Asymmetric organocatalytic Michael addition of Meldrum’s acid to nitroalkenes: probing the mechanism of bifunctional thiourea organocatalysts

The asymmetric Michael addition of Meldrum’s acid to nitroalkenes was studied using a novel type of Cinchona alkaloid-based bifunctional thiourea organocatalyst. The functionality of the thiourea catalysts was also probed by preparing and testing thiourea-N -methylated analogues of the well-known bis-(3,5-trifluoromethyl)phenyl-substituted catalyst.


Introduction
Enantioselective carbon-carbon bond formation is one of the lasting challenges in organic chemistry.Though well-established methods and strategies for stereocontrolled bond formation have been developed, the modern chemist should also consider 'green' values and hence avoid metal-based catalysis when possible.Thus asymmetric organocatalysis is in high demand and the subject of rigorous study. 1 A relatively recent addition to the library of organocatalysts is the family of thiourea-based bifunctional catalysts that combine a hydrogen bond-donor with a tertiary amine attached to a chiral scaffold.Thiourea-based bifunctional catalysis has indeed been extensively studied in recent years, with a great number of reports showing its applicability to a wide range of reactions. 2,3The stereochemical potential of Cinchona alkaloids has also been successfully exploited in the elaboration of thiourea-based organocatalysts. 4Other recently reported Cinchona-based H-bond donor catalysts include squaramide 5 and sulfonamide derivatives. 6e recently developed a concise, enantioselective synthesis for (S)-pregabalin (Lyrica ® ). 7he key stereoselective Michael addition of Meldrum's acid 1 to the nitroalkene precursor 2a was performed with a novel 9-amino-(9-deoxy)-epi-quinidine thiourea catalyst 3, which contains the sterically bulky trityl group as a thiourea substituent, instead of the widely used 3,5bis(trifluoromethyl)phenyl group.Moreover, the original Cinchona-derived catalyst 4, described independently and in rapid succession along with its close analogues by the groups of Chen, Sóos, Connon and Dixon, 8 was found to be severely lacking in stereoinduction in the key step of the synthesis, in spite of complete conversion and a reasonable reaction time (eight hours in model studies).The Michael adduct 5a was attained using 3 with an enantiomeric excess of 75%, whereas 4 gave an ee of only 32% (Scheme 1).In this present study we set out to elaborate further the behavior of 3 in Michael addition reactions of Meldrum's acid to nitroalkenes.We also attempted to probe the general reactivity of thiourea-type catalysts by perturbing the bicoordinating ability of the thiourea moiety.

Results and Discussion
We began our study on the activity of 3 by performing solvent and substrate screening.The results for the solvent screen are shown in Table 1.There is little significant difference in enantioselectivity between CH2Cl2, Et2O and THF.MeCN gave a slightly lower selectivity, whereas with DMF asymmetric induction was nearly completely suppressed.This is more or less expected for a solvent that is a good hydrogen bond acceptor, as it would conceivably compete for the donor sites on the catalyst.Chlorobenzene was found to give a slightly higher enantioselectivity, but the incomplete and slow solubility of Meldrum's acid to chlorobenzene led us to complete the substrate screening using CH2Cl2.Meldrum's acid functions as a protected enol equivalent of acetic acid.Easy deprotection and further functionalization makes it an interesting nucleophile. 10However, examples of asymmetric reactions with Meldrum's acid as a nucleophile are scarce, and furthermore the reports have not been encouraging, as was previously shown by Kleczkowska and Sas 11 .They reported poor enantioselectivities in the Michael addition of Meldrum's acid to β-nitrostyrene 2b, when catalyzed by Cinchona alkaloids or other chiral amines such as (-)-sparteine or (-)brucine.Takemoto also reported mediocre enantioselectivity for the well-known Takemoto's catalyst in this reaction. 12In a control experiment, we measured an ee of 31% when catalyst 4 was used.
For catalyst 3, we likewise found that the reaction between Meldrum's acid and aromatic and heteroaromatic nitroalkenes uniformly gave mediocre enantioselectivities with no clear trends based on electronic properties of the aromatic ring, as shown in Table 2. Surprisingly, aliphatic substrates turned out to give slightly higher selectivities, though with ample room for further improvement.This seems to be a reversal of the common tendency of aliphatic nitroalkenes to give poorer stereoselectivity in thiourea-catalyzed nitro-Michael additions.However, the real surprise was the complete inactivity of 3 in the reaction between dimethyl malonate (7) and β-nitrostyrene.After 48 hours, racemic 8 was isolated in 14% yield, presumably the result of a slow background reaction (Scheme 2).With acetylacetone 9 as the nucleophile, a comparable reaction gave 10 in 45% yield with an ee slightly below 50%.For comparison, when catalyzed by 4 the reaction gave complete conversion in eight hours and an ee of 85%.Before addressing the question of this inactivity, the mechanism of bifunctional thiourea catalysts should be discussed briefly.The widely accepted qualitative rationalization presented e.g. by Takemoto et al. assumes the electrophile to be activated via H-bond assisted coordination of the nitro group to the thiourea (Scheme 3).12b The protonated tertiary amine then directs the nucleophile to attack from the less hindered face.In a theoretical study, Pápai and Sóos showed 13 that while calculations do support this, and also explain the preferred stereochemistry of the product, an opposite coordination scheme is not only possible, but in fact energetically slightly more favorable.
Possible coordination schemes of the catalyst-nucleophile-electrophile ternary complex according to Takemoto (left), Soós (middle), and nucleophile deprotonation (right).Takemoto 11b and Jacobsen 14 have also recently published X-ray structures of their respective thiourea catalysts.In both cases the thiourea N-H protons are found in a trans-conformation with respect to the sulfur atom across the N-Csp2 bond.Conversely, in our earlier study the thiourea moiety was found to be in trans/cis-form for a phenylfluorenyl-substituted Cinchona-thiourea catalyst. 7A general survey of published thiourea crystal structures revealed that the trans/cis form is by far preferred in solid state, with less than 20% of published compounds showing trans/trans-conformation. 15 Though solid state structures obviously do not directly correlate with solvated structures, it is possible that the additional demand of energy for finding the optimal conformation would pose problems for catalysts with substituents that force the thiourea moiety to retain its trans/cis-conformation.
The unusually high acidity of Meldrum's acid (pKa 4.97 in H2O, 7.32 in DMSO) 16 shows that it can be deprotonated by an amine alone, as shown by the use of e.g.(-)-sparteine as a chiral base in the Michael addition (see above).The tertiary amine (pKaH 11/9.8 in H2O/DMSO) of the Cinchona-type thiourea catalysts might not be basic enough for the deprotonation of less acidic nucleophiles.Coordination to the thiourea H-bond donor sites is required to acidify the -proton sufficiently for deprotonation to take place.In the case of 3, either insufficient acidity of the thiourea N-H protons and thus their poorer ability to act as H-bond donors, or steric clashes from the bulky trityl substituent could account for the inactivity of the catalyst with regard to dimethyl malonate.
To investigate this, the 'Southern' substituent of the thiourea moiety was changed from trityl to cyclohexyl to give catalyst 11, which (like 3) has a significantly less acidic thiourea moiety than 4, and is sterically less demanding than 3. Catalyzed by 11, the reaction of dimethyl malonate and nitrostyrene gave 8 in 73% yield after 24 hours, with an ee of 73% (Scheme 4).Thus it seems that neither the acidity of thiourea N-Hs nor of the dimethyl malonate -proton is a main factor for the inactivity of 3. The general mechanistic assumptions of bifunctional thiourea catalysts are based on the 1,3dicarbonyl nucleophile deprotonation taking place on the more acidic enol form.There are practically no reports on the enolization constants of malonate diesters, and they are usually found to contain no enol form in solution (an NMR evaluation in CD2Cl2 confirmed this).Hence the tentative conclusion is that the steric bulk of the trityl group disrupts crucial H-bond coordination and  proton acidification between dimethyl malonate and 3, thus strongly suppressing the deprotonation step.On the other hand, acetylacetone is normally enolized to a considerable extent, especially in nonpolar solvents, stabilized by the formation of an intramolecular H-bond. 17This, together with the difference of several pKa units in the respective acidities of acetylacetone and dimethyl malonate  protons (9/13.3 vs. 13/15.9H2O/DMSO) makes the former significantly easier to deprotonate, which most likely accounts for the difference in activities.
A recent report suggested that the bicoordinating ability of the H-bond donor site is not completely essential; according to a study by Lu, 6 the sulfonamide analogues of Cinchonaderived thiourea catalysts are quite able to catalyze the Michael reaction of -ketoesters and nitroalkenes.We set out to test how the perturbation of the bicoordinating ability of the thiourea affects the activity of the catalyst.Based on the results of Lu, our tentative hypothesis was that perhaps one site would be enough, and this site should be closer to the Cinchona backbone to allow for interaction with the tertiary amine.This was examined by modifying the structure of 4 in order to prepare derivatives 12 and 13, in which both H-bond donors of the thiourea group are in turn substituted with methyl groups.
The catalysts were prepared by condensation of the corresponding isothiocyanate and methylated amine (Scheme 5).The free amine 14 was prepared according to a slightly modified method of Brunner. 7,18The methylated catalyst 12 was prepared by first forming the corresponding ethyl carbamate 15 from 14.The carbamate was subjected to LiAlH4 reduction to attain the amine 16, which was subsequently condensed with 3,5-bis(trifluoromethyl)phenyl isothiocyanate.Catalyst 13 was prepared via reversing the original route.The free amine 14 was converted into the corresponding isothiocyanate 17 in good yield. 7The methylated aniline 18 was easily prepared from the commercially available 3,5-bis(trifluoromethyl)aniline via trifluoroacetylation, methylation and hydrolysis, giving a practically pure product in excellent yield. 19The condensation step was found to be somewhat problematic, since 18 is an extremely electron-poor nucleophile.Catalyst 13 was obtained by deprotonating 18 with n-BuLi at -78 °C and then adding the isothiocyanate dissolved in DMF.
We fond that, for both compounds 12 and 13, the catalytic activity was strongly diminished.After 48 hours the product was isolated in 36% and 45% yields, respectively, with practically no asymmetric induction.Hence this would seem to confirm that the bicoordinating ability is crucial for the functioning of thiourea-based catalysts (Scheme 6).There are other reports probing the activity of bifunctional thiourea catalysts by adding, blocking or removing additional hydrogen bond donor sites.Wang reported that catalyst activity in nitro-Michael addition is significantly boosted by addition of H-bond donor sites (19, 20), giving improved yields and enantioselectivities. 20 Conversely, the enantioselectivity is significantly eroded when the additional coordination sites around the thiourea moiety are blocked or removed.In another study, Liang and Ye tested a hybrid Cinchona-thiourea catalyst 21 in which chiral 1,2-diaminocyclohexane was used as the other thiourea substituent, thus incorporating an additional primary amine in the catalyst structure. 21The catalyst was found to be effective in Michael addition of diethyl malonate to both cyclic and open chain enones.Dimethylation of the free amino group completely suppressed the catalytic activity, but the authors did not further address the mechanistic questions this raises.

Conclusions
Chiral bifunctional thiourea organocatalysts are a new and effective way to conduct asymmetric C-C-bond formation.Among the several promising examples are Cinchona-alkaloid derived thiourea catalysts that have been found to be effective in a wide range of reactions.In this study we have elaborated the activity of a novel trityl-substituted 9-amino-(9-deoxy)epi-quinidine thiourea catalyst and the properties of Meldrum's acid as a nucleophile in asymmetric Michael addition.Furthermore, we have probed the functionality of Cinchona-derived thiourea catalysts in general by preparing and testing N-methylated thiourea analogues.Blocking the thiourea N-H-sites with methyl groups led to severe lack of catalytic activity in the Michael-addition of dimethyl malonate to β-nitrostyrene, thus underlining the importance of the bicoordinating Hbond donor functionality for bifunctional thiourea-tertiary amine catalysts.

Experimental Section
General.Reagents were used as purchased from suppliers, unless otherwise indicated.Dry solvents (CH2Cl2, THF, Et2O, MeCN) were obtained with MBraun MB-SPS 800 solvent drying system.Chlorobenzene was distilled from P2O5 on 4 Å molecular sieves.Reactions requiring inert conditions were performed in flame-dried glassware under a positive pressure of argon.The nitroalkenes were synthesized from corresponding aldehyde and nitromethane by Henry reactions (2b-c, 2f-k), followed by elimination in case of aliphatic substrates, or by Knoevenagel condensation (2d-e).Compounds 3, 4, 5a, 6a, 14 and 17 were synthesized according to published methods 7,8 and spectral data was found to correspond with literature.Spectral data for 5b and 5c, 22  General procedure for screening reactions Nitroalkene (0.8 mmol, 100 mol%) was loaded in a small vial equipped with a magnetic stirrer.
Compound 1 (230 mg, 1.6 mmol, 200 mol%) was added, and the mixture was dissolved in CH2Cl2 (0.5 mL).Catalyst 3 (50 mg, 0.08 mmol, 10 mol%) was added, the vial was capped and the mixture was stirred at room temperature until completion.The solvents were evaporated and the crude mixture analyzed by 1 H NMR and used as such in the derivatization step.

General procedure for derivatization
The crude product (100 mg) was loaded in a 5 mL flask and dissolved in dry DMF (2.5 mL).Distilled aniline (0.25 mL) was added and the mixture was heated to 100 °C for 3 hours.The mixture was partitioned between 10 mL EtOAc and 10 mL 1 M HCl, and washed further 3 times with 1 M HCl (à 10 mL).The organic phase was dried over MgSO4, filtered and evaporated.The residue was purified by flash chromatography (EtOAc in hexane) ISSN 1551-7012 Page 217  ARKAT USA, Inc.   (15).To a solution of 14 (715 mg, 2.2 mmol, 100 mol%) in dry THF (10 mL) was added Et3N (330 L, 2.4 mmol, 110 mol%), followed by slow addition of ethyl chloroformate (230 L, 2.4 mmol, 110 mol%).The mixture was stirred at room temperature under Ar for 2 hours.The solvent was partly evaporated, and the mixture was partitioned between H2O (30 mL) and Et2O (30 mL).The aqueous phase was washed with Et2O (3  20 mL).The organic phases were combined and washed with brine, then dried over MgSO4, filtered and evaporated.Purification by flash chromatography gave 16 as a white foam (463 mg, 53%
8c,d Thus Meldrum's acid does not seem to function well with chiral bifunctional catalysts, and a different solution to establishing its use as a nucleophile in asymmetric synthesis awaits discovery (see Ref. 7 for a screening of different thiourea catalysts).

2b Scheme 4. Reaction of dimethyl malonate and nitrostyrene catalyzed by 11.
8,23and 10 24 was found to correspond with literature data.Reactions were monitored by thin layer chromatography using SiO2 (silica gel 60 F254, Merck, coated aluminum plates), and visualizing by UV light or by aqueous KMnO4 or ninhydrin solutions.Flash chromatography was carried out on SiO2 (silica gel 60 F254, 230-400 mesh ASTM, Merck).1Hand13CNMR spectra were recorded with a Bruker Avance 400 ( 1 H: 399.98 MHz; 13 C: 100.59 MHz) spectrometer.Chemical shifts are reported in ppm relative to TMS internal standard ( = 0.00) in CDCl3, or residual solvent signal in MeOD-d4 ( = 3.31) for 1 H NMR spectra.For 13 C NMR spectra, solvent residual peaks ( = 77.0ppm for CDCl3,  = 49.0 ppm for MeOD-d4) were used as internal standards.Abbreviation of multiplicities is as follows: s (singlet), d (doublet), t (triplet), q (quadruplet), m (multiplet).The prefix br is used when the signal is broadened, and app when the signal resolution is not good enough to determine the true multiplicity (e.g.dd becomes t).High-resolution mass spectrometric data was recorded with Waters LCT Premierspectrometer at Helsinki University of Technology.IR spectra were recorded with Perkin-Elmer Spectrum One FTIR instrument.Optical rotations were obtained with a Perkin-Elmer 343 polarimeter ( = 589 nm) using a 1 dm cell.Chiral HPLC analysis was performed with Waters 501 pump and Waters UV 2487 dual absorbance detector.The exact conditions are reported in connection with each analyzed substance.HPLC analyses were performed before crystallization steps to exclude possible additional enantioenrichment.Elemental analyses were performed with Perkin-Elmer (PE) 2400 Series II CHNS/O Analyzer.Melting points were recorded with Stuart SMP3 melting point apparatus in open capillary tubes.

-Methylamino-(9-deoxy)-epi-quinidine (16). Lithium
, 19 mmol, 300 mol%) was added, and the mixture was stirred at 0 °C for 20 minutes.The solvents were evaporated, and the residue was dissolved in acetone (15 mL).Anhydrous K2CO3 (1.77 g, 12.8 mmol, 200 mol%) and iodomethane (1.2 mL, 19 mmol, 300 mol%) were added, and the mixture was heated to mild reflux for 2 hours.The mixture was filtered and evaporated, the residue was dissolved in H2O:MeOH (5 mL : 25 mL), and anhydrous K2CO3 (880 mg, 6.4 mmol, 100 mol%) was added.The mixture was stirred at room temperature for one hour.The solvents were partly evaporated, and the residue partitioned between H2O (25 mL) and CH2Cl2 (30 mL).The aqueous phase was extracted with CH2Cl2 (30 mL).The combined organic phases were dried over MgSO4, filtered and evaporated to give 19 as a yellowish oil (1.41 g, 90%).The product was practically pure with some solvent residues present.No further purification was done since the compound was somewhat volatile at reduced pressures and could be lost in rigorous drying.