An investigation of the observed, but counter-intuitive, stereoselectivity noted during chiral amine synthesis via N-chiral-ketimines

The default explanation for good to high diastereomeric excess when reducing N-chiral imines possessing only mediocre cis/trans-imine ratios (>15% cis-imine) has invariably been in situ cis-to-trans isomerization before reduction; but until now no study unequivocally supported this conclusion. The present study co-examines an alternative hypothesis, namely that some classes of cis-imines may hold conformations that erode the inherent facial bias of the chiral auxiliary, providing more of the trans-imine reduction product than would otherwise be expected. The ensuing experimental and computational (DFT) results favor the former, pre-existing, explanation.

then hold 5 min), Program B: 50 °C (hold 1 min), 280 °C (14 °C/min, then hold 5 min) and Program C: 50 °C (hold 1 min), 100 °C (5 °C/min), then up to 280 °C (40 °C/min, then hold 2 min). All reagents were obtained from Sigma-Aldrich and used without further purification. 97% grade Ti(OiPr) 4 (catalog number, 205273) was used for all screening reactions (2-5 mmol). (S)-PEA (catalog number, 115568) was of 98% chemical purity and 98% ee. Platinum 5% on activated carbon (catalog number, 205931), Palladium 5% (dry basis) on activated carbon 50% water (catalog number, 276707), Palladium 5% on CaCO 3 with Lead (Lindlar's catalyst) (catalog number, 62145), Palladium on barium sulfate 5% Pd (catalog number, 76022) were all used as heterogeneous hydrogenation catalysts. The Raney-Nickel (in water) was purchased from Fluka (catalog number, 83440). All of the hydrogenation reactions were performed with reagents kept under nitrogen atmosphere and anhydrous solvents. When considering the -CH 2 R' carbonyl substituents (Figure 2 above), no -branching is present, e.g. n-butyl (imine 2d) or ibutyl (imine 2c), it is apparent that in the cis isomer steric interactions between the PEA phenyl group and the R' group will favor the cis--R' (II) conformation over the cis--R' (I) conformation. This stereochemical relay effect should have the consequence of eroding the expected facial selectivity for the cis-imine (if cis--R' (I) and cis--R' (II) are reduced at similar reaction rates. In the corresponding trans-imine series (Figure 2), the effect is expected to be minimal or non-existent due to a lack of proximity between the Ph and R' groups, i.e. conformations trans--R' (IV) and trans--R' (III) would be expected to be similar in energy. We therefore proposed that cis-N-phenylethylimines might experience facial erosion, or even inversion of facial selectivity, due to a strong steric effect imposed by the phenyl group of PEA on the proximal carbonyl substituent. In the corresponding trans series the effect is minimal. On changing to -branched carbonyl substituents, -CHR' 2 ( Figure 1S, above), e.g. chexyl (imine 2b) and phenyl (imine 2a), the conformational implications change dramatically. Again, allylic 1,2-strain is expected to force the -hydrogen of the -CHR' 2 substituent to be eclipsed with the methyl group ( Figure A). This conformation has the effect of locating one R' substituent on one face of the imine double bond and the other R' group on the opposite face of the imine double bond. When the R' groups are equivalent, both faces of the imine double bond would be similarly hindered. An immediate consequence would be little or no change in the 'expected' facial selectivity for both the cis-imine and trans-imine. Thus carbonyl substituents with -branching should largely lead to a linear relationships between cis/trans imine ratios and the product amine diastereomeric ratios, and our experimental data strongly supported this conclusion.       3 Pt/C c 0.10 full 36.5 a determined by GC, based on ketone, imine and amine product peaks b determined by GC c used as purchased from Sigma-Aldrich, dry d 50% wet and then dried under high vacuum overnight at 50°C Synthesis of the Rosenmund catalyst (Pd/BaSO 4 poisoned with quinoline-sulfur) 1.0 g of sulfur (Sigma-Aldrich, catalog number, 215198) and 5.0 g of quinoline (Sigma-Aldrich, catalog number, 453544) were refluxed for 5 h. The resulting dark brown liquid was diluted to 70 mL with xylene. To that solution was added an equimolar quantity of Pd/BaSO 4 as compared to the quinoline. This procedure was adapted from references [1S,2S].

General procedure for asymmetric reductive amination
In an anhydrous solvent (0.50-0.60 M) a prochiral ketone 1 (2.5 or 5.0 mmol), titanium tetraisopropoxide (1.25 equiv), and (S)-PEA (1.10 equiv) were combined and stirred at room temperature for 30 min. Raney-Ni (100 wt %; the catalyst was first triturated with EtOH (x 3) and then with the anhydrous reaction solvent (3 x 0.5 mL) before addition to the reaction) was then added and the vessel pressurized at 120 psi (8.3 bar) of hydrogen. After 9 h, the reaction was stopped and the mixture was stirred with aqueous NaOH (1.0 M, 10 mL) for 1 h. The heterogeneous mixture was then filtered through a bed of celite, and the celite washed with CH 2 Cl 2 or EtOAc. The filtrate was concentrated (rotary evaporator) to remove the low boiling organics and the remaining aqueous solution was then extracted with CH 2 Cl 2 (3 x 15 mL). The combined organic extracts were dried (Na 2 SO 4 ), filtered, concentrated (rotary evaporator), and GC analysis was performed to determine chemical purity and de. For isolated yield data see reference [S3].

Synthesis of (S)-phenylethylimine (2c) in methanol at room temperature
Ketone 1c (6.00 mmol) and (S)-PEA (1.10 equiv) are added to a round bottom flask containing anhydrous methanol (0.80 M). The reaction mixture was stirred for 24 h at room temperature. Concentration of the mixture (rotary evaporation) and high vacuum drying yields the respective ketimine (534 mg, 44% crude yield, > 98% GC purity).

Experimental procedure for (S)-phenylethylimine reduction
The (S)-PEA(2a-e) (1.25 mmol) was added to an organic solvent as indicated in the text of this article (0.60 M). Raney-Ni (100 wt %, triturated prior addition three times with EtOH and three further times with the reaction solvent) or Pt/C (0.50 mol %) or Pd/C (0.50 mol %) or Pd/CaCO 3 with Lead (Lindlar's catalyst, 0.50 mol %) was then added, and the mixture was hydrogenated at room temperature under 8.3 bar of hydrogen for 9 h (unless specified otherwise in the text). As a precautionary measure work up involved stirring with aqueous NaOH solution (1.0 M, 20.0 mL) for 1 h to ensure that any traces of the ketimine were hydrolyzed. The heterogeneous mixture was then filtered through a bed of celite, and the celite washed with CH 2 Cl 2 or EtOAc. The filtrate was concentrated (rotary evaporator) to remove the low boiling organics and the remaining aqueous solution was extracted with CH 2 Cl 2 (15 mL x 3). The combined organic extracts were dried (Na 2 SO 4 ), filtered, concentrated (rotary evaporator), and GC analysis was performed to determine chemical purity and de.

Deuteration of (S)-shenylethylimine
Under a nitrogen atmosphere the (S)-phenylethylimine of 2a or 2c (1.0 mmol) is added to a vessel by means of a syringe, followed by CD 3 OD (0.80 M). The mixture is stirred for 12 h and then concentrated (rotary evaporation and then under high vacuum) at room temperature. Multiple deuteration was confirmed by 1 H NMR analysis.

Effect of heterogeneous hydrogenation catalyst on (S)-phenylethylimines 2a-e
Catalyst activation: The hydrogenation catalyst (for Raney-Ni, trituration with EtOH (x 3) and the reaction solvent (x 3) with the reaction solvent) was deposited in a hydrogenation vessel with the reaction solvent. The vessel was then pressurized with hydrogen (120 psi, 8.3 bar) and stirred for 9 h. Concentrations and temperatures were in accordance with the imine reduction protocol.
(S)-Phenylethylimine 2a-e (1.0 mmol) was added under nitrogen atmosphere to a hydrogenation vessel containing the reaction solvent (0.60 M) and the hydrogenation catalyst (activated or non-activated) by means of a syringe, in the same amounts as called for by the imine reduction protocol. The mixture was left to stir for 10 h and then filtered and concentrated under high vacuum for 6 h at room temperature. cis/trans Ratios were determined by 1 H NMR analysis of this material.
General procedure for the preparation of (S)-phenylethylimines 2a-d p-Toluensulfonic acid (2.00 or 4.00 mol %) was added to a double-neck 100 mL round bottom flask equipped with a Dean-Stark trap (the water collection arm had spherical 4 Å molecular sieves added to it) and reflux condenser. To this was added toluene (0.50 M), ketone (32.00 mmol), and (S)-PEA (1.10 equiv). The mixture was refluxed for 24-60 h, then cooled to room temperature and concentrated (rotary evaporator). The residue was dissolved in hexane (or CH 2 Cl 2 ) and filtered to remove the precipitated p-TsOH, and the hexane briefly washed with aqueous NaHCO 3 (1.0 M, 40 mL). The hexane was then washed with brine solution, dried over MgSO 4 , filtered, and concentrated (rotary evaporator). The resulting oil was then dried under high vacuum at 50 °C (unless specified otherwise) for 24 h to remove any starting materials giving the crude product. GC analysis of the crude product showed less than 2% of the starting ketone and <1% (S)-PEA. cis/trans Ratios were determined on the crude product by 1 H NMR.
Note: none of the cis and trans-imines are separable by GC.
Below specific examples of (S)-phenylethylimines synthesized by this method are presented.
A representative example follows. To a 150 mL double neck round bottom flask containing EtOAc (0.75 M, 32 mL), the following were added under nitrogen in this order: Ti(OiPr) 4 (1.25 equiv, 30 mmol, 8.88 mL), butyrophenone (1.00 equiv, 24 mmol, 3.48 mL), (S)-PEA (1.10 equiv, 26.4 mmol, 3.37 mL). The mixture was stirred for 6 h at room temperature, and then quenched by stirring with aqueous NaOH (1.0 M, 30 mL) for 1 h. [Note: although this imine is reasonably resistant to these hydrolysis conditions, this cannot be said for the other imines studied here. If imines 2a-d are synthesized in this manner the work-up would require the less basic conditions of Na 2 CO 3 (0.5 M) and with shortened exposure times, e.g. 15 min. The basic conditions convert Ti(OiPr) 4 trapped intermediates and products to TiO 2 , without this filtration and sedimentation is frequently observed with reduced yields] The organic phase was collected and the aqueous phase was again extracted with EtOAc (3 x 30 mL). All of the organic extracts were combined and dried (Na 2 SO 4 ), filtered, concentrated (rotary evaporator), and the resulting oil was placed under high vacuum at 80 °C for 24 h. [Note 80 °C was required to remove the high boiling butyrophenone (bp 228-230 °C). The yields ranged from 55-70%, with 60% occurring most often, 2.26 g). These low yields imply imine hydrolysis during the stirring with aqueous NaOH. Again, the cis/trans ratio was determined using the crude product

CONTINUES ON THE NEXT PAGE…
Determination of trans-imines vs cis-imines. The next 20 pages are devoted to the NOESY spectra of imines 2a-e Due to 1,3-allylic strain, the methine hydrogen (H A ), of the α-MBA moiety, is known to lie in the plane of the ketimine C-N double bond [S9-11]. As a consequence, identification of the ketimine as either cis or trans should be possible due to the proximity of proton H A and a proton (H B ) on one of the substituents of the carbonyl carbon. The ketimine of acetophenone is shown below (See Figure 2S); a nOe would be expected between H A and H B in the trans-isomer only. Regarding all other substrates, the effect between the H A and the proton H B is clearly noticeable for the major isomer in all substrates, indicating thus their proximity in space. The data strongly suggests that the major ketimine diastereomer is trans-2a.   Figure S10: NOESY spectrum of ketimine 2c. Cross peak labeled at 1. The major quartet at 4.58 ppm (A) correlates with the doublet of the methyl at 1.47 ppm (C). Cross peak labeled at 2. The same correlation can be seen for the minor quartet (A') with the minor doublet for the resonance at 1.44 ppm (C'). Cross peak labeled at 3. The major quartet (A) correlates with the singlet at 1.81 ppm (B). Cross peak labeled at 4. The minor quartet (A') correlates with the methylene protons at 2.18 ppm (D) This combination of data strongly suggests that the major diastereomer is trans-2c and simultaneously confirms the minor ketimine as cis-2c. Cross Peak labeled 2. The same correlation can be seen for the minor quartet (A'), although the doublet for the minor resonates at δ 1.45 (C'). Cross Peak labeled 3. The major quartet (A) also correlates with the singlet at δ 1.82 (B) clearly suggesting that the major diastereomer is trans-2d. Cross Peak labeled 4. Correlation of the minor singlet at δ 2.03 (B') and the minor quartet (A') cannot be seen; this confirms the minor diastereomer to be the cis-2d.  Cross Peak labeled 2. The same correlation can be seen for the minor quartet (A'), although the doublet for the minor resonates at 1.37 ppm (C'). Cross Peak labeled 3. The major quartet (A) also correlates with a major multiplet at 2.71 ppm (B) clearly suggesting that the major diastereomer is trans-2e. Cross Peak labeled 4. Correlation of the minor multiplet at 2.55 ppm (B') and the minor quartet (A') cannot be seen; this confirms the minor diastereomer to be the cis-2e.   Fig.19S (y-axis 0.6-7.9 ppm; x-axis 0.78-4.51 ppm). Cross Peak labeled 1. The major quartet at 4.29 ppm (A) correlates with the doublet of the methyl at 1.32 ppm (C). Cross Peak labeled 2. Multiplet at 2.78 ppm (B) correlates with the multiplet at 1.10 ppm (E). Cross Peak labeled 3. The major quartet (A) also correlates with resonances in the aromatic area (D), clearly suggesting that the major diastereomer is cis-2f.