–rhodium complex promotes the diastereo- and enantioselective reaction of arylboronic acids with alkynyl 1,3-diketones, for which both acyclic and cyclic 1,3-diketones are effective substrates.

Scheme 1 Catalytic arylative cyclizations involving 1,4-metal migration.

Our experiments began with the arylative cyclization of alkynyl 1,3-diketone 1a with PhB(OH)₂ (eqn (1) and Table 1). Application of conditions identical to those described in our

previous study,⁵ using an iridium–(R)-difluorphos complex, led to no reaction and only unreacted starting material was recovered (eqn (1)). Pleasingly, however, reaction of 1a with PhB(OH)₂ (1.5 equiv.) in the presence of [Rh(C₂H₄)₂Cl]₂ (5 mol%), (R)-BINAP (L1, 10 mol%), and KF (1.5 equiv.) in THF/H₂O (9 : 1) at 70 °C for 24 h gave arylative cyclization product ent-2a in 54% yield (as determined by ¹H NMR analysis using 1,4-dimethoxybenzene as an internal standard) as a single diastereomer (>19 : 1 dr) in 80% ee (Table 1, entry 1).⁹ Higher enantioselectivity was obtained using (S)-DTBM-SEGPHOS (L2), which gave 2a in 52% NMR yield and 91% ee (entry 2). Changing the protic additive from H₂O to t-AmOH (1.5 equiv.) further increased the enantioselectivity (entry 3). The yield of 2a was increased further by raising the temperature to 80 °C (entry 4) and using 2.0 equivalents of PhB(OH)₂ (entry 5). Conducting the reaction on a larger scale using 0.30 mmol of 1a gave 2a in 78% yield and 98% ee (entry 6). This experiment also gave a side-product 3a in 5% yield.¹⁰ It should be noted that the use of PhB(OH)₂ free from triphenylboroxine is very important for good results, as otherwise lower enantioselectivities are observed.¹¹ Finally, repeating the conditions of entry 5 but using [Ir(coe)₂Cl]₂ in place of [Rh(C₂H₄)Cl₂] led to no reaction, and only unreacted starting material was recovered (entry 7).

Table 1 Evaluation of reaction conditions^a

Entry	Ligand	Additive	Solvent	Temp. (°C)	Yield^b (%)	ee^c (%)
a Reactions were conducted with 0.05 mmol of 1a in 1 mL of solvent. b Determined by ^1H NMR analysis using 1,4-dimethoxybenzene as an internal standard. c Determined by HPLC analysis on a chiral stationary phase. d The major enantiomer was ent-2a. e Using 2.0 equivalents of PhB(OH)2. f Using 0.30 mmol of 1a in THF (6 mL). g Value in parentheses refers to the yield of side-product 3a, which was also isolated from this experiment. h Using [Ir(coe)Cl2]2 in place of [Rh(C2H4)2Cl2]. i n.r. = no reaction.
1	L1	—	THF : H2O (9 : 1)	70	54	−80^d
2	L2	—	THF : H2O (9 : 1)	70	52	91
3	L2	t-AmOH	THF	70	54	95
4	L2	t-AmOH	THF	80	67	96
5^e	L2	t-AmOH	THF	80	70	96
6^e^,^f	L2	t-AmOH	THF	80	78 (5)^g	98
7^e^,^h	L2	t-AmOH	THF	80	n.r.^i	—

---

With effective conditions identified (Table 1, entry 6), the scope of this process with respect to the alkynyl acyclic 1,3-diketone 1 was investigated in reactions with PhB(OH)₂ (Table 2). Arylative cyclization products 2a–2r were obtained as single observable diastereomers (>19 : 1 dr as determined by ¹H NMR analysis of the crude reaction mixtures) in 27–82% yield and 56–99% ee. Side-products analogous to 3a (see Table 1) were generally detected but not isolated. Changing the α-substituent R² between the two ketones from methyl (2a) to ethyl (2b), n-butyl (2c), benzyl (2d), or 4-methoxybenzyl (2e) is tolerated. The low yield of 2c results from a low conversion as significant unreacted starting material was observed. The process is also compatible with a range (hetero)aryl groups Ar¹ at the alkynyl position, such as 4-substituted phenyl (2f–2i), 2-fluorophenyl (2j), 3,5-dimethylphenyl (2k), 3,4-(methylenedioxy)phenyl (2l), 2-naphthyl (2m), 1-naphthyl (2n), 2-thienyl (2o), and 2-pyridyl (2p) groups. Finally, the ketone substituents can also be varied from methyl to ethyl (2q) or phenyl groups (2r), although the enantioselectivity dropped substantially in the latter case.

Table 2 Evaluation of alkynyl acyclic 1,3-diketones 1^a

a Reactions were conducted with 0.30 mmol of 1 in THF (6 mL). Yields are of isolated products. Enantiomeric excesses were determined by HPLC analysis on a chiral stationary phase. PMB = para-methoxybenzyl.

---

The process is not limited to the use of PhB(OH)₂, as shown by the arylative cyclizations of 1a with different arylboronic acids to give 2s–2x in 88–96% ee (Table 3). Various 4-substituted phenylboronic acids containing methyl (2s), methoxy (2t), fluoro (2u), or chloro groups (2v) reacted successfully. When 3-methylphenylboronic acid was employed, 1,4-Rh(I) migration occurred to the least sterically hindered site, para to the methyl group (2x). However, 2-methylphenylboronic acid did not provide any of the arylative cyclization product 2y, and returned mainly unreacted starting material along with what appeared to be small quantities of alkyne hydroarylation products.

Table 3 Evaluation of arylboronic acids with alkynyl acyclic 1,3-diketone 1a^a

a Reactions were conducted with 0.30 mmol of 1a in THF (3 mL). Yields are of isolated products. Enantiomeric excesses were determined by HPLC analysis on a chiral stationary phase. b Unreacted starting material was returned along with a trace of what appeared to be alkyne hydroarylation products.

---

Our attention then turned to the reaction of alkynyl cyclic 1,3-diketones 4, substrates employed in our prior study using iridium catalysis (Tables 4 and 5).⁵ With toluene as the solvent, these more reactive substrates generally allowed the use of a decreased catalyst loading of 5 mol% and a lower temperature of 50 °C. Furthermore, in most cases, acceptable results were obtained using only 1.5 equivalents of the arylboronic acid. Various substrates 4a–4f underwent arylative cyclization with PhB(OH)₂ to give products 5a–5f in 39–74% yield and 61–99% ee (Table 4). Small quantities of side-products resulting from arylrhodation of the alkyne with the regioselectivity opposite to that seen in the formation of products 5 were also observed but generally not isolated (see ESI† for details). As with the acylic 1,3-diketones (Table 2), a range of aryl substituents at the alkyne are tolerated, including phenyl (5a and 5f), 4-methoxyphenyl (5b), 4-chlorophenyl (5c), and 3-methylphenyl (5d). The lower yield of 5b results from the formation of products of alkyne hydroarylation without cyclization. The cyclization of a 2-cyanophenyl-containing substrate 4e proceeded smoothly using a 10 mol% catalyst loading but the product 5e was formed in a modest 61% ee. A six-membered cyclic 1,3-diketone also underwent arylative cyclization with PhB(OH)₂ to give 5f in 60% and 97% ee.

Table 4 Evaluation of alkynyl cyclic 1,3-diketones 4^a

a Reactions were conducted with 0.30 mmol of 4 in toluene (3 mL). Yields are of isolated products. Enantiomeric excesses were determined by HPLC analysis on a chiral stationary phase. b Using 5 mol% of [Rh(C₂H₄)₂Cl]₂, 10 mol% of L2. c Using 2.0 equiv. of PhB(OH)₂.

---

Table 5 Evaluation of arylboronic acids with alkynyl cyclic 1,3-diketone 4a^a

a Reactions were conducted with 0.30 mmol of 4 in toluene (3 mL). Yields are of isolated products. Enantiomeric excesses were determined by HPLC analysis on a chiral stationary phase. b Using 5 mol% of [Rh(C₂H₄)₂Cl]₂ and 10 mol% of L2. c Using 2.0 equivalents of the arylboronic acid. d Using 2.4 equivalents of the arylboronic acid. e Unreacted starting material was returned.

---

The scope of the arylative cyclization of alkynyl cyclic 1,3-diketones with respect to the arylboronic acid was then explored in reactions with substrate 4a (Table 5). These reactions proceeded in 44–69% yield and gave products 5g–5p in 95–99% ee. The process tolerates diverse 4-substituted phenylboronic acids containing methyl (5g), halide (5h and 5i), methoxy (5j), acetoxy (5k), or carboethoxy groups (5l). 3-Substituted phenylboronic acids (5m and 5n), 3,4-dimethoxyphenylboronic acid (5o), and 2-naphthylboronic acid (5p) also react effectively. Again, where 1,4-Rh(I) migration could occur to two different positions, migration to the sterically less-hindered side was observed (5m–5p). An attempt to form 5q with 2-methylphenylboronic acid was unsuccessful, and returned only unreacted starting material.

Interestingly, the reaction of substrate 4a with 3-thienylboronic acid gave two products 5ra and 5rb resulting from 1,4-Rh(I) migration to different positions of the thienyl ring before cyclization (eqn (2)).

Finally, this method is not restricted to the use of 1,3-diketone-containing substrates; substrate 6 containing a single methyl ketone also underwent arylative cyclization to give 7 in 94% yield and 85% ee (eqn (3)).

Scheme 2 illustrates a possible catalytic cycle for these reactions, using 1a and PhB(OH)₂ as example substrates. First, upon mixing [Rh(C₂H₄)Cl]₂, L2, KF, and t-AmOH, a chiral complex 8 consisting of one bisphosphine bound to one rhodium atom is formed, which could have a chloride, fluoride, or tert-amyl counterion. Transmetalation of 8 with PhB(OH)₂ gives an arylrhodium species 9, which can then undergo migratory insertion with the alkyne of 1a to give alkenylrhodium intermediate 10. Alkenyl-to-aryl 1,4-rhodium(I) migration of 10 then provides arylrhodium species 11. The relative configuration of products 2 can be explained by a tentative stereochemical model where cyclization proceeds through a conformation similar to 12, in which: (i) rhodium(I) has a square pyramidal coordination geometry; (ii) the ketone undergoing nucleophilic attack is coordinated to rhodium such that the carbonyl group is aligned with the arylrhodium bond to enable subsequent migratory insertion; and (iii) the second ketone is coordinated to rhodium in an axial position. The relative configuration of products 5 (Tables 4 and 5) is more straightforward to rationalize; because of geometric constraints, nucleophilic addition of the arylrhodium group must occur to the same face of the cyclic 1,3-diketone as that from which the tether connecting the two reacting components projects (as in 14 to give representative product 5a, for example). However, as to exactly how the chiral ligand controls the absolute configuration of the products is not clear at the present time.

Scheme 2 Possible catalytic cycle and rationalization of diastereochemical outcomes.

In conclusion, we have reported rhodium(I)-catalyzed arylative cyclizations of alkynyl 1,3-diketones with arylboronic acids, which involve an alkenyl-to-aryl 1,4-Rh(I) migration as a key step. By using a chiral rhodium(I) complex based upon (S)-DTBM-SEGPHOS, the formation of side-products observed previously⁵ with [Rh(cod)Cl]₂ is significantly reduced, and catalytic activity is greatly increased compared with chiral iridium complexes.⁵ These desymmetrization reactions provide densely functionalized polycarbocycles with high diastereo- and enantioselectivities, and notably, both acyclic and cyclic 1,3-ketones are effective substrates.¹²

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

This work was supported by the Engineering and Physical Sciences Research Council [grant number EP/M50810X/1]; the Leverhulme Trust [grant RPG-2016-341]; the Natural Science Foundation of Jiangsu Province [grant number BK20161307]; the Startup Foundation for Introducing Talent of NUIST (grant number 2016r044); the University of Nottingham; and GlaxoSmithKline.

Notes and references

1. For selected reviews of remote functionalization reactions, For selected reviews: (a) R. Breslow, Acc. Chem. Res., 1980, 13, 170–177; (b) H. Schwarz, Acc. Chem. Res., 1989, 22, 282–287; (c) H. Jiang, L. Albrecht and K. A. Jorgensen, Chem. Sci., 2013, 4, 2287–2300; (d) I. Franzoni and C. Mazet, Org. Biomol. Chem., 2014, 12, 233–241; (e) M. Vilches-Herrera, L. Domke and A. Börner, ACS Catal., 2014, 4, 1706–1724; (f) G. Qiu and J. Wu, Org. Chem. Front., 2015, 2, 169–178; (g) A. Vasseur, J. Bruffaerts and I. Marek, Nat. Chem., 2016, 8, 209–219; (h) J. Bruffaerts, D. Pierrot and I. Marek, Org. Biomol. Chem., 2016, 14, 10325–10330; (i) H. Sommer, F. Juliá-Hernández, R. Martin and I. Marek, ACS Cent. Sci., 2018, 4, 153–165.
2. Seminal examples: (a) K. Oguma, M. Miura, T. Satoh and M. Nomura, J. Am. Chem. Soc., 2000, 122, 10464–10465; (b) T. Hayashi, K. Inoue, N. Taniguchi and M. Ogasawara, J. Am. Chem. Soc., 2001, 123, 9918–9919.
3. For an early review, see: S. Ma and Z. Gu, Angew. Chem., Int. Ed., 2005, 44, 7512–7517.
4. Selected, recent examples of 1,4-rhodium(I) migrations: (a) R. Shintani, S. Isobe, M. Takeda and T. Hayashi, Angew. Chem., Int. Ed., 2010, 49, 3795–3798; (b) K. Sasaki, T. Nishimura, R. Shintani, E. A. B. Kantchev and T. Hayashi, Chem. Sci., 2012, 3, 1278–1283; (c) J. Zhang, J.-F. Liu, A. Ugrinov, A. F. X. Pillai, Z.-M. Sun and P. Zhao, J. Am. Chem. Soc., 2013, 135, 17270–17273; (d) R. Shintani, R. Iino and K. Nozaki, J. Am. Chem. Soc., 2014, 136, 7849–7852; (e) H. B. Hepburn and H. W. Lam, Angew. Chem., Int. Ed., 2014, 53, 11605–11610; (f) T. Johnson, K.-L. Choo and M. Lautens, Chem. Eur.–J., 2014, 20, 14194–14197; (g) A. Masarwa, M. Weber and R. Sarpong, J. Am. Chem. Soc., 2015, 137, 6327–6334; (h) A. Claraz, F. Serpier and S. Darses, ACS Catal., 2017, 7, 3410–3413; (i) B. M. Partridge, M. Callingham, W. Lewis and H. W. Lam, Angew. Chem., Int. Ed., 2017, 56, 7227–7232; (j) M. Callingham, B. M. Partridge, W. Lewis and H. W. Lam, Angew. Chem., Int. Ed., 2017, 56, 16352–16356; (k) J. L. Ming and T. Hayashi, Org. Lett., 2018, 20, 6188–6192; (l) J. L. Ming, Q. Shi and T. Hayashi, Chem. Sci., 2018, 9, 7700–7704; (m) S.-S. Zhang, T.-J. Hu, M.-Y. Li, Y.-K. Song, X.-D. Yang, C.-G. Feng and G.-Q. Lin, Angew. Chem., Int. Ed., 2019, 58, 3387–3391; (n) A. Selmani, F. Serpier and S. Darses, J. Org. Chem., 2019, 84, 4566–4574; (o) L. O'Brien, S. N. Karad, W. Lewis and H. W. Lam, Chem. Commun., 2019, 55, 11366–11369; (p) A. Selmani and S. Darses, Org. Lett., 2019, 21, 8122–8126.
5. B. M. Partridge, J. Solana González and H. W. Lam, Angew. Chem., Int. Ed., 2014, 53, 6523–6527.
6. For other examples of 1,4-iridium(I) migration, see ref. 4j and; R. E. Ruscoe, M. Callingham, J. A. Baker, S. E. Korkis and H. W. Lam, Chem. Commun., 2019, 55, 838–841..
7. J. Yan and N. Yoshikai, ACS Catal., 2016, 6, 3738–3742.
8. For selected examples of other types of enantioselective arylative cyclizations of alkynyl electrophiles, see ref. 4h, n–p and: (a) R. Shintani, K. Okamoto, Y. Otomaru, K. Ueyama and T. Hayashi, J. Am. Chem. Soc., 2005, 127, 54–55; (b) T. Miura, M. Shimada and M. Murakami, J. Am. Chem. Soc., 2005, 127, 1094–1095; (c) T. Miura, T. Sasaki, H. Nakazawa and M. Murakami, J. Am. Chem. Soc., 2005, 127, 1390–1391; (d) R. Shintani, A. Tsurusaki, K. Okamoto and T. Hayashi, Angew. Chem., Int. Ed., 2005, 44, 3909; (e) J. Song, Q. Shen, F. Xu and X. Lu, Org. Lett., 2007, 9, 2947–2950; (f) X. Han and X. Lu, Org. Lett., 2010, 12, 108–111; (g) Z.-T. He, B. Tian, Y. Fukui, X. Tong, P. Tian and G.-Q. Lin, Angew. Chem., Int. Ed., 2013, 52, 5314–5318; (h) J. Keilitz, S. G. Newman and M. Lautens, Org. Lett., 2013, 15, 1148–1151; (i) Y. Li and M.-H. Xu, Org. Lett., 2014, 16, 2712–2715; (j) T. Johnson, K.-L. Choo and M. Lautens, Chem. Eur.–J., 2014, 20, 14194–14197; (k) F. Serpier, B. Flamme, J.-L. Brayer, B. Folléas and S. Darses, Org. Lett., 2015, 17, 1720–1723; (l) C. Clarke, C. A. Incerti-Pradillos and H. W. Lam, J. Am. Chem. Soc., 2016, 138, 8068–8071; (m) C. Yap, G. M. J. Lenagh-Snow, S. N. Karad, W. Lewis, L. J. Diorazio and H. W. Lam, Angew. Chem., Int. Ed., 2017, 56, 8216–8220; (n) S. N. Karad, H. Panchal, C. Clarke, W. Lewis and H. W. Lam, Angew. Chem., Int. Ed., 2018, 57, 9122–9125.
9. The relative and absolute configurations of 2j and 5g were determined by X-ray crystallography, and those of the remaining products were assigned by analogy. CCDC 1959877 and 1959878 contain the ESI crystallographic data for this paper.† Furthermore, the spectroscopic data of 2r are consistent with those reported previously (ref. 7), which suggests the relative configurations of products 2 are identical to those reported in ref. 7.
10. Ketone 3a appears to be the result of arylrhodation of the alkyne of 1a with the regioselectivity opposite to that seen in the formation of arylative cyclization product 2a, combined with a retro-Claisen condensation. However, the order of these steps is not currently known.
11. See the ESI† for purification of arylboronic acids.
12. The research data associated with this publication can be found at: http://dx.doi.org/10.17639/nott.7036.

---

Footnotes

† Electronic supplementary information (ESI) available: Experimental procedures, full spectroscopic data for new compounds, and crystallographic data for 2j and 5g. CCDC 1959877 and 1959878. For ESI and crystallographic data in CIF or other electronic format see DOI: 10.1039/c9sc06309a

‡ To whom enquires regarding X-ray crystallography should be addressed.

---