A General, Enantioselective Synthesis of Protected Morpholines and Piperazines

Abstract

A short, high yielding protocol has been developed for the enantioselective and general synthesis of C2-functionalized, benzyl protected morpholines and orthogonally N,N′-protected piperazines from a common intermediate.

Morpholines and piperazines are saturated aza-heterocycles commonly employed as bases in organic synthesis.¹ These heterocycles have also become key components of pharmaceutical compositions, typically with chiral C-functionalization at C2.^2,3 However, chemistry to access enantiopure C2-functionalized morpholines and piperazines is limited, relying either on the chiral pool, stoichiometric auxilaries or HPLC resolution of racemic mixtures.^1–6 Based on our earlier efforts to access chiral β-fluoroamines and N-termial aziridines via organocatalysis,^7–9 we applied this strategy to the enantioselective synthesis of C2-functionalized morpholines and piperazines (Fig. 1). Here, an organocatalytic, enantioselective chlorination of aldehyde 1 produced 2,^10,11 which was used without purification. A subsequent reductive amination step with an amine containing an embedded ‘O’ or ‘N’ nucleophile (3 or 4), such that after based-induced cyclization of either 5 or 6, N-benzyl protected morpholines 7 and orthogonally N,N′-protected piperazines 8, respectively, were prepared with C2-functionalization.¹² While this result was gratifying, the methodology suffered from two key limitations: 1) the 3-step overall yield was low (13–50%) and 2) the % ee was variable (55–98% ee) due to the epimerization prone α-chloroaldehyde 2. In fact, one trial where the α-chloroaldehyde 2 was left on the bench for hours prior to subsequent reductive amination led to an % ee erosion of greater than 30%. Furthermore, four separate trials of this reaction sequence, with immediate use of the chloroaldehyde, afforded 78–94% ee immediately upon generation. In addition to varying ee %, the incipient imine could also be attacked by the latent oxygen nucleophile to form an undesired hemiaminal that further compromised yields, leading to recovered aldehyde starting material upon work-up. Thus, in this Letter, we report a general, high yielding solution for the enantioselective synthesis of these valuable aza-heterocycles that overcomes the limitations of the first generation approach affording good overall yields and high enantioselectivities.

Figure 1.

First generation organocatalytic approach for the enantioselective synthesis of C2-functionalized, N-protected morpholine and orthogonally N,N′-protected piperazines.

In the original Jørgensen methodology for the organocatalytic α-chlorination of aldehydes,^10,11 the aldehydes could be immediately reduced with NaBH₄ to the corresponding 2-chloro alcohols 9 without any loss in enantioselectivity; moreover, the alcohol derivatives were configurationally stable. Thus, if we could convert the hydroxyl moiety of 2-chloro alcohol into an efficient leaving group 10, followed by a chemoselective displacement by 3 or 4, substrates 5 or 6 would result which could be smoothly cyclized to form either 7 or 8 (Fig. 2). This envisioned approach was attractive as it would eliminate the variable % ee, avoid the undesired hemiaminal formation and thus, improve the overall yields of 7 and 8.

Figure 2.

Envisioned second generation organocatalytic approach for the enantioselective synthesis of C2-functionalized, N-protected morpholine and orthogonally N,N′-protected piperazine.

To test this new approach, we prepared various 2-chloro alcohol substrates 9a–d under standard conditions in high yield (72–83%) over two steps and in high enantioselectivity (80–98% ee) as expected from literature precedent (Scheme 1).^10–12 Now the challenge was to convert 9a–d into the appropriate bis-electrophile that would allow for a chemoselective displacement of the primary leaving group – something not yet reported in the literature. After surveying a number of potential primary leaving groups (mesylate, tosylate and iodide), the triflate emerged as the optimal moiety to deliver congeners of 5 and 6. Here, treatment of 9a–d with triflic anhydride in DCM with lutidine at −78 °C, smoothly generated the corresponding triflates, which were then immediately exposed to either secondary amine 3 or 4 to generate 5a–d and 6a–d (Scheme 2) in good yields (63–87%) for the two step sequence.

Scheme 1.

Organocatalytic, enantioselective synthesis of 2-chloro alcohols 9a–d.

Scheme 2.

Synthesis of Cyclization Substrates 5a–d and 6a–d.

With 5a–d in hand, we employed our optimized cyclization conditions (KO^tBu, CH₃CN, −20 °C) to afford N-benzyl protected morpholines 7a–d (Scheme 3) in good yield (65–74%) and with excellent enantioselectivities (80–98% ee). The only marginal % ee was in the silyl ether substrate 7d, which resulted from the intial α-chlorination step, and was expected based on literature precedent.^10,11 Overall yields for 7a–d from the commercial aldehydes ranged from 35–46% for the new five step sequence, a notable improvement over the 13–19% overall yields of the first generation, three-step approach as well as improved % ee.¹²

Scheme 3.

Enantiospecific Cyclization to Afford Morpholines

In a similar fashion (Scheme 4), but employing DMF as the solvent, orthogonally N,N′-protected piperazines 8a–d were arrived at in good yields (66–91%) and with high enantioselectivities (75–95% ee). As discussed earlier, the one low ee was due to the substrate. Once again, this new methodology afforded comparable or improved overall yields (35–60%) for the five step sequence and uniformly high % ee relative to the first generation approach (15–50% overall yields and 55–96% ee). Thus, this new five step sequence for the enantioselective synthesis of C2-functionalized, N-protected morpholines and piperazines affords access to these valuable aza-heterocycles that often cannot be accessed readily.

Scheme 4.

Enantiospecific Cyclization to Afford Orthogonally N,N′-Protected Piperazines 8a–d.

Finally, we applied this new methodolgy to a pharmaceutically relevant morpholine target with antipsychotic activity from the patent literature.¹³ Chiral morpholine 15, reported to be a specific dopamine subtype 4 (D₄) antagonist,⁶ was previously prepared in three steps, including a preparative chiral HPLC separation, in 9.9% overall yield (Scheme 5). Though they claim a single enantiomer of 15 to be more prefered, they did not disclose the absolute sterochemistry or the differences in D₄ potency.

Scheme 5.

Published Synthesis of Morpholine 15.

Therefore, we took advantage of already synthesized enantiopure (R)-morpholine 7b, removed the benzyl protecting group via hydrogenation and alkylated with 14 to deliver (R)-15 (Scheme 6). In contrast to the known route, our methodology provided enantiopure (R)-15 in 98% ee and in 35% overall yield, a significant improvement. In a similar manner, racemic 15 was prepared, according to Figure 2 utilizing D,L-proline as the organocatalyst, as well as (S)-15, and all three were evaluated against the full dopamine family of receptors, D₁-D₄, in both radioligand binding (K_i) and functional (IC₅₀) assays (Table 1).¹⁴ Racemic (±)-15, is devoid of activity at D₁ and D₂ (K_i and IC₅₀s >100 μM), and highly selective for D₄ versus D₃ (K_i and IC₅₀ >10 μM). Upon evaluation of the pure enantiomers (R)-15 and (S)-15, enantiospecific activity is clearly present. As shown in Table 1, (S)-15 is uniformly inactive against D₁-D₄ (IC₅₀s > 25 μM); however, (R)-15 is twice as potent (D₄ K_i = 0.07 μM, IC₅₀ = 0.18 μM) as racemic (±)-15, indicating that all of the activity of the racemate is due to the (R)-enantiomer. These data further highlight the utility of our methodology to afford high yielding, enantioenriched access to chiral, C2-functionalized morpholines and piperazines.

Scheme 6.

Enantioselective Synthesis of (R)-Morpholine 15.

Table 1.

Biological Actvity Data at D₁–D₄ for Racemic and Enantiopure Isomers of Morpholine 15.

compd	D1a		D2a		D3a		D4a
	Ki	IC50	Ki	IC50	Ki	IC50	Ki	IC50
(±)-15	>100	>100	>100	>100	10.8	31.8	0.14	0.36
(R)-15	>100	>100	>100	>100	15.7	46.2	0.07	0.18
(S)-15	>100	>100	>100	>100	25.9	76.4	>100	>100

^aK_i and IC₅₀ values are in μM, and represent at least three measurements.

In summary, we have developed an optimized five-step procedure for the enantioselective synthesis of N-benzyl protected morpholines and orthogonally N,N′-protected piperazines with chiral alkyl groups installed at the C2 position of each heterocyclic core via organocatalysis. This methodology allows for the rapid preparation of functionalized, pharmaceutically relevant morpholines and piperazines in 35–60% overall yields and in 75–98% ee. This new methodology addresses the major short-comings (variable % ee and low overall yields) of our first generation approach. Of major significance, this methodology does not rely on the chiral pool; instead we can employ simple aldehydes and commercial organocatalysts, thereby allowing access to either enantiomer of the corresponding morpholines and piperazines. Application of this new methodology to the synthesis and biological evaluation of a known D₄ antagonist further highlights the power of the methodology, and sheds light on enantioselective inhibition of dopamine receptors. Additional refinements are under development and will be reported in due course.