A General Method for a-Functionalization of Enol Derivatives through the Umpolung Cross-Nucleophile Coupling Mediated by a Single Iodine(III) Reagent

In this paper, we present an unprecedented and general umpolung protocol that allows the functionalization of silyl enol ethers and of 1,3-dicarbonyl compounds with a large range of heteroatomic nucleophiles, including carboxylic acids, alcohols, primary and secondary amines, azide, thiols, and also anionic carbamates derived from CO2. The scope of the reaction also extends to carbon-based nucleophiles. The reaction relies on the use of 1-bromo-3,3-dimethyl-1,3dihydro-1l3[d][1,2]iodaoxole, which provides a key a-brominated carbonyl intermediate. The reaction mechanism has been studied experimentaly and by DFT, and we propose formation of an unusual enolonium intermediate with a halogen-bonded bromide. Introduction The introduction of functional groups at the a carbon of carbonyl compounds is a common transformation in synthetic organic chemistry. a-Functionalized ketones are substructures found in many natural products, pharmaceuticals, and other functional organic compounds.[1] The functionalization reaction relies on the inherent nucleophilicity of the a carbon of the enol (or enolate) derivative of the carbonyl compound, which reacts with an electrophilic reaction partner in this process. Many carbon-based electrophiles can be used, leading to the formation of C-C bonds,[2] but the use of heteroatomic electrophiles becomes challenging. This is due to the high reactivity, and therefore limited functional-group compatibility of these species. They are typically strong oxidants, and this can lead to the formation of by-products, such as overfunctionalized or oxidized compounds.[3] The structural variety of these species is also limited, and it is difficult to reconcile this with the idea of producing structurally diverse target compounds. Nevertheless, there are a number of heteroatomic electrophiles that can be used in such reactions, designed for specific transformations and with specific functional-group tolerances.[4] An alternative approach is to use nucleophiles rather than electrophiles to react with enol derivatives. Iodine(III) compounds have been used in this context to mediate the coupling of the two nucleophilic reactants through twoelectron oxidations, thereby inverting the polarity of one of the reagents.[5] This strategy has recently been termed “cross-nucleophile coupling”.[6] This area has evolved significantly[7] since the first report. [8] When it comes to heteroatomic nucleophiles, the reaction usually requires a Lewis acid, as well as a low reaction temperature, in a one-pot two-step procedure, to avoid formation of by-products (Figure 1a).[9] First the I(III) reagent and the enol derivative react at low temperature to form an enolonium intermediate.[9a] This ensures that the enol nucleophile is consumed before the second nucleophile is added at higher temperature (Figure 1a). In this way, side reactions such as homocoupling of the enol derivative or a-functionalization with other nucleophiles derived from the I(III) reagent (e.g., OAc) are minimized.[9a] Using this protocol, Szpilman et. al elegantly observed O-enolonium species for the first time using 13C NMR spectroscopy.[9a] The formation of by-products is closely related to the outstanding leaving ability of the I(III) functional group (106 better than triflate).[10] In general, the nature of the second nucleophile is somewhat limited, but excellent results have been reported for arylations,[9b, 9c, 9e,11] azidations,[7c] cyanations,[7b] and acetoxylations.[7d,8,12] The somewhat narrow scope is partly due to the fact that the nucleophile may need to be incorporated into the structure of the I(III) reagent.[7b,11,12,13] General methods for the intermolecular reaction of ketones or enol derivatives with a variety of nucleophiles, using a non-designer I(III) reagent, are scarce. The Wirth group developed an effective approach to the formation of nitrogenand oxygen-a-substituted ketones through an internal umpolung strategy mediated by PhI(OAc)2 (Figure 1b).[7a] Their strategy relied on the use of a tethered nucleophile, i.e., the nucleophile was attached to the silicon center of the enol ether substrate (Figure 1b). Importantly, they were able to extend the scope of the reaction to the synthesis of chiral a-substituted ketones when using chiral I(III) reagents. More recently, the Gulder group 2 reported another elegant approach aimed at expanding the range of nucleophiles that can be used in such reactions. Here, 2-pyridyl ketones react by an umpolung coupling process mediated by a λ3-fluoro iodane (Figure 1c).[14] It was proposed that a noncovalent interaction between the F atom in the iodane and the pyridine moiety in the ketone substrate plays a key role in this reaction. Although this reaction is quite limited in terms of the ketone structure, a large number of nucleophiles could be coupled. Considering the ketone component, the majority of reported examples, with the exception of the pyridyl ketones used by Gulder,[14] rely on the use of silyl enol ethers.[7a-c,7e,9a,9b,9e] Our own group contributed to this area of research with an umpolung protocol using allylic alcohols as enol synthons, in a reaction mediated by iridium catalysts (Figure 1d). This method gave a-methoxy ketones from allylic alcohols, or 3(2H)-furanones from carbonyl-functionalized allylic alcohols. For all the examples, 1-fluoro-3,3-dimethyl-1,3-dihydro-1l3-benzo[d][1,2]iodaoxole was used as an oxidant.[15] Figure 1. Strategies for the umpolung a-functionalization of ketones and enol derivatives. a) Cross-nucleophile coupling of silyl enol ethers with nitrogen nucleophiles.[7c] b) Intramolecular cross-nucleophile coupling of nucleophile-functionalized silyl enol ethers.[7a] c) Crossnucleophile coupling of pyridyl ketones.[14] d) Cross-coupling of allylic alcohols with nucleophiles.[15] d) This work: general method for the cross-nucleophile coupling of silyl enol ethers with nucleophiles. In this paper, we report the results of our investigations into the development of a method for the general reaction of unbiased silyl enol ethers (1) with a wide variety of heteronucleophiles, including carboxylic acids, thiols, alcohols, amines, azides, and even CO2 for the first time. Conveniently, in all instances, the same I(III) reagent is used. Thus the synthesis of substrate-specific iodanes is avoided, which contributes to the generality and applicability of the method. The mechanism of the reaction has been studied experimentally and by DFT calculations, and an unusual enolonium intermediate with a halogen-bonded bromide atom is proposed. From this


Introduction
The introduction of functional groups at the a carbon of carbonyl compounds is a common transformation in synthetic organic chemistry. a-Functionalized ketones are substructures found in many natural products, pharmaceuticals, and other functional organic compounds. [1] The functionalization reaction relies on the inherent nucleophilicity of the a carbon of the enol (or enolate) derivative of the carbonyl compound, which reacts with an electrophilic reaction partner in this process. Many carbon-based electrophiles can be used, leading to the formation of C-C bonds, [2] but the use of heteroatomic electrophiles becomes challenging. This is due to the high reactivity, and therefore limited functional-group compatibility of these species. They are typically strong oxidants, and this can lead to the formation of by-products, such as overfunctionalized or oxidized compounds. [3] The structural variety of these species is also limited, and it is difficult to reconcile this with the idea of producing structurally diverse target compounds. Nevertheless, there are a number of heteroatomic electrophiles that can be used in such reactions, designed for specific transformations and with specific functional-group tolerances. [4] An alternative approach is to use nucleophiles rather than electrophiles to react with enol derivatives. Iodine(III) compounds have been used in this context to mediate the coupling of the two nucleophilic reactants through twoelectron oxidations, thereby inverting the polarity of one of the reagents. [5] This strategy has recently been termed "cross-nucleophile coupling". [6] This area has evolved significantly [7] since the first report. [8] When it comes to heteroatomic nucleophiles, the reaction usually requires a Lewis acid, as well as a low reaction temperature, in a one-pot two-step procedure, to avoid formation of by-products (Figure 1a). [9] First the I(III) reagent and the enol derivative react at low temperature to form an enolonium intermediate. [9a] This ensures that the enol nucleophile is consumed before the second nucleophile is added at higher temperature ( Figure 1a). In this way, side reactions such as homocoupling of the enol derivative or a-functionalization with other nucleophiles derived from the I(III) reagent (e.g., OAc) are minimized. [9a] Using this protocol, Szpilman et. al elegantly observed O-enolonium species for the first time using 13 C NMR spectroscopy. [9a] The formation of by-products is closely related to the outstanding leaving ability of the I(III) functional group (10 6 better than triflate). [10] In general, the nature of the second nucleophile is somewhat limited, but excellent results have been reported for arylations, [9b, 9c, 9e,11] azidations, [7c] cyanations, [7b] and acetoxylations. [7d,8,12] The somewhat narrow scope is partly due to the fact that the nucleophile may need to be incorporated into the structure of the I(III) reagent. [7b,11,12,13] General methods for the intermolecular reaction of ketones or enol derivatives with a variety of nucleophiles, using a non-designer I(III) reagent, are scarce. The Wirth group developed an effective approach to the formation of nitrogen-and oxygen-a-substituted ketones through an internal umpolung strategy mediated by PhI(OAc)2 ( Figure  1b). [7a] Their strategy relied on the use of a tethered nucleophile, i.e., the nucleophile was attached to the silicon center of the enol ether substrate (Figure 1b). Importantly, they were able to extend the scope of the reaction to the synthesis of chiral a-substituted ketones when using chiral I(III) reagents. More recently, the Gulder group reported another elegant approach aimed at expanding the range of nucleophiles that can be used in such reactions. Here, 2-pyridyl ketones react by an umpolung coupling process mediated by a λ 3 -fluoro iodane ( Figure  1c). [14] It was proposed that a noncovalent interaction between the F atom in the iodane and the pyridine moiety in the ketone substrate plays a key role in this reaction. Although this reaction is quite limited in terms of the ketone structure, a large number of nucleophiles could be coupled. Considering the ketone component, the majority of reported examples, with the exception of the pyridyl ketones used by Gulder, [14] rely on the use of silyl enol ethers. [7a-c,7e,9a,9b,9e] Our own group contributed to this area of research with an umpolung protocol using allylic alcohols as enol synthons, in a reaction mediated by iridium catalysts (Figure 1d). This method gave a-methoxy ketones from allylic alcohols, or 3(2H)-furanones from carbonyl-functionalized allylic alcohols. For all the examples, 1-fluoro-3,3-dimethyl-1,3-dihydro-1l 3 -benzo[d][1,2]iodaoxole was used as an oxidant. [15] [7a] c) Crossnucleophile coupling of pyridyl ketones. [14] d) Cross-coupling of allylic alcohols with nucleophiles. [15] d) This work: general method for the cross-nucleophile coupling of silyl enol ethers with nucleophiles.
In this paper, we report the results of our investigations into the development of a method for the general reaction of unbiased silyl enol ethers (1) with a wide variety of heteronucleophiles, including carboxylic acids, thiols, alcohols, amines, azides, and even CO2 for the first time. Conveniently, in all instances, the same I(III) reagent is used. Thus the synthesis of substrate-specific iodanes is avoided, which contributes to the generality and applicability of the method. The mechanism of the reaction has been studied experimentally and by DFT calculations, and an unusual enolonium intermediate with a halogen-bonded bromide atom is proposed. From this enolonium, an a-brominated carbonyl intermediate is formed, which is key for the high efficiency and the broad scope of the umpolung reaction.
Further optimizations were carried out with reagent 2c (Table 1). When the number of equiv. of NaH was lowered to 1.5, the yield increased substantially (59%, Table 1, entry 3 vs entries 1-2). Lowering the amount of NaH further, or lowering the amount of amine 3 did not have a significant effect on the yield (Table 1, entries 4-5). On the other hand, with 1.5 equiv. of 2c, a yield of 68% was obtained (Table 1, entry 6). Importantly, when the triisopropylsilyl group (TIPS, 1a-TIPS) was replaced by a tert-butyldimethylsilyl group (1a-TBS, Table 1, entry 7), product 5a was obtained in 77% yield. The less hindered trimethylsilyl group (1a-TMS, Table 1, entry 8) gave a lower yield. We also tested toluene, THF, 2-methyltetrahydrofuran, and acetone as reaction solvents, but the desired product was not observed (Table 1, entry 9). We then applied the optimal conditions (Table 1, entry 7) to a number of TBS-enol ethers (Scheme 2a). With electron-donating groups at the para position of the aryl group at R 1 , the corresponding carbamates 5b and 5c were obtained with high efficiency (80 and 60% yields, respectively). Also, F-substituted silyl enol ether 2d gave 5d in 52% yield.Thiophene 1e gave the corresponding carbamate 5e in a good yield of 64%. The reaction was found to be sensitive to steric effects on R 2 . Thus, with an ethyl group at R 2 , 5f was obtained in a lower yield (49%). For aliphatic tert-butyldimethylsilyl enol ethers 1g and 1h, yields of 51 and 46%, respectively, were obtained. To assess the generality of the reaction, we tested different secondary amines to form the carbamate. Symmetrically and unsymmetrically substituted dialkyl amines reacted smoothly to give good yields (5i-5k, 60-80%). Moreover, pyrrolidine substituted carbamate 5l was also obtained under the reaction conditions in 58% yield. However, carbamates derived from primary amines gave a mixture of unidentified products (not shown), which represents a major limitation of this approach. Significantly, this umpolung strategy is not limited to silyl enol ethers, but could be extended to the use of 1,3-dicarbonyl compounds. A b-ketoester 6a reacted smoothly under the same conditions to give 7a in 82% yield. 1,3-Dicarbonyl compound 6b reacted with carbamate 4 to give 7b in good yield (56%). The less nucleophilic b-amidoester 6c and malonate 6d gave carbamoyl derivatives 7c and 7d in good yields (68 and 78%, respectively).
We went on to examine the generality of the cross-nucleophile coupling of silyl enol ethers with a variety of other nucleophiles (Schemes 2b and 2c). Using both electron-poor and electron-rich benzoic acids as nucleophiles, acarboxylate-carbonyl compounds 8a-8c were obtained in excellent yields (80-98%). When alcohols were tested as nucleophiles under otherwise identical reaction conditions, complex mixtures of unidentified by-products were formed. However, these difficulties were overcome by modifying the protocol; first silyl enol ether 1a was treated with I(III) reagent 2c. This was followed by the addition of the alcohol nucleophile (Method B). Using this procedure, a-phenolate 8d was obtained in 66% yield. Thiols were also well tolerated, and thiophenol gave 8e in 60% yield using the standard procedure described above (i.e., Method A), and in 82% yield using Method B. An alkylsubstituted thiol gave a quantitative yield (8f, Method B). Importantly, primary amines are also well tolerated, and benzylamine derivative 8g was obtained in 67% yield when using Method B. Cyclopropylamine reacted smoothly to give 8h in 40% yield, which could be improved to 60% by using Method B. Piperidine gave a-aminoketone 8i in 53% yield, and morpholine derivative 8j was obtained in 60% yield. Interestingly, carbon nucleophiles such as malonates also reacted smoothly, and 8k was formed in 50% yield. a-Azido carbonyl compound 8l was obtained in 50% isolated yield.   We went on to test a number of natural products and pharmaceuticals as nucleophiles in the reaction with 1a-TBS. These compounds all contained carboxylic acid moieties in their structures (Scheme 2c). A BOC-protected glycine derivative gave 8m in 55% yield. With biotin as the nucleophile, compound 8n was obtained in 44% yield. Acetyl salicylic acid gave 8o in quantitative yield (98% isolated yield), and ibuprofen gave 8p in 87% yield. With the aliphatic oleic acid, 8q was formed in 70% yield.
Next, we focused our attention on studying the mechanism of the reaction. In an attempt to understand the dramatic effect of DMF as the reaction solvent, a variety of other polar aprotic solvents were tested. When the reaction of 1a, CO2, and diethylamine with 2c was run in MeCN, 2-bromo-1-phenylpropan-1-one (9) was detected as the sole product (Scheme 3a). a-Bromo carbonyl compound 9 was also formed from 1a and 2c when the reaction was run in DMF in the absence of the nucleophile, in quantitative yield (Scheme 3b). We found that a-bromo carbonyl derivative 9 reacted with the carbamate anion to give 5a in quantitative yield (see Supporting Information). Therefore, it is reasonable to suggest that the reaction might proceed by umpolung bromination followed by a nucleophilic substitution step with the second nucleophile, in this instance the carbamate generated from CO2. The formation of a-substituted carbonyl derivatives as intermediates that can react with nucleophiles in SN2-type reactions has been previously studied by Maulide [20] and Jorgensen [21] among others, [22] as a way to circumvent the inconvenience of using electrophilic reactants. Therefore, 2c seems to be an unusually mild brominating agent for the bromination of silyl enol ethers and 1,3-dicarbonyl compounds. Other brominating agents such as NBS or Br2 lead to the formation of polybrominated products in their reactions with silyl enol ethers.
[23] We also tested NBS and Br2 (see Scheme S2) under the standard conditions described above, but the desired umpolung products, the a-carbamoyl carbonyl compounds, were not detected in any case.
In the presence of radical scavengers, silyl enol ether 1a gave 5a in yields similar to those obtained in their absence (Scheme 3c and 3d vs Scheme 2), suggesting a non-radical pathway. We then tested the selectivity of the reaction towards the formation of a-bromoketone 9 by adding OH -, as another potential nucleophile, to the reaction mixture (Scheme 3e). It has been shown before that water is able to displace I(III) in the enolonium intermediates, and form a-hydroxy ketones (i.e. 10). However, when adding a large excess of OH -, under otherwise identical reactions conditions, only a-bromoketone 9 was formed in 99% yield.
To further understand the mechanism of the cross-nucleophile-coupling reaction mediated by bromobenzoidoxole 2c, we turned to DFT calculations at the B3LYP [24] and M06 [25] functional levels using the Gaussian 16 software [26] (see Supporting Information for more details). The calculations were carried out using the sum of the energies of enolate species [11] I and iodine(III) reagent 2c as the reference point (G = 0 kcal/mol) of the energy profile ( Figure  2, middle). The enolate can attack the iodine atom either through the oxygen or through the carbon of the enolate, leading to formation of at least three possible low-energy enolonium intermediates (O-enolonium-II (-2.0 kcal/mol), C-enolonium-II (-3.0 kcal/mol), and C-enolonium-II´ (0.2 kcal/mol). In all cases, the formation of the enolate-iodine bond occurs in the position trans to the phenyl-I substituent, inducing an elongation of the Br-I bond. In the starting compound 2c, the Br-I bond is short, i.e., 2.9 Å. This bond length is increased in O-enolonium-II (3.17 Å), and even more so in C-enolonium-II´ (3.20 Å) and C-enolonium-II (3.29 Å). These distances are similar to those found in structures with a halogen bond between iodine and bromine atoms. [27] The three calculated intermediates are almost isoenergetic and they could, therefore, be in equilibrium. The direct reductive ligand coupling of the bromine and enolate fragments of these enolonium intermediates was calculated to be feasible. [27b] These species then undergo a reductive coupling step to give a-bromoketone 9. These Br-I interactions might explain the lack of by-products. As mentioned above, the formation of by-products such as dimers (Nu 2 = 1) or a-hydroxy ketones (Nu 2 = OH -, Scheme 3e) has been a major limitation in similar I(III)mediated umpolung approaches. We have also observed by-product formation when using other I(III) reagents, e.g., with 2a in our previous work, [15] and with 2e in the optimization experiments described in this paper (Scheme 1). Alternatively, pathway C can occur via C-enolonium-III, which is formed by ligand rearrangement around the I(III) center of C-enolonium-II. In pathway C, the bromide atom is not interacting with the I(III) center (Br-I distance is 5.16 Å) of C-enolonium-III. From here, a nucleophilic attack of the Bron the very electrophilic [10] a-C of Cenolonium-III also leads to a-bromoketone 9 (16.2 kcal/mol for isomerization and 15.8 kcal/mol for the SN2 step). The DFT calculations do not show a strong preference for any of the three pathways explored. However, if Cenolonium-III is involved (pathway C), we would expect the highly electrophilic enolonium [10] (C-enolonium-III) to react also with any of the nucleophiles present in the reaction mixture, such as starting enol 1 or HO -(Scheme 3e) to form by-products. The control experiment in Scheme 3b also shows that a-bromoketone 9 is the sole product formed in the absence of the heteronucleophile (Nu 2 ). The absence of by-products may be explained through a reaction pathway in which an intramolecular reaction results in formation of the C-Br bond formation (pathways A and B), rather than through an intermolecular nucleophilic displacement of I(III) on the enolonium intermediates (pathway C).

Conclusion
We have developed an umpolung method for the cross-nucleophile coupling of enol derivatives with a variety of nucleophiles, using a single iodine(III) reagent. The reactions occur with good efficiency, and by-products are not formed. Using the method described here, CO2 has been used, in the form of anionic carbamates, in their reaction with silyl enol ethers for the first time. Additionally, benzoic acids, alcohols, thiols, primary and secondary amines, malonate and an azide have been used as nucleophiles in the transformation. We have also used this approach to derivatize natural products and drugs. Mechanistic evidence and DFT calculations have shown that abromoketones are formed as reaction intermediates. DFT calculations indicate that the mechanism may proceed via enolonium species containing I-Br halogen bonds. This may be key for the generality and high selectivity of the reaction.

Experimental Section
General procedures for the cross-nucleophile-coupling reaction: Method A: A solution of the nucleophile (0.2 mmol, 2 equiv.) in DMF (0.33 mL or 1 mL, 0.33 M or 0.1 M of 1) was added to a vial containing 1 (0.1 mmol, 1 equiv.), NaH (0.15 mmol, 1.5 equiv.), and 2c (0.15 mmol, 1.5 equiv.). The reaction mixture was stirred at room temperature for 18 h. After this time, the mixture was extracted with EtOAc (3 ´ 5 mL). The organic phases were combined and washed with water three times. The organic phase was then dried (MgSO4), and the solvent was evaporated under reduced pressure. The residue was purified by flash column chromatography using silica gel as the stationary phase, eluting with a pentane/EtOAc mixture (5 to 100% EtOAc), to give the desired products.
Method B: DMF (0.5 mL, 0.2 M) was added to a vial containing 1 (0.1 mmol, 1 equiv.) and 2c (0.15 mmol, 1.5 equiv.). The reaction mixture was stirred at room temperature for 18 h. After this time, the formation of 9 was complete. A solution of the nucleophile (0.2 mmol, 2 equiv.) and NaH (0.15 mmol, 1.5 equiv.) in DMF (0.5 mL, to reach 0.1 M of 1) was added to the mixture. The reaction mixture was stirred at room temperature for a further 4 h. After this time, the mixture was extracted with EtOAc (3 ´ 5 mL), and the organic phases were combined and washed with water several times. The organic phase was then dried (MgSO4), and the solvent was evaporated under reduced pressure. The residue was purified by flash column chromatography using silica gel as the stationary phase, eluting with a pentane/EtOAc mixture (5 to 100% EtOAc), to give the desired products.
All other experimental data and characterization is provided in the Supporting Information.