β,β-Dialkyl γ-amino γ-trifluoromethyl alcohols from trifluoromethyl (E)-aldimines by a one-pot solvent-free Mannich-type reaction and subsequent reduction

Abstract

The synthesis of trifluoromethylated γ-amino alcohols through an eco-friendly one-pot self-catalysed Mannich-type reaction between N-protected trifluoromethyl aldimines and suitable cyclic or acyclic α,α-dialkyl aldehydes has been developed. Good yields, mild reaction conditions and simple experimental work-up procedures are some of the advantages of this method. Starting from optically pure trifluoromethyl aldimines, target compounds, also having a quaternary stereocentre, can be obtained in good to excellent diastereoselectivities.

---

Introduction

As is well known, the Mannich reaction is among the most significant carbon–carbon bond forming reactions in synthetic organic chemistry¹ and a standard method for the synthesis of β-amino carbonyl compounds,² key intermediates for the preparation of numerous nitrogen-containing derivatives and pharmaceutical products as well as biologically active compounds. The corresponding γ-amino alcohols are also very useful molecules for their biological properties and their application as ligands in asymmetric syntheses.³

Nevertheless, catalytic asymmetric synthesis of fluorinated molecules, particularly those containing nitrogen functions, is one of the most important challenging topics in pharmaceutical⁴ and materials chemistry,⁵ but only a few papers have considered the Mannich-type approach to obtain fluorinated β-amino carbonyl compounds and derivatives. Starting from suitable N-protected trifluoromethyl aldimines, highly diastereo- and enantioselective organocatalysed Mannich-type reactions were reported, achieving syn-⁶ or anti-γ-amino alcohols⁷ using L-proline or Jørgensen–Hayashi's aryl prolinols as catalysts, respectively. Also starting from a chiral fluorinated oxazolidine, a Mannich-type reaction occurred in good yield and stereoselectivity, involving a chiral fluorinated iminium ion.⁸

Interested in trifluoromethyl aldimine chemistry,⁹ we recently reported an unexpected diastereoselective outcome of solvent-free L-proline-catalysed Mannich-type reaction depending from the reaction temperature.¹⁰ Starting from chiral trifluoromethyl aldimines, diastereomerically pure fluorinated syn- or anti-γ-amino alcohols can be obtained only by changing the reaction temperature (40 or 0 °C, respectively). While a resident stereocentre on aldimine leads only to the facial stereoselective control of nucleophilic attack, the added L-proline allows us to control also the syn/anti stereochemistry through two different transition states, whose stability is a function of temperature.

The study has been subsequently extended to trifluoromethyl aldimines bearing L-α-amino ester moieties. Performing an L-proline catalysed Mannich-type reaction under solvent-free conditions by using isovaleraldehyde as suitable carbonyl compound, the same temperature control has been observed, once again.¹¹ In particular, working at low temperatures (−20 °C), a high diastereoselectivity was observed, obtaining only one of four possible anti diastereomers.

The importance of quaternary stereocentres and their implementation in drug discovery has prompted us to study a possible synthetic strategy to obtain fluorinated γ-amino alcohols with a quaternary motif.¹² Thus, α,α-dialkyl aldehydes were chosen as carbonyl partners in a one-pot solvent-free eco-friendly Mannich-type reaction (Scheme 1).

Scheme 1 Strategy to trifluoromethylated γ-amino alcohols.

Results and discussion

Starting from our reported data,¹³ N-benzyl protected aldimine (E)-1a¹⁴ and 2-methylbutanal (2, 1.5 equiv.) were chosen as suitable reactants to find the best reaction conditions (Table 1).

Table 1 Optimisation of the Mannich-type reaction

Entry	Catalyst	Time (h)	T (°C)	Solvent	Conversion^a (%)
a Determined by ^1H and ^19F NMR analyses performed on the crude mixture.b N-Methyl-2-pyrrolidone.
1		17	25	—	57
2		75	−20 → 0	NMP^b	43
3		18	25	—	56
4		73	−20 → 0	NMP^b	41
5	TEA	4	−20 → 25	CH2Cl2	Complex mixture
6	DABCO	4	−20 → 25	CH2Cl2	Complex mixture
7	Cs2CO3	3	−20 → 25	CH2Cl2	Complex mixture
8	KF	4	−20 → 25	CH2Cl2	Complex mixture
9	—	18	25	CH2Cl2	37
10	—	18	25	PhCH3	30
11	—	18	25	NMP^b	Trace
12	—	18	25	THF	Trace
13	—	18	25	—	68
14	—	72	−20 → 0	—	50

---

First the reactions were performed in the presence of L-proline (entries 1 and 2) or of (S)-α,α-diphenylprolinol (entries 3 and 4) as catalyst (20 mol%), working in the presence or in the absence of solvent and at different temperatures. ¹H and ¹⁹F NMR spectra showed a satisfactory conversion of imine 1a in the Mannich adducts but, unexpectedly, organocatalysts seem to not affect the diastereoselectivity, the geometric diastereomers being obtained always with the same unsatisfactory ratio (6/4), even working at lower temperatures (from −20 to 0 °C). Then, we considered different organic (entries 5 and 6) or inorganic bases (entries 7 and 8), hoping to get at least an increase in yields. However, only complex crude mixtures were observed in all cases, even after shorter reaction times. Finally, considering that the presence of a methyl substituent in α-position to aldehyde group should favour the carbonyl enolizability,¹⁵ Mannich-type addition was attempted in the absence of catalyst, working with different solvents (entries 9 and 12) or under solvent-free conditions (entry 13). Once again, a strong solvent influence was observed, the best yields being obtained performing the reaction under solvent-free conditions (entry 13). With the hope of getting a relevant increase in the stereoselective induction, we repeated the solvent-free Mannich reaction lowering the temperature (−20 → 0 °C, entry 14), but the diastereoselectivity did not change, as shown by ¹⁹F NMR analyses, and only a decrease of both conversion and reaction rate were observed.

Then, fixed the best reaction conditions, starting from trifluoroacetaldehyde ethyl hemiacetal and benzylamine, a one-pot Mannich-type addition was performed both without added catalysts and in the presence of organocatalysts (Scheme 2). After reduction under classical conditions of Mannich adducts,¹⁶ the expected β,β-dialkyl γ-amino γ-trifluoromethyl alcohols 3,3′a were obtained (Scheme 2).

Scheme 2 Comparison between organocatalysed and uncatalysed additions.

After separation by flash chromatography on silica gel of two diastereomers 3,3′a formed under all conditions reported in Scheme 2, we considered the possible enantioselective induction which may have been obtained in the organocatalysed reactions. Regrettably, chiral HPLC analysis showed that amino alcohols were formed as a racemate in all cases (see ESI†).

Considering the stereoselectivity, the reaction time of both catalysed and uncatalysed reactions and their yields, it can be concluded that, probably, the Mannich addition always takes place through a very fast and uncontrollable self-catalysed pathway favoured by the enolate stability (Fig. 1), without involving added organocatalysts.

Fig. 1 Mannich addition through a self-catalysed pathway.

Then, we considered the possibility to reduce the reaction times performing the addition in the presence of zirconium tetrachloride¹⁷ as a suitable and efficient Lewis acid, compatible with the –CF₃ group¹⁸ and successfully used by us as catalyst in some addition reactions.¹⁹ In effect, performing the Mannich step in the presence of ZrCl₄ (20 mol%) NMR spectra showed the disappearance of reactants after only 15 min, but, surprisingly, unexpected trifluoromethyl 1,3-diamino compounds 4,4′a were isolated as major products after reduction step, while compounds 3,3′a were present only in trace (Table 2, entry 1).

Table 2 ZrCl₄-catalysed Mannich-type reaction

Entry	ZrCl4 (mol%)	Solvent	Time (h)	Yield of 3,3′a^a (%)	Yield of 4,4′a^a (%)
a Determined after flash chromatography on silica gel.
1	20	—	0.25	<5	37
2	20	THF	0.25	<5	34
3	20	PhCH3	0.25	<5	32
4	20	CH2Cl2	0.25	<5	35
5	5	—	15	36	10

---

The same product distribution was achieved even working in the presence of different solvents (Table 2, entries 2–4) and only when ZrCl₄ was used in the lowest molar percent (entry 5), we are able to obtain 3,3′a as the major product, but in lower yields and longer times (15 h).

A hypothesis can be made to explain these last results. ZrCl₄ is promoting, as expected, the Mannich addition, but it also greatly accelerates the imine hydrolysis.^17b Therefore, after coordination of catalyst to form I and deprotonation of the enolisable aldehyde (II), an intramolecular attack leads to the intermediate III (Fig. 2). At the same time, however, zirconium(IV) appears to promote the imine hydrolysis so generating in situ benzylamine. The latter, capable of giving a nucleophilic addition to activated carbonyl of III, leads to imine IV. The subsequent reduction with NaBH₄ furnished trifluoromethyl 1,3-diamino compounds 4,4′a as major products.

Fig. 2 Possible pathway for Zr-catalysed Mannich-type reaction.

Completed the reaction condition screening, the one-pot procedure was extended also to other aldehydes to test the influence of different α-alkyl substituents on the addition reaction (Table 3).

Table 3 One-pot Mannich-type addition with different α-hindered aldehydes

Entry	Aldehyde	R	R′	Product	Yield^a (%)	anti/syn^b
a After purification by flash chromatography on silica gel.b Determined by NMR analysis (see ESI).
1	5	Me	Ph	—	—	—
2	6	–(CH2)5–		9a	70	—
3	7	Et	Bu	10a	72	99 : 1
4	8	Me	Me	11a	62	—

---

As reported in Table 3, the reaction failed when a phenyl group was present (entry 1). In effect, the extended conjugation of enolate derived from aldehyde 5 can drastically lower its nucleophilicity enough to make it unreactive. On the contrary, in all other cases, stable enolate species act as good nucleophiles and give the expected fluorinated γ-amino alcohols 9a–11a, through an easy and eco-friendly one-pot process. Furthermore, an increase of stability difference between the E and Z enolate isomers leads to the formation of only the more stable E isomer and allows us to obtain a complete diastereoselective control (entry 3), the only β,β-dialkyl γ-amino γ-trifluoromethyl alcohols anti-10a (see ESI†) being formed as a pair of enantiomers (Fig. 3).

Fig. 3 Stereoselective pathway of addition.

Continuing our studies on optically pure fluorinated γ-amino alcohols, we decided to attempt the Mannich-type reaction starting from aldimine (R,E)-1b, derived from (R)-1-phenylethylamine and trifluoroacetaldehyde ethyl hemiacetal, hoping that the presence of a chiral resident centre on the electrophilic species can bring to facial stereoselectivity control of the reaction (Table 4).

Table 4 Synthesis of β,β-dialkyl γ-amino γ-trifluoromethyl alcohols

Entry	Aldehyde	R	R′	Product	Yield^a (%)	anti/syn^b	dr^b
a After purification by flash chromatography on silica gel.b Determined by NMR analysis (see ESI).
1	2	Me	Et	3,3′b	70	1 : 1	99 : 1
2	6	–(CH2)5–		9b	70	—	99 : 1
3	7	Et	Bu	10b	71	99 : 1	99 : 1
4	8	Me	Me	11b/12b	65	—	70 : 30

---

Contrary to what happened starting from aldehyde 7 (entry 3), the syn/anti control failed using aldehyde 2 as nucleophilic source (Table 4, entry 1). Nevertheless, very good yields and a stereofacial control of attack, from satisfactory (entry 4) to very high (entries 1–3), were obtained in all cases. After chromatography on silica gel, β,β-dialkyl γ-amino γ-trifluoromethyl alcohols anti-3b, anti-10b, syn-3′b, bearing a quaternary chiral centre, and 9b, 11b and 12b were obtained as diastereomerically pure compounds.

To determine the stereochemistry of addition pathways, we turned our attention to assign the absolute configuration of the new chiral centres. So, following our already reported methodology,²⁰ two-dimensional nuclear Overhauser effect spectroscopic analyses (2D NOESY) coupled with computational studies (see ESI†) allowed us to assign the (R,S) absolute configuration to 9b and 11b and the (R,R) absolute configuration to 12b. Starting from these data, further 2D NOESY analyses and computational studies permitted to assign the (R,S,R) absolute configuration to 3b and 10b and the (R,S,S) absolute configuration to 3′b (Fig. 4).

Fig. 4 β,β-Dialkyl γ-amino γ-trifluoromethyl alcohols as optically pure diastereomers.

Based on stereochemical data, it is possible to conclude that (E)-enolates added only to prochiral Si face of aldimine 1b, a like attack giving the optically pure anti isomer, while an unlike attack leads to the optically pure syn diastereomer (Fig. 5).

Fig. 5 Stereochemical outcome of self-catalysed Mannich-type addition.

Conclusions

In summary, in this paper a one-pot self-catalysed solvent-free Mannich-type addition of α-alkyl substituted aldehydes on trifluoromethyl (E)-aldimines was reported. After reduction reactions, β,β-dialkyl γ-amino γ-trifluoromethyl alcohols were obtained in good yields. Linear or cyclic alkyl α-substituted aldehydes successfully reacted with trifluoromethyl aldimines, which first act as a base, generating the nucleophilic species, and then as an activated electrophile.²¹ The presence of an electron donating group (EDG) in the α-position to the aldehyde function drastically modified the reaction. In fact, while working with linear aldehydes¹⁰ the reaction took places only in the presence of an organocatalyst, the presence of just a methyl in the α-position increased the aldehyde reactivity that it seems impossible to control the self-catalysed reaction.

In reactions described here, no temperature effect was observed, unlike the case of linear aldehydes and the geometric syn/anti stereoselectivity was controlled by the different stability of E and Z isomers of enolates, while the facial stereoselectivity control depends on the presence of a resident chiral centre on the electrophilic aldimine. Interestingly, adding ZrCl₄ to the reaction mixture, N,N′-dibenzyl protected trifluoromethyl 1,3-diamines were obtained, probably due to the aldimine hydrolysis promoted by the same Lewis acid.

Experimental section

General remarks

IR spectra were recorded on a Perkin-Elmer 1600 FT/IR spectrophotometer in CHCl₃ as solvent. ¹H NMR and ¹³C NMR spectra were recorded on a VARIAN XL-300 spectrometer at 300 and 75 MHz or on a Bruker Avance III at 400 and 101 MHz, respectively, at room temperature. CDCl₃ was used as solvent and CHCl₃ and CDCl₃ as internal standard for ¹H and ¹³C, respectively. ¹⁹F NMR spectra were recorded on a VARIAN XL300 spectrometer at 282.2 MHz, using CDCl₃ as solvent and C₆F₆ as internal standard. The NOESY experiments were performed with a Bruker Avance III spectrometer at 400 MHz using CDCl₃ as solvent and CHCl₃ as internal standard and used to assist in structure elucidation.²² Enantiomeric ratios were determined by HPLC analyses performed with a Varian 9002 instrument equipped with a Varian 9050 UV/Vis detector using an analytical IA Chiralcel column, HPLC grade hexane/2-propanol = 95 : 5 as eluent, low 1.2 mL min⁻¹. HR-MS analyses were performed using a Micromass Q-TOF Micro quadrupole-time of flight (TOF) mass spectrometer equipped with an ESI source and a syringe pump. The experiments were conducted in the positive ion mode. Optical rotation was determined at 25 °C with a JASCO DIP-370 polarimeter at a wavelength of 589 nm, using a quartz cell of 1 cm length. All trifluoromethyl (E)-aldimines are known compounds.¹³

Synthesis of trifluoromethyl γ-amino alcohols. General procedure

An equimolar solution (1 mmol) of trifluoroacetaldehyde ethyl hemiacetal and primary amine was heated to 120 °C for 3 h (¹⁹F NMR). After bringing the reaction mixture to room temperature, aldehyde (1.5 mmol) was fast added. Then, the reactions were stirred under solvent-free conditions at different temperatures (25, 0 or −20 °C). Then (16–72 h, see Tables 1, 3 and 4), anhydrous MeOH (3 mL) and NaBH₄ (2 mmol) were added at 0 °C. After 1 h of stirring, the mixtures were quenched with a saturated aqueous NH₄Cl solution, extracted with Et₂O and dried on Na₂SO₄. After removal of solvents in vacuo, expected trifluoromethyl γ-amino alcohols were obtained as pure compounds by flash chromatography on silica gel (hexane/EtOAc = 9 : 1).

Organocatalysed Mannich-type reactions. General procedure

A mixture of trifluoromethyl aldimine (E)-1a (1 mmol), 2-methylbutanal (2, 1.5 mmol) and L-proline or (S)-α,α-diphenylprolinol (5–20 mol%, see Table 1) was stirred at different temperatures (−20, 0 or 25 °C) in NMP or under solvent-free conditions, respectively. After reduction reaction with NaBH₄ (2 mmol) in MeOH (3 mL) at 0 °C, the crude mixtures were quenched with a saturated aqueous NH₄Cl solution and extracted with Et₂O. Collected organic layers were dried on anhydrous Na₂SO₄, the solvent was evaporated in vacuo, and the residue purified by flash chromatography on silica gel (eluent hexane/ethyl acetate = 9 : 1).

ZrCl₄-Catalysed Mannich-type reactions. General procedure

ZrCl₄ (5–20 mol%, see Table 2) was added to a mixture of trifluoromethyl aldimines (E)-1a (1 mmol) and 2-methylbutanal (2, 1.5 mmol). After stirring at room temperature (0.25–15 h) and reduction reaction with NaBH₄ (2 mmol) in MeOH (3 mL), the crude mixtures were quenched with a saturated aqueous NH₄Cl solution and extracted with Et₂O. Collected organic layers were dried over anhydrous Na₂SO₄, the solvent was evaporated in vacuo, and the residue purified by flash chromatography on silica gel (eluent hexane/ethyl acetate = 9 : 1).

anti-3-(Benzylamino)-2-ethyl-4,4,4-trifluoro-2-methylbutan-1-ol (3a). Yellow oil (39%, 107 mg). IR: 3347 cm⁻¹. ¹H NMR (CDCl₃): δ 0.83 (t, J = 7.5 Hz, 3H), 0.85 (s, 3H), 0.92–1.25 (br, 2H), 1.31–1.51 (m, 1H), 1.58–1.79 (m, 1H), 2.77 (dd, J = 12.9, 69.3 Hz, 2H), 3.69 (dd, J = 12.9, 30.7 Hz, 2H), 3.80 (q, J = 8.5 Hz, 1H), 7.15–7.31 (m, 5H). ¹³C NMR (CDCl₃): δ 7.4, 19.4, 29.5, 37.8, 54.3, 56.4, 76.4 (q, J = 27.8 Hz), 126.3 (q, J = 285.1 Hz), 127.6, 128.3 (2C), 128.7 (2C), 138.3. ¹⁹F NMR (CDCl₃): δ −71.9 (d, J = 7.7 Hz). HR-MS (ESI Q-TOF) (m/z) [M + H]⁺ calcd for C₁₄H₂₁F₃NO⁺ 276.1575, found 276.1570.

syn-3-(Benzylamino)-2-ethyl-4,4,4-trifluoro-2-methylbutan-1-ol (3′a). Yellow oil (26%, 72 mg). IR: 3347 cm⁻¹. ¹H NMR (CDCl₃): δ 0.86 (t, J = 7.5 Hz, 3H), 1.11 (s, 3H), 1.18–1.39 (br, 1H), 1.47 (q, J = 7.5 Hz, 2H), 2.01–2.07 (br, 1H), 2.78 (dd, J = 12.7, 93.8 Hz, 2H), 3.72–3.82 (m, 2H), 3.88 (q, J = 8.0 Hz, 1H), 7.26–7.37 (m, 5H). ¹³C NMR (CDCl₃): δ 7.7, 21.0, 26.3, 38.3, 54.3, 56.9, 77.8 (q, J = 28.1 Hz), 125.8 (q, J = 284.9 Hz), 127.6, 128.3 (2C), 128.7 (2C), 138.2. ¹⁹F NMR (CDCl₃): δ −72.1 (d, J = 8.1 Hz). HR-MS (ESI Q-TOF) (m/z) [M + H]⁺ calcd for C₁₄H₂₁F₃NO⁺ 276.1575, found 276.1570.

N¹,N³-Dibenzyl-2-ethyl-4,4,4-trifluoro-2-methylbutane-1,3-diamine (4,4′a). Yellow oil (37%, 135 mg). 3330 cm⁻¹. ¹H NMR (CDCl₃): δ 0.82 (t, J = 7.5 Hz, 3H), 0.85–0.92 (m, 6H), 0.96 (s, 3H), 1.31–1.80 (m, 8H), 3.13 (q, J = 8.37 Hz, 1H), 3.21 (q, J = 8.35 Hz, 1H), 3.36–3.54 (m, 4H), 3.98 (dd, J = 12.2, 115.9 Hz, 8H), 7.26–7.40 (m, 20H). ¹³C NMR (CDCl₃): δ 7.7, 7.8, 17.4, 19.9, 23.4 (2C), 29.0 (2C), 40.6 (2C), 53.9, 54.0, 63.6 (q, J = 24.4 Hz), 66.8 (q, J = 24.3 Hz), 69.5 (2C), 127.3 (q, J = 288.6 Hz), 127.4 (q, J = 288.3 Hz), 127.7 (4C), 128.7 (16C), 138.6 (2C), 138.9 (2C). ¹⁹F NMR (CDCl₃): δ −65.3 (d, J = 10.0 Hz), −66.1 (d, J = 12.9 Hz). HR-MS (ESI Q-TOF) (m/z) [M + H]⁺ calcd for C₂₁H₂₈F₃N₂⁺ 365.2205, found 365.2213.

{1-[1-(Benzylamino)-2,2,2-trifluoroethyl]cyclohexyl}methanol (9a). Yellow oil (70%, 211 mg). IR: 3343 cm⁻¹. ¹H NMR (CDCl₃): δ 1.25–1.75 (m, 11H), 1.80–1.95 (br, 1H), 2.93 (dd, J = 13.0, 36.4 Hz, 2H), 3.75 (dd, J = 12.8, 32.3 Hz, 2H), 4.00 (q, J = 7.9 Hz, 1H), 7.25–7.39 (m, 5H). ¹³C NMR (CDCl₃): δ 20.9, 21.3, 25.9, 31.8, 32.5, 37.8, 54.5, 54.8, 77.3 (q, J = 27.5 Hz), 126.6 (q, J = 285.8 Hz), 127.5, 128.2 (2C), 128.6 (2C), 138.6. ¹⁹F NMR (CDCl₃): δ −71.5 (d, J = 8.0 Hz). HR-MS (ESI Q-TOF) (m/z) [M + H]⁺ calcd for C₁₆H₂₃F₃NO⁺ 302.1732, found 302.1741.

2-[1-(Benzylamino)-2,2,2-trifluoroethyl]-2-ethylhexan-1-ol (10a). Yellow oil (72%, 228 mg). IR: 3351 cm⁻¹. ¹H NMR (CDCl₃): δ 0.80 (t, J = 7.3 Hz, 3H), 0.91 (t, J = 7.1 Hz, 3H), 1.14–1.43 (m, 8H), 1.59–1.64 (m, 2H), 2.84 (dd, J = 6.3, 12.6 Hz, 2H), 3.82 (dd, J = 9.6, 12.4 Hz, 2H), 3.96 (q, J = 8.0 Hz, 1H), 7.30–7.38 (m, 5H). ¹³C NMR (CDCl₃): δ 7.5, 14.1, 23.3, 24.1, 24.5, 31.0, 40.6, 53.7, 54.1, 75.7 (q, J = 28.3 Hz), 126.2 (q, J = 285.1 Hz), 127.9, 128.6 (2C), 128.8 (2C), 137.3. ¹⁹F NMR (CDCl₃): δ −71.5 (d, J = 10.8 Hz). HR-MS (ESI Q-TOF) (m/z) [M + H]⁺ calcd for C₁₇H₂₇F₃NO⁺ 318.2045, found 318.2053.

3-(Benzylamino)-4,4,4-trifluoro-2,2-dimethylbutan-1-ol (11a). Yellow oil (62%, 162 mg). IR: 3343 cm⁻¹. ¹H NMR (CDCl₃): δ 1.05 (s, 3H), 1.14 (s, 3H), 1.26–1.30 (br, 1H), 1.65–2.33 (br, 1H), 2.78 (dd, J = 12.6, 69.7 Hz, 2H), 3.67–3.88 (m, 3H), 7.29–7.37 (m, 5H). ¹³C NMR (CDCl₃): δ 21.6, 25.2, 35.5, 54.3, 59.6, 78.7 (q, J = 28.2 Hz), 125.6 (q, J = 294.8 Hz), 127.6, 128.3 (2C), 128.7 (2C), 138.3. ¹⁹F NMR (CDCl₃): δ −72.7 (d, J = 9.5 Hz). HR-MS (ESI Q-TOF) (m/z) [M + H]⁺ calcd for C₁₃H₁₉F₃NO⁺ 262.1419, found 262.1425.

anti-(2R,3S)-2-Ethyl-4,4,4-trifluoro-2-methyl-3-{[(R)-1-phenylethyl]amino}butan-1-ol (3b). Yellow oil (35%, 101 mg). [α]_D: −8.7 (c = 1.9 g/100 mL, CHCl₃). IR: 3347 cm⁻¹. ¹H NMR (CDCl₃): δ, 0.83 (t, J = 7.5 Hz, 3H), 0.87 (s, 3H), 0.94–1.30 (br, 1H), 1.42 (d, J = 6.7 Hz, 3H), 1.46–1.57 (m, 1H), 1.68–1.80 (m, 1H), 1.88–2.30 (br, 1H), 2.62 (dd, J = 12.74, 112.15 Hz, 2H), 3.72 (q, J = 6.74 Hz, 1H), 3.83 (q, J = 8.1 Hz, 1H), 7.22–7.38 (m, 5H). ¹³C NMR (75 MHz, CDCl₃) δ 7.4, 19.4, 23.3, 29.2, 37.8, 54.5, 58.9, 76.4 (q, J = 28.0 Hz), 126.3 (q, J = 285.1 Hz), 126.3 (2C), 127.6, 128.8 (2C), 143.2. ¹⁹F NMR (CDCl₃): δ −71.7 (d, J = 7.7 Hz). HR-MS (ESI Q-TOF) (m/z) [M + H]⁺ calcd for C₁₅H₂₃F₃NO⁺ 290.1732, found 290.1763.

syn-(2S,3S)-2-Ethyl-4,4,4-trifluoro-2-methyl-3-{[(R)-1-phenylethyl]amino}butan-1-ol (3′b). Yellow oil (35%, 101 mg). [α]_D: −5.9 (c = 1.0 g/100 mL, CHCl₃). IR: 3347 cm⁻¹. ¹H NMR (CDCl₃): δ 0.85 (t, J = 7.6 Hz, 3H), 1.06 (s, 3H), 1.18–1.32 (br, 1H), 1.41 (d, J = 6.7 Hz, 3H), 1.49–1.80 (m, 2H), 1.81–2.25 (br, 1H), 2.58 (dd, J = 12.7, 169.8 Hz, 2H), 3.72 (q, J = 6.8 Hz, 1H), 3.82 (q, J = 8.2 Hz, 1H), 7.29–7.39 (m, 5H). ¹³C NMR (CDCl₃): δ 7.7, 20.9, 23.5, 26.4, 38.2, 55.4, 58.9, 77.6 (q, J = 28.3 Hz), 125.8 (q, J = 285.0 Hz), 126.4 (2C), 127.5, 128.8 (2C), 143.4. ¹⁹F NMR (CDCl₃): δ −72.0 (d, J = 8.6 Hz). HR-MS (ESI Q-TOF) (m/z) [M + H]⁺ calcd for C₁₅H₂₃F₃NO⁺ 290.1732, found 290.1763.

{1-[(S)-2,2,2-Trifluoro-1-{[(R)-1 phenylethyl]amino}ethyl]cyclohexyl}methanol (9b). Yellow oil (70%, 221 mg). [α]_D: +66.8 (c = 3.5 g/100 mL, CHCl₃). IR: 3343 cm⁻¹. ¹H NMR (CDCl₃): δ 1.08–1.63 (m, 14H), 1.70–1.98 (br, 1H), 2.72 (dd, J = 13.0, 78.1 Hz, 2H), 3.71 (q, J = 6.7 Hz, 1H), 4.00 (q, J = 8.5 Hz, 1H), 7.21–7.38 (m, 5H). ¹³C NMR (CDCl₃): δ 20.9, 21.2, 23.3, 25.9, 31.8, 32.4, 37.7, 53.2, 58.8, 76.3 (q, J = 27.8 Hz), 126.1 (2C), 126.6 (q, J = 285.9 Hz), 127.4, 128.8 (2C), 143.7. ¹⁹F NMR (CDCl₃): δ −71.4 (d, J = 8.7 Hz). HR-MS (ESI Q-TOF) (m/z) [M + H]⁺ calcd for C₁₇H₂₅F₃NO⁺ 316.1888, found 316.1876.

anti-(R)-2-Ethyl-2-[(S)-2,2,2-trifluoro-1-{[(R)-1-phenylethyl]amino}ethyl]hexan-1-ol (10b). Yellow oil (69%, 229 mg). [α]_D: +106.1 (c = 3.5 g/100 mL, CHCl₃). IR: 3347 cm⁻¹. ¹H NMR (CDCl₃): δ 0.77 (t, J = 7.5 Hz, 3H), 0.91 (t, J = 7.3 Hz, 3H), 1.01–1.37 (m, 7H), 1.45 (d, J = 6.7 Hz, 3H), 1.50–1.71 (m, 3H), 2.63 (dd, J = 13.0, 82.5 Hz, 2H), 3.75 (q, J = 6.7 Hz, 1H), 3.91 (q, J = 8.1 Hz, 1H), 7.25–7.40 (m, 5H). ¹³C NMR (CDCl₃): δ 7.5, 14.1, 23.1, 23.2, 24.2, 24.4, 30.7, 40.4, 52.3, 59.1, 75.8 (q, J = 28.0 Hz), 126.4 (2C), 126.4 (q, J = 285.5 Hz), 127.7, 128.9 (2C), 142.8. ¹⁹F NMR (CDCl₃): δ −71.4 (d, J = 9.6 Hz). HR-MS (ESI Q-TOF) (m/z) [M + H]⁺ calcd for C₁₈H₂₉F₃NO⁺ 332.2201, found 332.2231.

(S)-4,4,4-Trifluoro-2,2-dimethyl-3-{[(R)-1-phenylethyl]amino}butan-1-ol (11b). Yellow oil (45.5%, 125 mg). [α]_D: −96.6 (c = 6.2 g/100 mL, CHCl₃). IR: 3345 cm⁻¹. ¹H NMR (CDCl₃): δ 1.01 (s, 3H), 1.09 (s, 3H), 1.10–1.38 (br, 2H), 1.40 (d, J = 6.6 Hz, 3H), 2.58 (dd, J = 12.5, 103.3 Hz, 2H), 3.67–3.77 (m, 2H), 7.25–7.39 (m, 5H). ¹³C NMR (CDCl₃): δ 21.6, 23.5, 25.0, 35.4, 58.0, 58.8, 78.6 (q, J = 28.2 Hz), 125.7 (q, J = 284.6 Hz), 126.4 (2C), 127.5, 128.8 (2C), 143.5. ¹⁹F NMR (CDCl₃): δ −72.5 (d, J = 7.9 Hz). HR-MS (ESI Q-TOF) (m/z) [M + H]⁺ calcd for C₁₄H₂₁F₃NO⁺ 276.1575, found 276.1563.

(R)-4,4,4-Trifluoro-2,2-dimethyl-3-{[(R)-1-phenylethyl]amino}butan-1-ol (12b). Yellow oil (19.5%, 54 mg). [α]_D: +105.9 (c = 5.0 g/100 mL, CHCl₃). IR: 3345 cm⁻¹. ¹H NMR (CDCl₃): δ 1.01 (s, 3H), 1.14 (s, 3H), 1.20–1.33 (br, 2H), 1.53 (d, J = 6.78 Hz, 3H), 2.62 (dd, J = 12.55, 74.18 Hz, 2H), 3.82–4.01 (m, 2H), 7.32–7.41 (m, 5H). ¹³C NMR (CDCl₃): δ 20.9, 24.9, 29.7, 35.6, 57.6, 59.0, 70.1 (q, J = 30.6 Hz), 125.7 (q, J = 284.0 Hz), 126.8 (2C), 128.1, 129.0 (2C), 139.2. ¹⁹F NMR (CDCl₃): δ −72.7 (d, J = 8.1 Hz). HR-MS (ESI Q-TOF) (m/z) [M + H]⁺ calcd for C₁₄H₂₁F₃NO⁺ 276.1575, found 276.1578.

Acknowledgements

We thank the Università degli Studi di Roma “La Sapienza” for financial support.

Notes and references

1. For reviews, see: (a) M. Benohoud and Y. Hayashi, Enamine Catalysis of Mannich Reactions, in Science of Synthesis, Asymmetric Organocatalysis 1; Lewis Base and Acid Catalysts, ed. B. List, Georg Thieme Verlag K. G., Stuttgart, 2012, p. 73; (b) A. Ting and S. E. Schaus, Eur. J. Org. Chem., 2007, 5797.
2. (a) S. Kobayashi, Y. Mori, J. S. Fossey and M. M. Salter, Chem. Rev., 2011, 111, 2626; (b) R. G. Arrays and J. C. Carretero, Chem. Soc. Rev., 2009, 38, 1940; (c) S. Fustero, J. F. Sanz-Cervera, J. L. Aceña and M. Sánchez-Roselló, Synlett, 2009, 525; (d) J. M. M. Verkade, L. J. C. van Hemert, P. J. L. M. Quaedflieg and F. P. J. T. Rutjes, Chem. Soc. Rev., 2008, 37, 29; (e) S. Mukherjee, J. W. Yang, S. Hoffmann and B. List, Chem. Rev., 2007, 107, 5471; (f) G. K. Friestad and A. K. Mathies, Tetrahedron, 2007, 63, 2541.
3. (a) C. A. de Parrodi and E. Juaristi, Synlett, 2006, 2699; (b) D. Enders, M. Moser, G. Geibel and M. C. Laufer, Synthesis, 2004, 2040; (c) T. Kochi, T. P. Tang and J. A. Ellman, J. Am. Chem. Soc., 2003, 125, 11276; (d) E. F. Kleinmann, in Comprehensive Organic Synthesis, ed. B. M. Trost and I. Fleming, Pergamon Press, New York, 1991, ch. 4.1, vol. 2.
4. (a) X.-J. Zhang, T.-B. Lai and R. Y.-C. Kong, Biology of fluoro-organic compounds, in Fluorous Chemistry, ed. I. T. Horváth, Topics in Current Chemistry, 2012, vol. 308, p. 365; (b) S. Purser, P. R. Moore, S. Swallow and V. Gouverneur, Chem. Soc. Rev., 2008, 37, 320.
5. (a) A. Abate, A. Petrozza, G. Cavallo, G. Lanzani, F. Matteucci, D. W. Bruce, N. Houbenov, P. Metrangolo and G. Resnati, J. Mater. Chem. A, 2013, 1, 6572; (b) M. Hird, Chem. Soc. Rev., 2007, 36, 2070; (c) H. Xue, R. Verma and J. M. Shreeve, J. Fluorine Chem., 2006, 127, 159.
6. (a) S. Fustero, D. Jiménez, J. F. Sanz-Cervera, M. Sánchez-Roselló, E. Esteban and A. Simón-Fuentes, Org. Lett., 2005, 7, 3433; (b) S. Fustero, E. Esteban, J. F. Sanz-Cervera, D. Jiménez and F. Mojarrad, Synthesis, 2006, 4087.
7. S. Fustero, F. Mojarrad, M. D. Pérez Carrión, J. F. Sanz-Cervera and J. L. Aceña, Eur. J. Org. Chem., 2009, 5208.
8. F. Huguenot and T. Brigaud, J. Org. Chem., 2006, 71, 2159.
9. S. Fioravanti, Tetrahedron, 2016, 72, 4449.
10. S. Fioravanti, L. Parise, A. Pelagalli, L. Pellacani and L. Trulli, RSC Adv., 2015, 5, 29312.
11. L. Parise, A. Pelagalli, L. Trulli, M. C. Vergari, S. Fioravanti and L. Pellacani, Chirality, 2015, 27, 571.
12. J. Nie, H.-C. Guo, D. Cahard and J.-A. Ma, Chem. Rev., 2011, 111, 455 and references therein.
13. (a) L. Parise, A. Pelagalli, L. Pellacani, F. Sciubba, M. C. Vergari and S. Fioravanti, J. Org. Chem., 2016, 81, 2864; (b) L. Parise, L. Pellacani, F. Sciubba, L. Trulli and S. Fioravanti, J. Org. Chem., 2015, 80, 8300.
14. L. Carroccia, S. Fioravanti, L. Pellacani and P. A. Tardella, Synthesis, 2010, 4096.
15. The uncatalysed addition of linear aldehydes failed on the same C-CF₃ aldimines, but easily occurs on the corresponding C-CH₃ aldimines: A. Pelagalli, L. Pellacani, E. Scandozza and S. Fioravanti, Molecules, 2016, 21, 723.
16. A. Córdova, Synlett, 2003, 1651.
17. (a) Z.-H. Zhang and T.-S. Li, Curr. Org. Chem., 2009, 13, 1; (b) G. Smitha, S. Chandrasekhar and C. S. Reddy, Synthesis, 2008, 829.
18. B. Morandi and E. M. Carreira, Angew. Chem., Int. Ed., 2011, 50, 9085.
19. (a) S. Fioravanti, L. Pellacani and M. C. Vergari, Org. Biomol. Chem., 2012, 10, 8207; (b) S. Fioravanti, L. Pellacani and M. C. Vergari, Org. Biomol. Chem., 2012, 10, 524.
20. S. Fioravanti, A. Pelagalli, L. Pellacani, F. Sciubba and M. C. Vergari, Amino Acids, 2014, 46, 1961.
21. An analogous behaviour of trifluoromethyl aldimines was already reported by us in the amination reaction with alkyl nosyloxycarbamates: L. Carroccia, S. Fioravanti, L. Pellacani, C. Sadun and P. A. Tardella, Tetrahedron, 2011, 67, 5375.
22. T. D. W. Claridge, High-Resolution NMR Techniques in Organic Chemistry, Elsevier Science, Amsterdam, The Netherlands, 3rd edn, 2016.

---

Footnote

† Electronic supplementary information (ESI) available: General procedures, analytical and spectroscopic data, ¹H, ¹³C, ¹⁹F, 2D NMR spectra and computational data of all new compounds. See DOI: 10.1039/c6ra22507a

---