A Novel Bis-Thiourea Organocatalyst for the Asymmetric Aza-Henry Reaction

Abstract

A novel bis-thiourea BINAM-based catalyst for the asymmetric aza-Henry reaction has been developed. This catalyst promotes the reaction of N-Boc imines with nitroalkanes to afford β-nitroamines with good yields and high enantioselectivities. This catalyst has the advantage that it can be prepared in a single step from commercially available materials. A model is proposed for the catalyst action where the both components of the reaction are activated simultaneously by hydrogen bonding. Regardless of the mechanism, the success of the present catalyst demonstrates the potential of bis-thioureas as an interesting class of relatively unexplored catalysts.

The nucleophilic addition of nitroalkanes to imines, known as the aza-Henry or nitro-Mannich reaction, is a fundamental C-C bond forming reaction in organic chemistry.¹ The resulting 1,2-nitro amines are useful precursors to a variety of nitrogen containing chiral building blocks such as vicinal diamines via reduction of the nitro group² and α-amino carbonyl compounds by means of the Nef reaction.³ Given the importance of this transformation, considerable effort has been directed towards the development of catalytic asymmetric aza-Henry reaction over the past several years. The first pioneering examples were reported by Shibasaki⁴ and Jorgensen⁵ and subsequently other metal-based catalysts have been reported.⁶ Beyond these metal-catalyzed variants, an increasing interest in organocatalysts in the last few years⁷ has led to the appearance of the first examples of enantioselective organocatalytic aza-Henry reactions. Takemoto et. al. have reported that the aza-Henry reaction of N-Boc imines with nitroalkanes can be promoted by a bifunctional asymmetric catalyst bearing both a thiourea function and an N,N-dimethylamino group leading to β-nitroamines with good enantioselectivities,⁸ while shortly thereafter, Yoon and Jacobsen were able to promote the highly stereoselective addition of a range of nitroalkanes to aromatic N-Boc imines using a new thiourea based bifunctional catalyst.⁹ A third type of thiourea organocatalyst was recently reported for this reaction by Ricci¹⁰ and by Schaus¹¹ which involved a thiourea-cinchona alkaloid hybrid. The most recent thiourea catalyst is a hybrid with a chiral oxazoline and is also effective for the aza-Henry reaction.¹² Apart from the thioureas, other organocatalysts have been reported for the aza-Henry reaction which are either based on cinchona alkaloids¹³ or on the triflate salts of bis(amidines).¹⁴

It is known that both imines and nitro compounds will interact with thioureas¹⁵ and yet a bis-thiourea catalyst has not been reported for the aza-Henry reaction. The first chiral bis-thiourea catalyst to be reported was compound 1 by Nagasawa which was based on trans-cyclohexane-1,2-diamine.¹⁶ This catalyst proved effective for the Morita-Baylis-Hillman reaction but not for a number of other reactions.¹⁷ Other bis-thiourea catalysts that have been reported include the interesting guanidine derivative 2 which has three double-hydrogen donors and was developed by Nagasawa for the Henry reaction.¹⁸ Finally, Berkessel has developed the bis-thiourea catalyst 4 for the Morita-Baylis-Hillman reaction¹⁹ and the catalyst 3 for the kinetic resolution of azlactones.²⁰ We report here that the bis-thiourea catalyst 12 based on the 2,2’-diaminobinapthalene (BINAM) chiral scaffold is an effective catalyst for the aza-Henry reaction.²¹

The aza-Henry reaction of the Boc imine 14a and nitromethane were screened with a number of thiourea catalysts and the results are presented in Table 1. This screen includes Nagasawa’s bis-thiourea 1 and the BINAM based thiourea 5 developed by Wang for the Morita-Baylis-Hillman reaction and subsequently found to be effacious for a number of Michael addition reactions.²² The bis-thioureas 6–13 are all preparable in one or two steps from the commercially available (R)-(+)-1,1’-binaphthyl-2,2’-diamine and the details can be found in the supporting information. The bis-thiourea 1 was found to be very effective in promoting the aza-Henry reaction of Boc imine 14a but the asymmetric induction for this reaction was only 8% ee (entry 1). The bifunctional thiourea (R)-5 was not capable of promoting the reaction in the absence of an added base and even in the presence of triethylamine the reaction only went to 50% completion in 12 h and gave 16a in 4% ee. From these results it can be inferred that dimethylamino group in (R)-5 is not sufficiently basic to deprotonate nitromethane. For this reason, the bifunctional catalyst (R)-6 was prepared which has two thiourea functions and two basic functions in the form of imidazoles. As shown by the data in Table 1 this catalyst was capable of promoting the reaction in the absence of triethylamine, but unfortunately, it was not able to provide any significant asymmetric induction (entry 4). The series of six bis-thiourea catalysts 7-12 were then prepared and evaluated and while a few of them were active catalysts only compound (R)-12 with the bis-3,5-trifluoromethylphenyl substituent on the thiourea units was capable of providing significant asymmetric induction (74% ee, entry 10). The macrocyclic bis-thiourea (R,R)-13 was also an active catalyst, but as with most of the catalysts screened for this reaction very low asymmetric induction was observed.

Table 1.

Screen of thioureas catalysts with Boc imine 14a.^a

entry	catalyst	Et3N (equiv)	% conversion b	% ee c
1	(1S,2S)-1	0.4	100	8
2	(R)-5	—	0	—
3	(R)-5	0.4	50	4
4	(R)-6	—	65	5
5	(R)-7	0.4	100	5
6	(R)-8	0.4	10	4
7	(R)-9	0.4	36	0
8	(R)-10	0.4	100	16
9	(R)-11	0.4	73	<5
10	(R)-12	0.4	100	74
11	(R,R)-13	0.4	100	6

^aReactions conducted with 10 equiv CH₃NO₂ and 20 mol % catalyst in toluene at 25 °C for 12 h except entry 10 which was 2 h.

^bDetermined by ¹H NMR spectrum of crude reaction mixture.

^cDetermined by HPLC on Chiralpak OJ-H column.

A brief survey of the reaction conditions for the reaction of imine 14a with nitromethane catalyzed by the chiral thiourea (R)-12 in the presence of 0.4 equiv of Et₃N was conducted and the data in Table 2 reveal the effects of changes in the solvent, temperature and catalyst loading. The least effective solvent was THF which was not unanticipated since the catalyst design was predicated on the assumption that there would be H-bonds involved in the association of the catalyst and substrate. Solvents that are effective H-bond acceptors would be expected to disrupt such an association in the catalyst-substrate complex. Toluene was identifed as the best of the solvents that were examined and lowering the temperature of the reaction lead to an initial significant improvement in the asymmetric induction when the temperature was dropped to -5 °C (82% ee, entry 5) but then only to a small improvement as the temperature was further reduced to −35 °C (86% ee, entry 6). Interestingly, the induction was not increased when the temperature was further lowered to −55 °C and the only effect observed was a decrease in the rate of the reaction (entry 7). Lowering the catalyst loading to 10 mol% at the optimal temperature of −35 °C for the reaction in toluene lead to a dropoff in rate as well as a slight lowering of the asymmetric induction (entries 6 & 11). Thus, the optimal conditions revealed by variations in the conditions for the reaction of imine 14a with nitromethane are those indicated in entry 6 of Table 2.

Table 2.

Optimization of the reaction of imine 14a with thioureas (R)-12.^a

entry	solvent	cat load (mol %)	temp (°C)	% conversion b	% ee c
1	CH2Cl2	20	25	82	59
2	THF	20	25	52	28
3	benzene	20	25	100	62
4	toluene	20	25	100	74
5	toluene	20	−5	100	82
6	toluene	20	−35	83	86
7	toluene	20	−55	50	84
8	mesitylene	20	−35	86	84
9	chlorobenzene	20	−35	93	84
10	toluene	10	−10	100	80
11	toluene	10	−35	53	78
12	mesitylene	10	−10	90	82

^aReactions conducted with 10 equiv CH₃NO₂ in 0.2 M solution in imine 14a with x mol% (R)-12 and 0.4 equiv Et₃N. Reaction time at 25 °C was 24 h and all other temperatures was 36 h.

^bDetermined by ¹H NMR spectrum of crude reaction mixture with Ph₃CH as internal standard.

^c% ee of (R)-16a determined by HPLC on Chiralpak OJ-H column.

With the preparation and screen of the eleven catalysts indicated in Table 1 and the survey of the reaction conditions summarized in Table 2 it was time to review the substrate compatiblity of the reaction with the set of ten different imines and three nitroalkanes presented in Table 3 which serves to bring the reaction scope into focus. The asymmetric induction for these reactions does not appear to be particularly sensitive to the presence of either electron donating or electron withdrawing substituents on the phenyl group of the imine. It was also of note to observe that the reactions of the methoxy substituted imines did not appear to be slower than the imine from benzaldehyde, however, the imines with electron withdrawing groups were more reactive and were complete in shorter reactions times. It was also of interest to find that the reaction will tolerate the presence of a pyridine ring in the imine substrate (entry 10). A decrease in induction was noted for the ortho substituted aryl imines 14d and 14g (entries 4 and 7) and an explanation for this is not exactly clear especially upon consideration that the 1-naphthyl imine 14i does not give an induction that is suppressed relative to that of imine 14a (entries 1 vs. 9). Finally, the asymmetric induction drops from 86% ee for nitromethane to 70% ee for nitroethane with the same imine (entries 1 vs. 11). Interestingly, the drop is much smaller for nitropropane (entry 12). In the last two entries of the Table, the major diastereomer is the syn adduct 16 with a syn:anti ratio of approximately 4:1.

Table 3.

Scope of the aza-Henry reaction of the N-Boc imine 14 catalyzed by the bis-thiourea (R)-12.^a

entry	Imine	Ar	time (h)	R	adduct	% yield b	% ee c
1	14a	C6H5	36	H	16a	55	86
2	14b	4-ClC6H4	15	H	16b	62	85
3	14c	3-ClC6H4	15	H	16c	53	91
4	14d	2-ClC6H4	15	H	16d	61	74
5	14e	4-BrC6H4	24	H	16e	50	78
6	14f	4-MeOC6H4	36	H	16f	50	89
7	14g	2-MeOC6H4	36	H	16g	40	65
8	14h	4-MeC6H4	36	H	16h	48	86
9	14i	1-napthyl	36	H	16i	65	85
10	14j	3-pyridyl	22	H	16j	63	81
11	14a	C6H5	36	Me	16k	59 d	70 e
12	14a	C6H5	36	Et	16l	63 d	80 f

^aReactions conducted with 10 equiv CH₃NO₂ with 0.2 M 14 in toluene at −35 °C with 0.4 equiv Et₃N and 20 mol % (R)-12.

^bIsolated yield after purification by chromatography on silica gel.

^cDetermined by HPLC on Chiralpak OJ-H column.

^dCombined yield of syn and anti adducts.

^e%ee of syn isomer; syn:anti = 77:23.

^f%ee of syn isomer; syn:anti = 80:20.

The absolute stereochemistry of the adducts 16 was determined by comparing the HPLC data and optical rotations of the products with the data reported for those derivatives of 16 that have been previously reported and then the remaining derivatives of 16 were assigned by correlation and the details can be found in the Supporing Information. To account for the absolute sense of the enantioselectivity of the reactions with the bis-thiourea catalyst 12 we propose structure 17 as a model for the catalyst-substrate complex. The proposed structure 17 shown in Figure 2 is generated from (S)-12 and would give the (S)-enantiomer of the adduct 16a. On the right side of structure 17 one of the thiourea groups is interacting with the N-Boc imine via two hydrogen bonds with the nitrogen and oxygen atoms of the imine. On the left side of structure 17 the other thiourea group is interacting with the anion of nitromethane via hydrogen bonds to each of the oxygens of the nitronate anion. This would lead to re-face addition to the imine and the formation of the (S)-enantiomer of 16a which is in accord with the experimental observation that the (R)-enantiomer of 12 gives rise to the (R)-enantiomer of 16a.

Figure 2.

Catalyst-substrate complex 17 as a proposed model for the reaction with catalyst (S)-12 as it is hydrogen bonded to imine 14a and to the anion of CH₃NO₂.

The imine could have been hydrogen-bonded to the catalyst in two different ways. The alternative to the binding mode for the imine shown in Figure 2 is to have the imine rotated 180° such that the oxygen and nitrogen atoms of the imine switch hydrogens on the thiourea. Our preference for the binding mode of the imine shown in Figure 2 is based on the following reasoning. The conformation of the 3,5-bistrifluoromethyl substituted arene ring is known to be co-planar with the thiourea as a result of hydrogen bonding of one of the ortho-hydrogens of the arene with the sulfur of the thiourea.²³ The consequence of this is that there would not be any obvious non-covalent positive interactions between the phenyl of the imine and the bis-3,5-trifluoromethyl phenyl group of the catalyst in the alternative binding mode. In contrast, the binding mode of the imine that is shown in Figure 2 could benefit from CH-π interactions between the edge of the phenyl group of the imine and the face of one of the naphthalene rings of the catalyst. In addition it is likely that the nitromethane is deprotonated after it is hydrogen bonded to the catalyst. Triethylamine is not sufficiently basic enough to completely deprotonate nitromethane since its pKa in DMSO is 17 and in water it is 10. The pKa of nitromethane should be lowered upon binding to the catalyst via two hydrogen bonds. This is consistent with the observation that at −35 °C the reaction only goes to 9% completion for the formation of 16a without catalyst under the conditions in entry 1 in Table 3. Therefore, the catalyst does have a significant effect on the rate of the reaction which may be the result of many factors. The coordination of the imine by the catalyst is supported by ¹H NMR experiments. The signal for the HC=N proton of the amine is shifted from 8.88 ppm in the uncomplexed imine to 8.65 ppm in the presence of two equivalents of the catalyst. The interpretation of this shift is not clear at this point and this will be one of many subjects of future efforts to gain additional experimental evidence towards the understanding of the mechanism of this catalyst in the aza-Henry reaction.

Experimental Section

General procedure for the asymmetric aza-Henry reaction (Table 3)

A flame-dried round bottom flask was loaded with 0.2 equiv. of catalyst 12 (0.2 mmol, 160 mg) and 1 equiv of imine 14a-j (1 mmol). The solid mixture was dissolved in 4 mL of toluene and then RCH₂NO₂ (10 equiv., 0.52 mL) was added at −35 °C. After 5 minutes, Et₃N (0.4 equiv, 56 μL) was added and then the mixture was stirred at −35 °C for 17-36 hours. The volatiles were evaporated and the crude product was purified by column chromatography on silica gel (20% acetone in hexanes) to afford products 16a – 16l.

For all the optimization studies of the aza-Henry reaction detailed in Tables 1 and 2, the experimental procedure is the same with slight modifications that are indicated in each Table. Whenever a % conversion is mentioned, this means the product was not purified, and the % conversion was determined from the ¹H NMR spectrum of the crude reaction mixture by integration of product peaks versus the C(=N)H proton if the imine. The only species observed in the crude ¹H NMR spectrum of the crude reaction mixture was the catalyst, the desired product and the starting imine.

Figure 1.

Chiral thioureas and bis-thiourea organocatalysts.