Expanding Complex Morpholines Using Systematic Chemical Diversity

Abstract

The morpholine heterocycle is a structural unit found in many bioactive compounds and FDA-approved drugs, but the generation of more complex C-functionalized morpholine derivatives remains considerably underexplored. Using systematic chemical diversity (SCD), a concept that guides the expansion of saturated drug-like scaffolds through regiochemical and stereochemical variation, we describe the synthesis of a collection of methyl-substituted morpholine acetic acid esters starting from enantiomerically pure amino acids and amino alcohols. In total, a complete matrix of 24 diverse substituted morpholines were produced that vary systematically in regiochemistry and stereochemistry (relative and absolute). These diverse C-substituted morpholines can be directly applied in fragment screening or incorporated as building blocks in medicinal chemistry and library synthesis.

Graphical Abstract:

Achieving chemical diversity within small molecule screening collections is a necessary goal for expanding screening outcomes across a spectrum of biological targets.^1–3 Such diversity can be obtained by populating a library with small molecules containing scaffolds that are highly structurally dissimilar to one another; however, such collections are practically difficult to achieve. Alternatively, building scaffold families having structurally diverse features around a common core scaffold is a synthetically more efficient means of expanding chemical diversity.^4–8 Systematic chemical diversity (SCD) is a synthesis guiding strategy aimed at achieving diversity within saturated scaffold families through regiochemical, absolute and relative stereochemical and appendage functionalization. SCD provides a more synthetically tractable route to compounds containing different vectors and substitution patterns for modulating a wide swath of unique biological targets. Additionally, because all the compounds are related through their common core scaffold, screening SCD-derived libraries deliver precise structure-activity relationships (SAR), which can be used to guide further optimization of the hits. We have previously documented the application of SCD for generating diverse compounds based on the well-known piperazine scaffold.^9–17 Here, we report the synthesis of a collection of morpholines inspired by SCD.

Morpholine-containing compounds are prevalent within both natural products and unnaturally occurring biologically active substances, including many FDA approved therapeutics.^18–25 The nitrogen and oxygen atoms housed within the morpholine scaffold endow it with properties unique from the all-carbon analog cyclohexane. For example, the nitrogen atom can be easily functionalized to generate diverse functional groups, and the oxygen atom augments solubility and potentially serves as a H-bond acceptor from a target protein. Despite these attributes, methods to produce enantiomerically pure C-substituted morpholines remain somewhat limited. We reasoned that applying SCD to generate enantiomerically pure C-substituted morpholines would lead to compounds that would serve as starting points for chemical probes and therapeutic leads.

We elected to extend our previous studies on piperazines^9–17 to produce a scaffold family of morpholine-acetic acid esters using SCD. Incorporating the acetic acid moiety into the scaffolds would serve the purposes of facilitating the morpholine ring construction, enhancing solubility for screening, and establishing a handle for further diversification. Because morpholine contains two different heteroatoms, it lacks an axis of symmetry relative to the piperazine heterocycle. Therefore, the acetic acid ester side chain could potentially reside at either the 2- or 3-position relative to the O atom creating the opportunity for further diversity based on regiochemistry. These considerations encouraged our goal to synthesize the methyl-substituted morpholine-2- and 3-acetic acid ester family depicted in Scheme 1. Practically, our plan for producing the complete matrix of morpholines was to generate each scaffold as a mixture of optically active diastereomers, which would subsequently be separated into their relative (cis and trans) stereoisomers.

Scheme 1.

Diverse methyl substituted morpholine-2- and 3-acetic acid esters.

We began by focusing on the morpholine 3-methyl-2-acetic acid ester. Our synthetic strategy hinged on constructing the desired morpholine ring through an annulation reaction of enantiomerically pure vicinal substituted amino alcohol and bromoethyldiphenylsulfonium triflate. Starting with N-Boc-protected (S)-alanine (1a), we conducted a Masamune condensation reaction to yield β-ketoester 2a.^{16, 26} Next, reduction of the ketone using sodium borohydride resulted in alcohol 3a as mixture of diastereomers. A key strategy in our labs for achieving full matrices of stereoisomers in a practical way has been to separate diastereomeric mixtures of final products to give enantiomerically pure species. Therefore without separation, 3a was subjected to N-Boc removal, followed by sulfonylation with 4-nitrobenzene sulfonyl chloride, leading to the protected amino alcohol 4a. Finally, subjection of 4a to annulation reaction conditions with bromoethyldiphenylsulfonium triflate and sodium hydride^{27, 28} afforded the anticipated 3-methyl morpholine-2-acetic acid esters as a mixture of diastereomers (5a/6a, cis/trans) in 2:1 ratio (based on crude NMR, Scheme 2). The diastereomeric morpholines were purified by silica gel column chromatography and the relative stereochemistry of the isolated compounds was determined using NMR techniques, including ¹H, ¹³C, and 2D NMR (COSY, HMBC, HSQC, and NOESY; Supplementary Information). In keeping with the goal of SCD, we next replicated the reaction sequence, commencing from Boc-protected (R)-alanine 1b, to yield diastereomeric products 5b/6b (opposite enantiomers of 5a/6a, Scheme 2). We evaluated the enantiomeric purity of the final products using chiral HPLC, which confirmed that no erosion of stereoisomeric purity occurred throughout the synthetic sequence.

Scheme 2:

Synthesis of 3-substituted morpholine-2-acetic acid esters; * = enantiomerically pure; Ns = 4-nitrobenzenesulfonyl.

We next proceeded to synthesize the 5-substituted morpholine-2- acetic acid ester series which we surmised could be prepared using an intramolecular oxa-Michael reaction. This possibility could be achieved by selectively establishing a Michael acceptor onto the nitrogen (rather than the oxygen) of an appropriately substituted amino alcohol. We pursued this strategy by starting with (S)-2-aminopropan-1-ol 7a, to synthesize sulfonamide 8a through a selective sulfonylation of the primary amine with 4-nitrobenzenesulfonyl chloride. Intermediate 8a was then selectively alkylated on the nitrogen using ethyl-4-bromocrotonate, resulting in the precursor (9a) for achieving the final morpholine products. An oxa-Michael conjugate addition²⁹ ensued upon treating 9a with potassium ter-butoxide at 0 °C, which produced the desired 2,5-substituted morpholines 10a/11a as ~1.8:1 (cis:trans) mixture of diastereomers (indicated by crude NMR) (Scheme 3). We additionally investigated the oxa-Michael reaction at −78 °C and 25 °C. The diastereomeric ratio was not significantly affected at lower temperature (~1.5:1, cis:trans), but at room temperature the reaction produced ~3:1 ratio of cis:trans isomers. These results highlight that at higher temperature a retro oxa-Michael reaction can ensue to favor the more thermodynamically stable cis isomer. The diastereomers were separated via column chromatography and we assigned the relative stereochemistry by utilizing ¹H, ¹³C, and 2D NMR (including COSY, HMBC, HSQC, and NOESY; Supplementary Information). To obtain the antipodes of 10a/11a, we initiated the synthesis using (R)-2-aminopropan-1-ol 7b. By repeating the identical synthetic sequence, we successfully generated 10b/11b (Scheme 3). Finally, chiral HPLC analysis was performed on the final products which verified that no loss of enantiomeric purity resulted from the synthesis pathway.

Scheme 3.

Synthesis of 5- and 6-substituted morpholine-2-acetic acid esters; * = enantiomerically pure; Ns = 4-nitrobenzenesulfonyl.

Next, we focused on the synthesis of 6-substituted morpholine-2- acetic acid esters. The oxa-Michael precursors 9c and 9d were synthesized in a two-step process starting from commercially available (S) or (R)-1-aminopropan-2-ol (7c and 7d) and employing the same synthetic reaction sequence used in the 2, 5-morpholine series (Scheme 3). In the key oxa-Michael reaction step, treatment of compounds 9c and 9d with potassium tert-butoxide at 0 °C delivered the targeted diastereomeric 2,6-substituted morpholine products 10c/11c and 10d/11d (~1:1 ratio by crude NMR). Next we conducted an oxa-Michael reaction at −78 °C and 25 °C and observed the trans product (11c) as major compound at lower temperature (~1:2.6, cis:trans), and cis diastereomer (10c) as the major product at room temperature (1:0.6, cis:trans). These different ratios are noteworthy given that, while SCD does not prioritize one diastereomer over another, the selective generation of diastereomers is warranted in target-oriented synthesis. The diastereomers 10c/11c and 10d/11d were effectively isolated through column chromatography and their relative stereochemical assignments were established by using ¹H, ¹³C, and 2D NMR (comprising COSY, HMBC, HSQC, and NOESY experiments; Supplementary Information). Chiral HPLC analysis (see SI for the details) confirmed no loss of enantiomeric purity of the 2,6-morpholine products resulting from the synthesis.

At this juncture, we had successfully demonstrated the synthesis of all 12 possible isomers of methyl morpholine-2-acetic acid esters. However, as a synthesis guiding paradigm, Systematic Chemical Diversity also encourages the complete synthesis of possible regioisomers. Thus, for the morpholine heterocycle, we were faced with the additional task of generating a different set of compounds having the acetic acid ester moiety anchored at the 3-position. Starting with the synthesis of the 2-methyl morpholine-3-acetic acid esters, our synthesis commenced from the protected amino acid 12a (Z-Asp (O^tBu)-OH). Our strategy involved converting 12a to an amino alcohol which would then be followed by a key annulation reaction similar to the approach employed for 3-methyl-2-morpholine acetic acid esters (5a/6a and 5b/6b, Scheme 2). To execute this plan, the methylketone (13a) was prepared via Weinreb amide (not shown) and subsequent Grignard reaction with methylmagnesium bromide.³⁰ The ketone 13a was subjected to reduction with sodium borohydride to yield the secondary alcohol as a mixture of diastereomers (~1:1 ratio) 14a. To facilitate the subsequent annulation reaction, we exchanged the Cbz protecting group with Ns by treating with hydrogenation (Pd/C, H₂) followed by sulfonylation with 4-nitrobenzenesulfonyl chloride, which provided 15a. Subjection of 15a to the annulation reaction conditions with bromoethyldiphenylsulfonium triflate and sodium hydride,^{27, 28} resulted in the desired final products 16a/17a (Scheme 4). Surprisingly, the diastereomeric mixture (16a/17a) proved inseparable using conventional silica gel chromatography, which necessitated the use of preparative chiral HPLC to separate the diastereomers (Chiralpak IA, heptane/EtOH). The relative stereochemistry of both diastereomers was established through ¹H, ¹³C, and 2D NMR techniques (COSY, HMBC, HSQC, and NOESY, Supplementary Information). After validating the synthetic route, we repeated the same reaction sequence, starting from Z-D-Asp (O^tBu)-OH 12b, to yield diastereomeric products 16b:17b (opposite enantiomers of 16a:17a, Scheme 4). We assessed the enantiomeric purity of the final products using chiral HPLC and confirmed that no erosion of stereoisomeric purity occurred during the synthetic process.

Scheme 4.

Synthesis of 2-methyl substituted morpholine-3-acetic acid esters; * = enantiomerically pure; Ns = 4-nitrobenzenesulfonyl.

We next proceeded to the synthesis of the 5-substituted-3-morpholine acetic acid esters, where we sought to employ an intramolecular aza-Michael reaction strategy as the key ring-forming step. Functionally, this would involve establishing the acrylate Michael acceptor through O versus N from a starting amino alcohol. To achieve this, we started from N-Boc-(S)-alaninol 18a and performed a selective O-allylation to afford 19a, which was next submitted to a cross metathesis reaction with ethylacrylate and Grubbs’ 2^nd generation catalyst to obtain the aza-Michael precursor 20a.^{10, 31} On removal of the N-Boc group with TFA followed by a basic workup, the desired aza-Michael reaction ensued to furnish the diastereomeric morpholine products in ~3:1 (cis:trans) ratio (from crude NMR) (Scheme 5). In this case, the diastereomers (S2a/S3a) were successfully separated at this stage by column chromatography; sulfonylating the amine with 4-nitrobenzenesulfonyl gave 21a/22a. Assignment of the relative stereochemistry was carried out by ¹H, ¹³C, and 2D NMR (COSY, HMBC, HSQC, and NOESY; Supplementary Information). To obtain the antipodes of 21a/22a, we repeated the synthesis using N-Boc-(R)-alaninol 18b, which successfully generated 21b/22b (Scheme 5). Evaluation of the enantiomeric purity of the 2,5-disubstituted morpholine products by chiral HPLC confirmed that the stereochemical configuration remained intact from the enantiomerically pure starting amino alcohols (details in Supplementary Information).

Scheme 5.

Synthesis of 5- and 6-substituted morpholine-3acetic acid esters; * = enantiomerically pure; Ns = 4-nitrobenzenesulfonyl.

Finally, we focused on 6-substituted-3-morpholine acetic acid esters, which we sought to achieve by switching the regiochemistry of the starting amino alcohols. Accordingly, (S)- and (R)-tert-butyl-(2-hydroxypropyl)carbamates (18c and 18d) were effectively converted into their corresponding aza-Michael precursors 20c and 20d in two steps with good yields using an identical synthetic protocol as described above. After N-Boc removal followed by a basic workup, the desired morpholine products were obtained, which were further sulfonylated to yield 21c/22c and 21d/22d as ~1:4 (cis:trans) mixture of diastereomers (Scheme 5). These diastereomers were efficiently separated by column chromatography and the assignment of the relative stereochemistry of each diastereomer was achieved using ¹H, ¹³C, and 2D NMR (COSY, HMBC, HSQC, and NOESY). Analysis of the enantiomeric purity of the final products by chiral HPLC confirmed no racemization occurred throughout the synthesis (for details see Supplementary Information).

In conclusion, we have successfully demonstrated the application of SCD for structurally expanding the biologically valuable morpholine heterocycle. Starting from readily obtained chiral amino acids (or amino alcohols), SCD drove the production of regiochemically (2,3-, 2,5-, 2,6- and 3,2-, 3,5-, 3,6-) and stereochemically (relative and absolute) diverse morpholine acetic acid esters. In total, 24 distinct morpholine scaffolds were produced. Together, the extensive 3-dimensional vector coverage afforded by these morpholines underscores that a single saturated scaffold can efficiently sample chemical space if comprehensively substituted. These compounds can be directly screened as low-molecular weight fragments (deprotected MW=187) or used as building blocks in medicinal chemistry or library synthesis. Rather than randomly diverse compounds, screening isomers having systematically varied structural relationships will lead to immediate and actionable structure-activity-relationship information. Additionally, these SCD derived morpholines may prove useful in interrogating proteins in conjunction with machine learning/AI algorithms. Efforts toward each of these areas are underway in our labs and will be reported in due course.

Supplementary Material

ASSOCIATED CONTENT

Supporting Information

Detailed experimental procedures and compound characterization data of all new compounds available for free of charge via the Internet at http://pubs.acs.org.

Data Availability Statement

The data underlying this study are available in the published article and its online Supporting Information.