Organocatalytic Asymmetric Aldol Reaction of Arylglyoxals and Hydroxyacetone: Enantioselective Synthesis of 2,3-Dihydroxy-1,4-diones

A highly efficient quinine-derived primary-amine-catalyzed asymmetric aldol addition of hydroxyacetone to arylglyoxals is described. Structurally diverse anti-2,3-dihydroxy-1,4-diones were generated in high yields, with good diastereoselectivities and enantioselectivities.


Optimization of Reaction Conditions
Phenylglyoxal monohydrate (1a) and hydroxyacetone (2a) were selected as model reactants for initial tests, with various natural amino acids, including l-Proline, l-Serine, and l-Threonine, screened as catalysts [22,23] and N,N-dimethyl formamide (DMF) as solvent. Unfortunately, desired product 4a was not obtained (Table 1, entries 1-3). Chiral pyrrolidine derivatives 3a and 3b were also ineffective in this reaction (entries 4 and 5). Next, chiral trans-N,N-dialkyl diaminocyclohexanes 3c-3e were used as catalysts to promote this reaction [24], giving desired product 4a in good yields, with good diastereoselectivities and excellent enantioselectivities (entries 6-8). Catalyst 3f, derived from (1S,2S)-1,2-diphenylethane-1,2-diamine, was also tested, but generated product 4a with poor diastereoselectivity and moderate enantioselectivity (entry 9). The stereoselectivity of 4a was slightly enhanced when quinine-derived primary amine 3g was used as catalyst (84% yield, 94% ee, 89:11 dr; entry 10). Based on these results, chiral primary amine 3g was selected as the best catalyst for further optimization of the reaction conditions. Next, the effect of Brønsted acid additives was studied [25]. When additive 3,5-dinitrobenzoic acid (DNBA) was replaced by a weaker acid, p-nitrobenzoic acid, the reaction was slower and the yield, diastereoselectivity, and enantioselectivity of 4a were decreased slightly (entry 11). Similar results were observed when DNBA was replaced with tosic acid (entry 12). These results indicated that DNBA was the most suitable acid additive for this reaction. Decreasing the reaction temperature clearly enhanced the stereoselectivity. For example, when the reaction was conducted at 0 • C, 4a was obtained with 96% ee and 93:7 dr (entry 13). With the aim to further enhance the experimental outcome of 4a by introducing another hydrogen donor group, catalyst 3h was prepared. However, both the yield and stereoselectivity of 4a were decreased slightly in the reaction promoted by 3h (entry 14). An investigation of the catalyst loading indicated that 10 mol% of 3g was sufficiently effective for this reaction (entry 15). Further decreasing the catalyst loading greatly diminished the reaction outcome (entry 16). Various solvents were examined using catalyst 3g, with the results indicating that CHCl 3 was the best solvent in terms of yield, diastereoselectivity, and enantioselectivity. Based on these results, the reaction conditions depicted in entry 15 were selected as optimal for further substrate scope investigations.

Substrate Scope Study
Under the optimized reaction conditions, anti-selective aldol reactions of hydroxyacetone 2a with various phenylglyoxals 1b-1p were examined ( Figure 1). In addition to phenylglyoxal monohydrate 1a, substituted arylglyoxal monohydrates were found to be good reaction partners in this transformation. The electronic properties and positions of substituents on the phenyl ring had almost no influence on the reactivity and stereoselectivity. For example, arylglyoxals bearing 3-Cl-, Br-, or MeO-substituted phenyl groups reacted with hydroxyacetone smoothly to give the corresponding 2,3-dihydroxy-1,4-diketone products 4b-4m in high yields (82-92%) with high enantioselectivities (86-93% ee). For arylglyoxals containing para-substituted phenyl groups, both

Scale-Up Experiment and Crystal Structure of Compound 4j
Notably, this reaction was successfully conducted on a 2-mmol scale (Scheme 2), with product 4j obtained in 86% yield with 95% ee and 89:11 dr. The relative and absolute configurations of 4j (2S,3R) were determined by X-ray crystallography (see the Supplementary Materials) [26]. The stereochemistry of the other aldol products was assigned by comparison with 4j.
Molecules 2020, 25, x 5 of 10 Notably, this reaction was successfully conducted on a 2-mmol scale (Scheme 2), with product 4j obtained in 86% yield with 95% ee and 89:11 dr. The relative and absolute configurations of 4j (2S,3R) were determined by X-ray crystallography (see the Supplementary Materials) [26]. The stereochemistry of the other aldol products was assigned by comparison with 4j. Scheme 2. Scaled-up 2-mmol reaction.

Plausible Reaction Mechanism and Transition States
A proposed catalytic cycle and transition state model [27]

General Information
Unless otherwise noted, commercial reagents were used as received. All reactions were monitored by TLC with silica gel coated plates. 1 H-NMR (600 MHz) and 13 C-NMR (150 MHz) spectra

Plausible Reaction Mechanism and Transition States
A proposed catalytic cycle and transition state model [27]  Notably, this reaction was successfully conducted on a 2-mmol scale (Scheme 2), with product 4j obtained in 86% yield with 95% ee and 89:11 dr. The relative and absolute configurations of 4j (2S,3R) were determined by X-ray crystallography (see the Supplementary Materials) [26]. The stereochemistry of the other aldol products was assigned by comparison with 4j. Scheme 2. Scaled-up 2-mmol reaction.

Plausible Reaction Mechanism and Transition States
A proposed catalytic cycle and transition state model [27]

General Information
Unless otherwise noted, commercial reagents were used as received. All reactions were monitored by TLC with silica gel coated plates. 1 H-NMR (600 MHz) and 13 C-NMR (150 MHz) spectra