A General Catalytic Enantioselective Transfer Hydrogenation Reaction of β,β-Disubstituted Nitroalkenes Promoted by a Simple Organocatalyst

Given its synthetic relevance, the catalytic enantioselective reduction of β,β-disubstituted nitroalkenes has received a great deal of attention. Several bio-, metal-, and organo-catalytic methods have been developed, which however are usually applicable to single classes of nitroalkene substrates. In this paper, we present an account of our previous work on this transformation, which implemented with new disclosures and mechanistic insights results in a very general protocol for nitroalkene reductions. The proposed methodology is characterized by (i) a remarkably broad scope encompassing various nitroalkene classes; (ii) Hantzsch esters as convenient (on a preparative scale) hydrogen surrogates; (iii) a simple and commercially available thiourea as catalyst; (iv) user-friendly procedures. Overall, the proposed protocol gives a practical dimension to the catalytic enantioselective reduction of β,β-disubstituted nitroalkenes, offering a useful and general platform for the preparation of nitroalkanes bearing a stereogenic center at the β-position in a highly enantioenriched form. A transition state model derived from control kinetic experiments combined with literature data is proposed and discussed. This model accounts and justifies the observed experimental results.


Introduction
The enantioselective reduction of pro-chiral β,β-disubstituted nitroalkenes is a powerful synthetic transformation. It provides a straightforward access to optically active nitroalkanes carrying a configurationally stable stereogenic center at the β-position of the nitro group. These compounds can be easily converted to broadly useful chiral building blocks (e.g., enantioenriched β-chiral amines) exploiting the renowned synthetic versatility of the nitro moiety [1][2][3]. Accordingly, this reaction has received considerable attention from the synthetic chemistry community, leading to several efficient protocols based on different catalytic approaches and encompassing various nitroalkene substrates 1-4, providing access to enantioenriched nitroalkanes 6-9 (Scheme 1).

Optimization and Development of the Transfer Hydrogenation Reaction with β-Trifluoromethylnitroalkenes 4
We started our investigation by studying the H-bond driven transfer hydrogenation reaction of β-trifluoromethyl nitroalkene 4a with Hantzsch ester 10a (Scheme 4) testing various typical H-bond donor catalysts (thioureas, ureas, squaramides, diols, phosphoric acids, etc.). Even if most catalysts 5 were able to afford the desired β-trifluoromethyl nitroalkane 9a with good conversions (Table 1, entries 1-7) only the 1,2-diaminocyclohexane derived amido-thiourea 5o gave a promising enantioselectivity in the reaction (entry 8). Accordingly, variations in the two thiourea N-substituents were undertaken exploiting the modularity of this catalyst structure.
Even if most catalysts 5 were able to afford the desired β-trifluoromethyl nitroalkane 9a with good conversions ( Table 1, entries 1-7) only the 1,2-diaminocyclohexane derived amido-thiourea 5o gave a promising enantioselectivity in the reaction (entry 8). Accordingly, variations in the two thiourea N-substituents were undertaken exploiting the modularity of this catalyst structure.

Optimization and Development of the Transfer Hydrogenation Reaction with β-Trifluoromethylnitroalkenes 4
We started our investigation by studying the H-bond driven transfer hydrogenation reaction of β-trifluoromethyl nitroalkene 4a with Hantzsch ester 10a (Scheme 4) testing various typical H-bond donor catalysts (thioureas, ureas, squaramides, diols, phosphoric acids, etc.). Scheme 4. Test transfer hydrogenation reaction used to identify an efficient catalyst 5.
Even if most catalysts 5 were able to afford the desired β-trifluoromethyl nitroalkane 9a with good conversions (Table 1, entries 1-7) only the 1,2-diaminocyclohexane derived amido-thiourea 5o gave a promising enantioselectivity in the reaction (entry 8). Accordingly, variations in the two thiourea N-substituents were undertaken exploiting the modularity of this catalyst structure.
Catalyst 5p (entry 9) bearing the same 1,2-diaminocyclohexane moiety as 5o afforded the product with lower conversion and very poor and opposite selectivity, showing that enantioselectivity is driven by the amide portion of the catalyst. Indeed, the simpler amido-thiourea catalyst 5g (entry 10) afforded results even better than 5o.

Entry Catalyst 5 Conversion 2 (%)
Whereas the utility of a 1,2-cyclohexanediamine derived catalysts (i.e., 5a, Scheme 2) for the asymmetric reactions of nitroalkenes with Hantzsch esters had been previously reported [32,33], the notable outcome of the simpler structure 5g was an unanticipated yet gratifying result. Having identified catalyst 5g as the optimal, a screening of reaction conditions and Hantzsch esters 10 was undertaken, performing the reactions under more concentrated conditions and at 40 °C, to ensure full conversion ( Table 2). The methyl, iso-butyl, and tert-butyl Hantzsch esters 10b-d were tested and compared with 10a (entries 1-4). The tert-butyl derivative 10d outclassed the other dihydropyridines 10a-c in term of enantioinduction, affording 9a in 89% ee and for this reason it was selected for further exploration. The superior behavior the tert-butyl Hantzsch ester 10d compared to other esters, a common feature to many organocatalytic reactions, had been previously rationalized invoking not only increased nonbonding interactions between the Hantzsch ester ring and the catalyst-substrate complex, but also electronic factors related to the boat-like conformation of the tert-butyl Hantzsch

Entry
Catalyst 5 Conversion 2 (%) ee 3 (%) Catalyst 5p (entry 9) bearing the same 1,2-diaminocyclohexane moiety as 5o afforded the product with lower conversion and very poor and opposite selectivity, showing that enantioselectivity is driven by the amide portion of the catalyst. Indeed, the simpler amido-thiourea catalyst 5g (entry 10) afforded results even better than 5o.
Whereas the utility of a 1,2-cyclohexanediamine derived catalysts (i.e., 5a, Scheme 2) for the asymmetric reactions of nitroalkenes with Hantzsch esters had been previously reported [32,33], the notable outcome of the simpler structure 5g was an unanticipated yet gratifying result. Having identified catalyst 5g as the optimal, a screening of reaction conditions and Hantzsch esters 10 was undertaken, performing the reactions under more concentrated conditions and at 40˝C, to ensure full conversion ( Table 2). The methyl, iso-butyl, and tert-butyl Hantzsch esters 10b-d were tested and compared with 10a (entries 1-4). The tert-butyl derivative 10d outclassed the other dihydropyridines 10a-c in term of enantioinduction, affording 9a in 89% ee and for this reason it was selected for further exploration. The superior behavior the tert-butyl Hantzsch ester 10d compared to other esters, a common feature to many organocatalytic reactions, had been previously rationalized invoking not only increased nonbonding interactions between the Hantzsch ester ring and the catalyst-substrate complex, but also electronic factors related to the boat-like conformation of the tert-butyl Hantzsch ester 10d [23]. However, it is also important to underline that tert-butyl ester 10d features improved solubility in apolar solvents, compared to 10a-c. Next, toluene was confirmed as the most appropriate solvent after a short screening (entries 4-7), and a satisfactory 94% ee value was accomplished by cooling the reaction mixture to´20˝C (entry 8). solubility in apolar solvents, compared to 10a-c. Next, toluene was confirmed as the most appropriate solvent after a short screening (entries 4-7), and a satisfactory 94% ee value was accomplished by cooling the reaction mixture to −20 °C (entry 8).
A variation of the amide N-substituents in catalyst 5 was then explored, by using the closely related thioureas 5q and 5r in the transfer hydrogenation reaction, at −20 °C in toluene. The N,N-diethyl derivative 5q performed rather poorly, while the N-methyl-N-benzhydryl amide 5r did not give any improvement compared to 5g (entries 9-10). The use of α,α,α-trifluorotoluene as solvent provided slightly better results than toluene, when catalyst 5g was employed (entry 11 vs. entry 8). Unluckily, a lower catalyst loading gave a small decrease in the enantiomeric excess of the product 9a (entry 12). Reverting to 10 mol % loading, slightly better results were achieved by lowering reaction concentration (entry 13). These conditions, namely catalyst 5g (10 mol%) Hantzsch ester 10d, α,α,α-trifluorotoluene as solvent, and −20 °C as reaction temperature, were applied on a preparative scale (Scheme 5) obtaining product 9a in 94% yield, with a small erosion in the enantioselectivity compared to the optimization scale reaction (95% instead of 97%).

Scheme 5. Optimized reaction on preparative scale.
A study of the scope of the reaction was undertaken (Scheme 6). Similarly to derivative 4a, the reactions with different substrates 4b-h bearing aromatic rings substituted at different positions with  A variation of the amide N-substituents in catalyst 5 was then explored, by using the closely related thioureas 5q and 5r in the transfer hydrogenation reaction, at´20˝C in toluene. The N,N-diethyl derivative 5q performed rather poorly, while the N-methyl-N-benzhydryl amide 5r did not give any improvement compared to 5g (entries 9-10). The use of α,α,α-trifluorotoluene as solvent provided slightly better results than toluene, when catalyst 5g was employed (entry 11 vs. entry 8). Unluckily, a lower catalyst loading gave a small decrease in the enantiomeric excess of the product 9a (entry 12). Reverting to 10 mol % loading, slightly better results were achieved by lowering reaction concentration (entry 13).
A study of the scope of the reaction was undertaken (Scheme 6). Similarly to derivative 4a, the reactions with different substrates 4b-h bearing aromatic rings substituted at different positions with either electron donating or electron withdrawing groups furnished a series of β-aryl-β-trifluoro nitroalkanes 9b-h with excellent results (>70% yield and >93% ee). Very good yields and enantioselectivities (93%-97% ee) can also be achieved when a 2-naphthyl and two heteroaromatic substituents are collocated in substrates 4i-k. Notably, the optimized reaction conditions could also be applied successfully to substrates 4l and 4m bearing aliphatic chains, which furnished the expected adducts 9l and 9m with very good results. The benefit of using α,α,α-trifluorotoluene as reaction medium with respect to toluene was demonstrated by performing few reactions using toluene as the solvent (Scheme 6, results in brackets): in all the examined cases, the results in terms of enantioselectivity were slightly lower. These conditions, namely catalyst 5g (10 mol%) Hantzsch ester 10d, α,α,α-trifluorotoluene as solvent, and −20 °C as reaction temperature, were applied on a preparative scale (Scheme 5) obtaining product 9a in 94% yield, with a small erosion in the enantioselectivity compared to the optimization scale reaction (95% instead of 97%). either electron donating or electron withdrawing groups furnished a series of β-aryl-β-trifluoro nitroalkanes 9b-h with excellent results (>70% yield and >93% ee). Very good yields and enantioselectivities (93%-97% ee) can also be achieved when a 2-naphthyl and two heteroaromatic substituents are collocated in substrates 4i-k. Notably, the optimized reaction conditions could also be applied successfully to substrates 4l and 4m bearing aliphatic chains, which furnished the expected adducts 9l and 9m with very good results. The benefit of using α,α,α-trifluorotoluene as reaction medium with respect to toluene was demonstrated by performing few reactions using toluene as the solvent (Scheme 6, results in brackets): in all the examined cases, the results in terms of enantioselectivity were slightly lower. Scheme 6. Scope of the transfer hydrogenation reaction of β-trifluoromethyl nitroalkenes 4.
To our delight, after carefully adjusting the reaction conditions, thiourea 5g was able to promote the transfer hydrogenation reaction of 2a and 2b in the presence of Hantzsch ester 10d as hydrogen donor, affording the corresponding β-amino nitroalkanes 7a and 7b in 97% and 90% yield respectively, with complete enantioselectivity (ee >99%) regardless of the protecting group installed on the amine function (Scheme 7). Although a higher reaction temperature was required with these substrates 2 compared to the trifluoromethyl derivatives 4, presumably due to stereoelectronic effects rendering amido substrates 2 less reactive than 4, full enantioselectivity was observed even at a lower (5 mol %) catalyst loading. Substrates 1a and 3a had been previously studied and organocatalytic protocols based on thioureas are present in the literature (see Scheme 2); on the contrary, prior to our work [40], β-amino nitroolefines 2 were utilized only in asymmetric hydrosilylation [21] or in metal-catalyzed asymmetric hydrogenations [17][18][19].
To our delight, after carefully adjusting the reaction conditions, thiourea 5g was able to promote the transfer hydrogenation reaction of 2a and 2b in the presence of Hantzsch ester 10d as hydrogen donor, affording the corresponding β-amino nitroalkanes 7a and 7b in 97% and 90% yield respectively, with complete enantioselectivity (ee >99%) regardless of the protecting group installed on the amine function (Scheme 7). Although a higher reaction temperature was required with these substrates 2 compared to the trifluoromethyl derivatives 4, presumably due to stereoelectronic effects rendering amido substrates 2 less reactive than 4, full enantioselectivity was observed even at a lower (5 mol %) catalyst loading. Eager to further verify the capacity of catalyst 5g in inducing enantioselectivity in the reaction of β,β-disubstitued nitroolefin substrates with Hantzsch ester 10d, we moved to investigate substrates 1a and 3a. A short screening of reaction conditions (solvent, dilution, temperature, amount of Hantzsch ester 10d), indicated dichloromethane as a more suitable solvent for these substrates, and 0 and −20 °C as optimal reaction temperatures for 1a and 3a, respectively. Under these newly found conditions, catalyst 5g was indeed able to provide the corresponding nitroalkanes 6a and 8a in high yields and enantioselectivities, as depicted in Scheme 8.

Proposed Reaction Model for the Transfer Hydrogenation Reactions Catalyzed by 5g with Nitroalkenes
Keeping the activation of nitro compounds by thioureas as the preliminary reasonable assumption, we aimed at gaining some insights on the mode of action of catalyst 5g in the reactions. We first Eager to further verify the capacity of catalyst 5g in inducing enantioselectivity in the reaction of β,β-disubstitued nitroolefin substrates with Hantzsch ester 10d, we moved to investigate substrates 1a and 3a. A short screening of reaction conditions (solvent, dilution, temperature, amount of Hantzsch ester 10d), indicated dichloromethane as a more suitable solvent for these substrates, and 0 and´20˝C as optimal reaction temperatures for 1a and 3a, respectively. Under these newly found conditions, catalyst 5g was indeed able to provide the corresponding nitroalkanes 6a and 8a in high yields and enantioselectivities, as depicted in Scheme 8. Substrates 1a and 3a had been previously studied and organocatalytic protocols based on thioureas are present in the literature (see Scheme 2); on the contrary, prior to our work [40], β-amino nitroolefines 2 were utilized only in asymmetric hydrosilylation [21] or in metal-catalyzed asymmetric hydrogenations [17][18][19].
To our delight, after carefully adjusting the reaction conditions, thiourea 5g was able to promote the transfer hydrogenation reaction of 2a and 2b in the presence of Hantzsch ester 10d as hydrogen donor, affording the corresponding β-amino nitroalkanes 7a and 7b in 97% and 90% yield respectively, with complete enantioselectivity (ee >99%) regardless of the protecting group installed on the amine function (Scheme 7). Although a higher reaction temperature was required with these substrates 2 compared to the trifluoromethyl derivatives 4, presumably due to stereoelectronic effects rendering amido substrates 2 less reactive than 4, full enantioselectivity was observed even at a lower (5 mol %) catalyst loading. Eager to further verify the capacity of catalyst 5g in inducing enantioselectivity in the reaction of β,β-disubstitued nitroolefin substrates with Hantzsch ester 10d, we moved to investigate substrates 1a and 3a. A short screening of reaction conditions (solvent, dilution, temperature, amount of Hantzsch ester 10d), indicated dichloromethane as a more suitable solvent for these substrates, and 0 and −20 °C as optimal reaction temperatures for 1a and 3a, respectively. Under these newly found conditions, catalyst 5g was indeed able to provide the corresponding nitroalkanes 6a and 8a in high yields and enantioselectivities, as depicted in Scheme 8.

Proposed Reaction Model for the Transfer Hydrogenation Reactions Catalyzed by 5g with Nitroalkenes
Keeping the activation of nitro compounds by thioureas as the preliminary reasonable assumption, we aimed at gaining some insights on the mode of action of catalyst 5g in the reactions. We first performed some simple control experiments, comparing in the reaction between 4a and 10d the

Proposed Reaction Model for the Transfer Hydrogenation Reactions Catalyzed by 5g with Nitroalkenes
Keeping the activation of nitro compounds by thioureas as the preliminary reasonable assumption, we aimed at gaining some insights on the mode of action of catalyst 5g in the reactions. We first performed some simple control experiments, comparing in the reaction between 4a and 10d the activity of catalyst 5g vs. simpler achiral thiourea derivatives 5s and 5t (Scheme 9).  These reactions were found to obey pseudo-second order kinetics according to the following equation [46]: where: The straight lines reported in Figure 3 display good correlation coefficients. Their slopes are the rate constant kobs (in M −1 ·h −1 ) for the reactions, from which a more quantitative comparison of the activity displayed by the three catalysts 5g, 5s, and 5t can be deduced. The evolution of the reactions was conveniently monitored by 19 F-NMR spectroscopy. Whereas thiourea 5s should feature a similar acidity to 5g (pK a ca. 12.57 in DMSO [44]), 5t bearing two aryl substituents is considerably more acidic (pK a ca. 8.5 in DMSO [44]). Considering only the coordination of the nitro group by the thiourea moiety acting as a general acid catalyst, and neglecting conformational effects, an increase in acidity should roughly correspond to an increase in catalyst activity [45]. However, as shown in Figure 2, catalyst 5g is considerably more efficient in reaction promotion than 5s and 5t. Furthermore, the less acidic 5s was found to be slightly more effective than 5t.  These reactions were found to obey pseudo-second order kinetics according to the following equation [46]: The straight lines reported in Figure 3 display good correlation coefficients. Their slopes are the rate constant kobs (in M −1 ·h −1 ) for the reactions, from which a more quantitative comparison of the activity displayed by the three catalysts 5g, 5s, and 5t can be deduced. These reactions were found to obey pseudo-second order kinetics according to the following equation [46]: lnp r4as 0ˆp r10ds 0´r 9asq r10ds 0ˆp r4as 0´r 9asq q " pr10ds 0´r 4as 0 qˆk obs t (1) where: The straight lines reported in Figure 3 display good correlation coefficients. Their slopes are the rate constant k obs (in M´1¨h´1) for the reactions, from which a more quantitative comparison of the activity displayed by the three catalysts 5g, 5s, and 5t can be deduced. The straight lines reported in Figure 3 display good correlation coefficients. Their slopes are the rate constant kobs (in M −1 ·h −1 ) for the reactions, from which a more quantitative comparison of the activity displayed by the three catalysts 5g, 5s, and 5t can be deduced.  Together with the previously determined mode of action of catalyst 5g in the Strecker reaction [41,42], these experiments suggest that the amide moiety of this catalyst effectively participates in the coordination/stabilization of a reaction transition state (TS) leading to the major enantiomer of product 9. The pseudo-second order kinetics followed by the reactions (Figure 3) indicates as a first approximation that interactions prior to TS are not relevant (i.e., Curtin-Hammett control). Thus, it can be surmised that, along with the coordination of the Lewis acidic thiourea moiety to the negatively charged nitro moiety, the amide oxygen acts as a Lewis base in this TS, coordinating the N-H proton of the Hantzsch ester having a positive charge density. Applying the previously determined most stable conformation of catalyst 5g [41], these considerations result in the model depicted in Scheme 10. Overall, a dipolar TS is effectively stabilized by the electrostatically complementary functionalities present in the catalyst structure, in a way reminding the mode of action of some macromolecular enzymes [47]. A coordination of both dipole charges (negative at the nitro moiety and positive at the Hantzsch ester) might be more efficient than an exclusive coordination of the negative part of the dipole (nitro moiety), justifying the lower activity offered by "mono-functional" catalysts 5s and 5t. After TS, an irreversible proton-transfer follows, leading to the products 9 and the pyridine co-product. We have previously demonstrated [40], using an α-substituted nitroalkene giving a pro-chiral nitronate intermediate, that catalyst 5g is not able to exert significant stereocontrol in this proton-transfer process [48,49]. Together with the previously determined mode of action of catalyst 5g in the Strecker reaction [41,42], these experiments suggest that the amide moiety of this catalyst effectively participates in the coordination/stabilization of a reaction transition state (TS) leading to the major enantiomer of product 9. The pseudo-second order kinetics followed by the reactions (Figure 3) indicates as a first approximation that interactions prior to TS are not relevant (i.e., Curtin-Hammett control). Thus, it can be surmised that, along with the coordination of the Lewis acidic thiourea moiety to the negatively charged nitro moiety, the amide oxygen acts as a Lewis base in this TS, coordinating the N-H proton of the Hantzsch ester having a positive charge density. Applying the previously determined most stable conformation of catalyst 5g [41], these considerations result in the model depicted in Scheme 10. Overall, a dipolar TS is effectively stabilized by the electrostatically complementary functionalities present in the catalyst structure, in a way reminding the mode of action of some macromolecular enzymes [47]. A coordination of both dipole charges (negative at the nitro moiety and positive at the Hantzsch ester) might be more efficient than an exclusive coordination of the negative part of the dipole (nitro moiety), justifying the lower activity offered by "mono-functional" catalysts 5s and 5t. After TS, an irreversible proton-transfer follows, leading to the products 9 and the pyridine co-product. We have previously demonstrated [40], using an α-substituted nitroalkene giving a pro-chiral nitronate intermediate, that catalyst 5g is not able to exert significant stereocontrol in this proton-transfer process [48,49]. The soundness of this model is confirmed by computational work on the transfer hydrogenation reaction with the related catalyst 5b [34]. In line with the established mode of action of thiourea catalysts in this type of reactions [50], enantioinduction is not due to repulsive interactions between catalysts and substrates, but rather to the good 3D geometrical fit between the polar functionalities of the catalyst and a TS giving the (R)-enantiomer of 9. Apparently, alternative TSs leading to (S)-9 are not matching well the 3D structure of the catalyst, and cannot thus be stabilized/promoted. The The soundness of this model is confirmed by computational work on the transfer hydrogenation reaction with the related catalyst 5b [34]. In line with the established mode of action of thiourea catalysts in this type of reactions [50], enantioinduction is not due to repulsive interactions between catalysts and substrates, but rather to the good 3D geometrical fit between the polar functionalities of the catalyst and a TS giving the (R)-enantiomer of 9. Apparently, alternative TSs leading to (S)-9 are not matching well the 3D structure of the catalyst, and cannot thus be stabilized/promoted. The tert-butyl moiety serves to control the 3D conformation of the catalyst. Furthermore, catalyst/substrates interactions in this TS are limited to the nitro and N-H groups, the nitroalkene β-substituents do not play an obvious role. Indeed, the reaction with nitroalkenes 4 proved to be applicable with very good results to both aromatic and aliphatic substrates (R in Scheme 10). Besides, these considerations justify the remarkable performances of this catalyst with the different nitroalkenes 1-4, as well as the attack of the hydride to the same pro-chiral face [51] of the nitroolefins, irrespective of the nature of the β-substituents (Scheme 11).
Asymmetric transfer hydrogenation β-amino nitroalkanes 2: in a screw cap round bottom vial, to a stirred solution of 2 (0.15 mmol), toluene (510 μL, 0.3 M), catalyst 5g (3.9 mg, 0.0075 mmol, 0.05 equiv), and Hantzsch ester 10d (56 mg, 0.18 mmol, 1.2 equiv) were added. The vial was saturated with nitrogen and closed with the cap. The reaction mixture was stirred for 14 h at 40 °C. The resulting mixture was purified by column chromatography to afford product 7. The spectroscopic data have been Scheme 11. Attack of the hydride to the same pro-chiral face of 1-4 according to the TS model.
Asymmetric transfer hydrogenation β-amino nitroalkanes 2: in a screw cap round bottom vial, to a stirred solution of 2 (0.15 mmol), toluene (510 µL, 0.3 M), catalyst 5g (3.9 mg, 0.0075 mmol, 0.05 equiv), and Hantzsch ester 10d (56 mg, 0.18 mmol, 1.2 equiv) were added. The vial was saturated with nitrogen and closed with the cap. The reaction mixture was stirred for 14 h at 40˝C. The resulting mixture was purified by column chromatography to afford product 7. The spectroscopic data have been previously published [40].

Conclusions
To summarize, we have presented a very general methodology for organocatalytic enantioselective transfer hydrogenation reactions of β,β-disubstituted nitroalkenes, making use of a simple thiourea catalyst, and Hantzsch esters as convenient hydrogen donors. It is the first time that a single catalyst can be used with all nitroalkene substrate classes; previous reports [32][33][34][35][36][37] were all focused on specific catalysts tailored for specific substrates. Some kinetic studies confirmed the generally accepted mode of action of this type of catalysts, allowing us to put forward a reasonable transition state model accounting for the remarkable substrate generality. We envision this synthetic platform might be useful for further synthetic applications and other studies related to thiourea based on catalytic asymmetric transformations [55]. names can be found in the reference section, and to Giacomo Foli for preliminary results with nitroalkenes 1a and 3a. Takashi Ooi is gratefully acknowledged for suggesting to use of a thiourea like 5s in the kinetic studies.