Enantioselective synthesis of hydantoins by chiral acid-catalysed condensation of glyoxals and ureas

Hydantoins are important scaffolds in natural products and pharmaceuticals, with only a few synthetic strategies available for their asymmetric preparation. We herein describe a single-step enantioselective synthesis of 5-monosubstituted hydantoins via condensation of glyoxals and ureas in the presence of a chiral phosphoric acid at room temperature. Products were formed in up to 99% yield and 98 : 2 e.r. Using mechanistic and kinetic studies, including time course 1H NMR monitoring, we revealed that the reaction likely proceeds via face-selective protonation of an enol-type intermediate.


Introduction
5][6] The presence of various vectors for functionalisation of the hydantoin scaffold, along with the sp 3 -hybridised stereocentre, renders them useful for investigation as "3D" fragments in early-stage drug discovery. 7Despite their importance, however, there are relatively few methods available for the asymmetric synthesis of 5-monosubstituted hydantoins from achiral precursors (Fig. 2a-c).
One effective strategy for enantioselective synthesis of 5monosubstituted hydantoins is the asymmetric hydrogenation of prochiral hydantoins 6 bearing exocyclic alkenes at the 5-position (Fig. 2a).This type of approach was rst reported by Takeuchi in 1987, 8 who used a Co catalyst in the presence of an amine ligand to prepare enantioenriched hydantoins 7 in up to 82% ee.More recently, precious metal catalysts have been used to effect the same overall transformation in the presence of chiral phosphine ligands, including Pd 9 (up to: 96% yield; 90% ee), Rh 10 (up to: 99% yield; 97% ee), and Ir 11 (up to: 99% yield; 98% ee).A limitation of this strategy is that it can only deliver hydantoins bearing aliphatic substituents at the 5-position.
Building upon an earlier protocol by Shi, 12 in 2018 Gong reported the enantioselective a-amination of pentauorophenyl esters 8 using diaziridinone 9, mediated by cooperative catalysis between Cu(I) and the chiral benzotetramisole catalyst 10 (Fig. 2b). 13The enantiodetermining step was postulated to proceed via face-selective attack of a urea-derived N-centred radical onto a chiral benzotetramisole-derived enamine intermediate.Some limitations of this approach are the need for pentauorophenyl esters 8 to achieve high ee's, high catalyst (and ligand) loadings, and the use of superheated solvent.
In 2021, Bach introduced an elegant photochemical deracemisation of hydantoins 11 mediated by hydrogen atom transfer in the presence of a chiral diarylketone 12 (Fig. 2c). 14,15g. 1 Examples of bioactive hydantoins.Some constraints of the approach are that it appears limited to hydantoins that bear an irremovable N-phenyl group at the 3position, and aliphatic groups at the 5-position.It also requires a uorinated solvent, and low reactant concentrations.
Given the demand for enantiopure hydantoins, the development of new methods for their asymmetric synthesis from easily accessible achiral precursors remains an important area of research.We envisioned that a new enantioselective synthesis of hydantoins could be achieved via chiral acidcatalysed condensation of glyoxals and ureas (Fig. 2d).
To the best of our knowledge, the acid-mediated condensation of arylglyoxals and ureas was rst reported by Arnold and Möbius in the patent literature in 1970. 16Prior to this, Ekeley and Ronzio had reported that the reaction was only successful under base-mediated conditions. 17Despite being known for >80 years, both the acid-and base-mediated condensations of glyoxals and ureas have received relatively little attention, with <150 reports in SciFinder to date.Signicantly, none of the reported condensations of substituted glyoxals and ureas in the literature are enantioselective.
The acid-mediated condensation of substituted glyoxals 13 and ureas 14 has been suggested 18,19 to occur via a reaction mechanism related to that established for the Biltz hydantoin synthesis from 1,2-diketones and ureas (Fig. 2d, mechanistic possibility 1). 20In this mechanism, glyoxals 13 and ureas 14 would rst react to form vicinal diol intermediates 15, which would then undergo 1,2-hydride migration (presumably in a stereospecic manner). 18,19Vicinal diol intermediates 15 have been isolated and fully characterised previously. 21,22However, we envisioned that an alternative mechanism may be possible (Fig. 2d, mechanistic possibility 2); elimination of the vicinal diol intermediate 15 would afford planar enol-type intermediate 16, which could then undergo protonation to provide the hydantoin product 17.
Regardless of whether the reaction proceeds via mechanistic possibility 1 or 2 (Fig. 2d), we envisioned that use of an appropriate chiral acid could enable an asymmetric condensation to give hydantoins 17, either by (i) controlling face-selective addition of ureas 14 to glyoxals 13 to give enantioenriched diols 15 (followed by stereospecic 1,2-hydride migration); or (ii) by faceselective protonation of enols 16.
Carrying out the reaction using (R)-H 8 -BINOL 20a in the absence of oxygen, and in the dark, provided the targeted  28,29 A 1 H NMR time course experiment over 48 h revealed >99% conversion to the target hydantoin 17a when the reaction was carried out in the absence of oxygen and light; only a marginal trace of 5-hydroxyhydantoin 21a formed (Fig. S1 †).Interestingly, in the presence of oxygen, but the absence of light, the 1 H NMR ratio of 17a:21a aer 48 h was 94 : 6, suggesting that a slow background reaction between triplet oxygen and enol 16a (Table S1, entry 6 †). 28,29ur initial results (entries 1-4) suggested that (i) a H 8 -BINOL backbone, or (ii) large bulky substituents at the 3 and 3 ′ positions, or (iii) both, were needed in order to prepare the hydantoin products in high e.r.To investigate these hypotheses, rst the use of alternate CPAs based on H 8 -BINOL were explored in the reaction (entries 5-11).Notably, H 8 -BINOL catalysts bearing large fused aromatic rings at the 3 and 3 ′ positions, e.g.1naphthyl and 1-pyrenyl rings (entries 9 and 10), provided the product in similar yields to (R)-H 8 -BINOL 20a (entry 4) but did not improve upon the e.r.achieved (81 : 19 and 80 : 20 e.r., respectively, vs. 85 : 15 for 20a).Secondly, the (R)-SPINOL CPA 19b, bearing a 9-anthracenyl substituent at the 6 and 6 ′ positions was explored in the reaction (entry 11).This showed preference for the formation of the (S)-enantiomer of hydantoin 17a, which was formed in 98% yield and 29 : 71 e.r.Since our investigations of the CPA structure did not furnish further improvements to the isolated yield and e.r. of the hydantoin product 17a, we chose to proceed and investigate optimisation of other reaction parameters using the anthracenyl-substituted (R)-H 8 -BINOL 20a catalyst.
We next chose to investigate the reaction solvent, which can markedly inuence the solubility and relative acidity (pK a ) of chiral phosphoric acids. 30Conversion to hydantoin 17a was highest in aprotic solvents (Table 2, entries 1-6), with complete conversion and the highest e.r.'s achieved in chlorinated solvents (entries 5-6), in which the catalyst was fully solubilised (cf.CH 3 CN).Most notably, when the reaction was run in CHCl 3 it proceeded to give hydantoin 17a in 99% yield and 96 : 4 e.r.(entry 6).In contrast, in polar protic EtOH, conversion to hydantoin 17a was sluggish, with cis-diol 15a being the major component of the reaction mixture at 48 h (entry 7).
Investigation of the reaction in CHCl 3 using the other front-running catalysts, 20f and 20g, did not lead to improvements in the e.r. of the hydantoin 17a, although the isolated yield essentially remained the same (entries 8-9).Further investigations therefore focused on optimising the  protocol using (R)-H 8 -BINOL 20a as the catalyst.Lowering the catalyst 20a loading to 2.0 mol% led to only slight erosion in e.r.(96 : 4 / 95 : 5), with 99% isolated yield, and the reaction was found to be complete at 20 h (entry 10).Further lowering of the catalyst 20a loading led to unacceptable erosion of the e.r. and conversion in 48 h (entries 11-13).Heating the reaction to 60 °C reduced the reaction time to two hours, expediently providing hydantoin 17a in 99% isolated yield and with slightly diminished e.r.(91 : 9, entry 14).To improve the e.r. of the product 17a, the method in entry 14 may be coupled with recrystallisation (see below).When the reaction temperature was lowered to 0 °C (entry 15), the e.r. of hydantoin product 17a was improved to 97 : 3, but at the expense of both reaction time (incomplete at 72 h) and isolated yield (96%).The presence of drying agents did not alter the enantioselectivity of the process (Table S2, entries 1-6; 8 †).Addition of activated molecular sieves slowed the reaction down (entries 3-6; 8), however, addition of 10 eq.H 2 O had no effect on the yield, e.r., or reaction time (entry 7).
Based on the high yield and e.r. of hydantoin 17a produced when 2.0 mol% (R)-H 8 -BINOL 20a was used at room temperature for 20 h (entry 10), these conditions were chosen to explore the substrate scope.
Substituents at the aryl 3-position were well tolerated in the reaction.3-OMe-phenylglyoxal hemihydrate gave hydantoin 17l in 96% yield and 94 : 5 e.r., while 3-Cl-phenylglyoxal hemihydrate gave 17m in 96% yield and 92 : 8 e.r.Electron decient 3,4-(diuoro)phenylglyoxal monohydrate reacted to give hydantoin 17n in 75% isolated yield and 93 : 7 e.r. at rt, although the reaction was incomplete at 72 h (the ratio of 15n to 17n was 23 : 77 when the crude reaction mixture was analysed by 1 H NMR at 600 MHz).At rt, the reaction only went to completion aer ve days, giving compound 17n in 95% yield and 91 : 9 e.r.However, heating the reaction to 60 °C provided compound 17n in 98% yield and 89 : 11 e.r.aer 4 hours.
Excitingly, preliminary studies with alkylglyoxals demonstrated that the corresponding hydantoins S17y-ac can be prepared enantioselectively, albeit that further optimisation is required in future (see Scheme S1 †).
Brief investigation of the urea component revealed that high yields and reasonable e.r.'s are maintained when 1,3-dimethylurea (/17u) and 1,3-diDMB-urea were used in the protocol (/17v).Interestingly, 1-DMB-urea performed sluggishly in the reaction, but regioselectively provided the 1-protected hydantoin 17w in 35% yield and 69 : 31 e.r.aer 72 h (incomplete conversion).When 1,3-dimethylthiourea was reacted with phenylglyoxal monohydrate at rt, thiohydantoin 17x was isolated in 95% yield and 71 : 29 e.r.When the reaction temperature was dropped to 0 °C, conversion to the thiohydantoin 17x was incomplete aer 72 h (the ratio of cis-diol 15x to thiohydantoin 17x was 11 : 89), and thiohydantoin 17x was isolated in 86% yield and 62 : 38 e.r.Curiously, however, heating the reaction to 60 °C for 0.5 h gave thiohydantoin 17x in 95% yield and 91 : 9 e.r.Time course 1 H NMR studies were used to investigate the kinetics of the condensation reaction in CDCl 3 over 24 hours, both in the presence and absence of catalyst 20a.A global optimisation algorithm was used to determine the two rate constants within the model by maximising the convergence of the modelpredicted reaction outcomes to the experimental data (Fig. 3a, i-iii).In the absence of catalyst 20a, the reaction between phenylglyoxal monohydrate and 1,3-dibenzylurea proceeded to make cis-diol 15a, which barely reacted further (Fig. 3a-i).The rst step was tted as a second-order process, and the intramolecular cyclisation as a rst-order process, to obtain rate constants of 4.9 M −1 h −1 and 0.0020 h −1 , respectively.However, with the addition of 2.0 mol% (R)-H 8 -BINOL 20a, both of these steps were considerably faster (Fig. 3a-ii).Second-order formation of the cis-diol 15a proceeded with a rate of 150 M −1 h −1 .Interestingly, the subsequent formation of hydantoin 17a in the rate-determining step displayed linear (zero-order) behaviour with a tted rate constant of 0.0070 M h −1 .By using the variable time normalisation analysis (VTNA) technique, 31 we graphically tted the kinetic plot for the (R)-H 8 -BINOL 20a-catalysed reaction with an 'articial zero' aer full conversion of the starting materials (from 5 h), and at a known concentration for intermediate 15a.Subsequently, reaction of intermediate 15a to form the corresponding product 17a in the rate-determining step could be tted with increased accuracy (rate constant = 0.0067 M h −1 ; and R 2 = 0.995, Fig. 3a-iii).This behaviour switched to rst-order when 15a was nearly consumed (see Fig. S13 †).][34][35] To gain a deeper understanding of the reaction mechanism, a series of experiments were carried out (Fig. 3b).First, we prepared and isolated racemic cis-diol 15a (see ESI †) and then exposed it to the optimised reaction conditions (Fig. 3b-i).We obtained the corresponding hydantoin 17a in 95 : 5 e.r., which is consistent with the e.r.obtained for the reaction between phenylglyoxal monohydrate and 1,3-dibenzylurea under the same conditions (cf.Scheme 1).
Finally, exposing racemic hydantoin 17a to (R)-H 8 -BINOL 20a for 48 hours resulted in recovery of racemic 17a (Fig. 3b-iii).Based on the above kinetic and experimental observations, we can explain the origin of enantioselectivity in the reaction (R)-H 8 -BINOL 20a catalyses formation of (scalemic) cis-diol 15 then, in the enantiodetermining step, converts cis-diol 15 to enantioenriched hydantoin 17, presumably via face-selective protonation of a transient, planar enol-type intermediate 16 (Fig. 3c).Further support for the reaction mechanism proceeding through a planar enol intermediate 16 is provided by the observed formation of 5hydroxyhydantoin 21a when the reaction is performed in the presence of singlet (and to a lesser extent triplet) oxygen.
To demonstrate the utility of the hydantoin products formed, we briey investigated their synthetic modication to provide various chiral building blocks (Scheme 2).Our attempts to remove the N-benzyl groups via hydrogenation and other conditions were unfortunately unsuccessful (Tables S17 and  S18 †).However, we were able to remove the 2,4-dimethoxybenzyl (DMB) group from N-1 of hydantoin 17v-a using TfOH at −40 °C to give compound 22v-a, albeit with some erosion in e.r.(89 : 11 / 85 : 15).Additionally, we were able to obtain imidazolidine 23a in 95% yield, and with complete retention of e.r., by reduction of enantioenriched hydantoin 17a using LiAlH 4 . 36ubsequent treatment of imidazolidine 23a with hydroxylamine revealed vicinal diamine 24a, a ligand scaffold used in enantioselective metal-catalysed reactions. 37,38

Conclusions
In summary, we have established a new asymmetric synthesis of hydantoins via the chiral phosphoric acid-catalysed condensation of glyoxals and ureas at room temperature.The reaction proceeds in high yields and enantioselectivities using a variety of substituted aryl glyoxals.Mechanistic investigations revealed that the enantiodetermining step likely arises from the faceselective protonation of a transient enol-type intermediate.Further development of this approach, for instance by modular structural variation of the chiral phosphoric acid catalyst, holds great promise for the broad application of this strategy in the enantioselective synthesis of hydantoins.

Fig. 2
Fig. 2 (a-c) State-of-the-art methods for the asymmetric synthesis of hydantoins; (d) chiral acid-catalysed condensation of glyoxals and ureas (this work), including possible mechanisms for enantioinduction (grey arrows/structures).

Table 1
Preliminary investigations and initial optimisations