Enantiospecific Synthesis of (3R,4R)-1-benzyl-4-fluoropyrrolidin-3

Subscriber access provided by Nottingham Trent University

Full Paper

Enantiospecific Synthesis of (3R,4R)-1-benzyl-4fluoropyrrolidin-3-amine Utilizing a Burgess-type Transformation Daniel W Widlicka, Alexander Gontcharov, Ruchi Mehta, Dylan Pedro, and Robert North Org. Process Res. Dev., Just Accepted Manuscript • DOI: 10.1021/acs.oprd.9b00245 • Publication Date (Web): 16 Jul 2019 Downloaded from pubs.acs.org on July 16, 2019

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 28 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Organic Process Research & Development

Enantiospecific Synthesis of (3R,4R)-1-benzyl4-fluoropyrrolidin-3-amine Utilizing a Burgesstype Transformation Daniel W. Widlicka*†, Alexander Gontcharov†, Ruchi Mehta††, Dylan J. Pedro†, Robert North†† †

Chemical Research and Development, Pfizer Worldwide Research and Development,

Groton Laboratories, Eastern Point Road, Groton, Connecticut 06340, United States ††

Analytical Research and Development, Pfizer Worldwide Research and Development,

Groton Laboratories, Eastern Point Road, Groton, Connecticut 06340, United States

ABSTRACT: Manufacture of an EGFR inhibitor required the asymmetric synthesis of a key 3,4-trans substituted pyrrolidine suitable for pilot plant scale. The initial synthetic route utilized reagents and intermediates that posed safety concerns due to their energetic potential and

ACS Paragon Plus Environment

Organic Process Research & Development 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

then required SFC chromatography to access the desired single enantiomer. Burgess-type reagents provide tremendous utility in organic synthesis but see limited use on large scale due to high cost and reagent instability. Nevertheless, extensive process development led to a scale friendly process where in-situ formation of a Boc-Burgess reagent enabled access to a chiral cyclic sulfamate from inexpensive materials. React IR monitoring was used to study intermediate stability and enabled processing on multi-kilo scale. The sulfamate was converted to the trans-3-fluoro-4-aminopyrrolidine 1 with complete stereospecificity. Intermediate crystallinity offered purity control points where byproducts and impurities were rejected avoiding the need for chromatography. KEY WORDS: Cyclic sulfamate, Burgess reagent, chiral pyrrolidine, in-situ IR monitoring INTRODUCTION The pyrrolidine moiety is a common building block in pharmaceutically active compounds. While many approaches to 3,4-disubstituted pyrrolidines have been developed, control of stereochemistry is often difficult as this substitution pattern is typically accessed through breaking the symmetry of olefins1, epoxides2 or aziridines3 resulting in racemic product mixtures. Isolation of the desired enantiomer then requires a salt resolution or chiral chromatography. Alternatively, 3,4-pyrrolidinediol has been mono-protected, and then transformed through SN2 substitution; however, this approach is challenged by modest selectivity.4 In our process development work towards a recent oncology candidate for treatment of EGFR-T790M (epidermal growth factor receptor) secondary mutations5, we encountered the difficult synthesis of trans-3-fluoro-4-


Page 2 of 28

Page 3 of 28 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


aminopyrrolidine 1. We describe our development efforts to design a robust, stereospecific process to manufacture 1 through an intermediate cyclic sulfamate. Due to the inherent reagent and intermediate instability of the Burgess inner salts, few examples of these transformations are reported on plant scale. In-situ IR monitoring was used to enable this transformation by evaluating reagent and intermediate stability as well as illustrating the kinetics of the process.

RESULTS AND DISCUSSION 1. Racemic Pyrrolidine Synthesis. The first-generation route to fluoro-pyrrolidine 1 was used to support exploratory toxicology and was processed on a 250 g scale.5 This racemic route proceeded through six linear steps and involved a supercritical fluid chromatographic separation (SFC) to isolate the desired enantiomer (Scheme 1). Starting from the Cbz protected 3-pyrroline 2, the epoxide 3 was formed by treatment with mCPBA followed by ring opening with sodium azide. The azide functionality enabled the alcohol 4 to be displaced by fluoride with retention of stereochemistry by treating with diethylaminosulfur trifluoride (DAST). The amine functionality was then unmasked by reducing the azide 5 with triphenylphosphine. Attempts to separate the trans-isomers of 6 through classical salt resolution were not initially successful, so chiral SFC chromatography was employed. The amine was Boc-protected to enhance the SFC separation, and then after isolation of the desired 3R,4R enantiomer 8, the Boc group was cleaved to give 9. Scheme 1. Racemic Synthesis of Pyrrolidine with SFC Separation



While each of these steps gave reasonable yields, the path was not suitable for processing on a multi-kilo scale. The 3-pyrroline 2 is costly and the route utilized high energy intermediates (epoxide and azide) and reagents (mCPBA, NaN3 and DAST) that are a concern for large scale production. The major driving factor to develop a new route came from the cost and extended lead-time required to process the SFC separation on multikilogram scale. We intended to design a route where we could access the requisite chirality through an inexpensive commodity starting material and then build up the required functionality under safe reaction conditions. 2. Exploring the Cyclic Sulfamate. As an alternative to the first-generation approach to the chiral fluoro-pyrrolidine, 1 can be accessed from a C-2 symmetric pyrrolidinediol 10 (Scheme 2). This diol is readily prepared from tartaric acid, providing an inexpensive and readily available chiral pool raw material for the target.6 One could envision a variety of approaches converting 10 to 1 via an activation and displacement of each of the hydroxyls to the target functionality with proper stereocontrol. Scheme 2 illustrates route scouting toward two of the more straightforward approaches to access the targeted compound. Efforts to enable these


Page 4 of 28

Page 5 of 28 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


routes for kilo-scale processing found mono-activation of the diol to be challenging due to poor selectivity, difficult isolations, and low overall yields. Scheme 2. Diol Selective Activation

A more efficient method for selective double displacement of the two hydroxyls was found in methodology developed by Nicolaou7 which involved conversion of the diol 10 to a cis-cyclic sulfamate 22 by reaction with 2 equivalents of the Burgess reagent 20 (Scheme 3). The reaction proceeds through formation of the bis-sulfamate 21 which, upon intramolecular nucleophilic substitution with inversion of configuration, gives the cyclic product 22. The C2-symmetric nature of the trans-diol results in formation of the single enantiomer of the sulfamate, while activating the remaining C-O bond for a second nucleophilic substitution by a fluoride nucleophile. This sequence delivered the targeted trans-aminofluoride derivative 23. Scheme 3. Cyclic Sulfamate via the Burgess Reagent.



Reduced to practice, diol 10 was treated with 3 eq of the commercially available Burgess reagent 20, resulting in clean conversion to the bis-sulfonate 21. Upon heating, 21 was converted to desired cyclic sulfamate 22 in 44% isolated yield.8 Treatment of 22 with a solution of TBAF at 25 °C led to a facile displacement of the sulfamate oxygen by fluoride9 and then, after sulfate hydrolysis, to the methyl carbamate 23. Analysis of the stereoisomeric composition of the product indicated clean formation of the desired trans(R,R) isomer. Although the chemistry worked as expected, three issues needed to be addressed. The Burgess reagent 20 had poor stability when heated under reaction conditions8 which led to poor conversion and formation of side products. The lacking commercial availability and relatively high cost of 20 on bulk scale were also limiting. Finally, deprotection of methylcarbamate 23 to reveal the amino group turned out to be quite difficult and led to decomposition of the product. The issue of difficult carbamate deprotection was addressed by switching from the traditional methylcarbamate reagent to one of the known analogs.10 The t-butyl analog,11 resulted in a Boc-derivative, was cleanly deprotected and had the advantage of providing crystalline intermediates, whereas others (e.g., benzyl- or t-amyl-) resulted in materials with poor physical properties. While the t-Bu analog of the Burgess inner salt is not


Page 6 of 28

Page 7 of 28 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


commercially available, its preparation from chlorosulfonyl isocyanate, t-BuOH and TEA has been described in the literature.11 In our hands, attempts to prepare and isolate the inner salt (similar to the preparation of the DABCO inner salt described by Armitage)12 resulted in mixtures of the desired zwitterion and triethylammonium chloride which were unstable and of variable composition. We proposed that these isolation issues could be avoided if the inner salt 25 was prepared in-situ and immediately reacted with diol 10. This indeed worked on a lab scale (up to 100 g), the t-butoxy Burgess reagent 2511 was generated by sequential addition of tbutanol and then triethylamine to a solution of chlorosulfonyl isocyanate 24. To this solution was then slowly added diol 10 (no more than 0.5 equivalents), resulting in good conversion to the bis-sulfonylated intermediate 26. Heating of the resulting mixture gave the desired cyclic sulfamate 27 with good conversion (82%) and respectable isolated yield (66%). In preparation for a larger scale campaign, we stressed extended diol addition time to accommodate for the large exotherm associated with that addition. This led to a substantial decrease in conversion to 26. This problem was attributed to the inherent instability of the inner salt under the reaction conditions (Scheme 4). We were able to track the decomposition of 25 through in-situ IR monitoring (Figure 1). The distinct signal at 1709 cm-1 attributed to the carbamate C=O stretch in the inner salt 25 was seen forming instantly upon addition of triethylamine and then losing about a third of its intensity over 1 h at 10 °C. Scheme 4. One Pot Process



Figure 1. Instability of Boc-Burgess Reagent

Isocyanate, 24

Carbamate, 28

Inner Salt, 25

Full conversion of the diol 10 to intermediate 26 was critical to enable a robust process. We investigated an alternate order of addition while using in-situ IR monitoring to determine intermediate stability and found that carbamate 28 was stable at 10 °C for up to 24 h (Figure 2) as seen by the stretch at 1768 cm-1. This intermediate was then added to a second vessel containing the diol 10 and excess TEA. The resulting rapid reaction gave the desired bis-sulfonylated intermediate 26. It was not clear whether this forms through direct sulfonylation of the diol or a transient formation of the inner salt 25, but in either case the instability of the zwitterion was mitigated. The exotherm was easily controlled


Page 8 of 28

Page 9 of 28 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


by the addition rate and upon complete addition of 28, the bis-sulfonyl adduct 26 was fully formed (Scheme 5). Figure 2. IR Trends – Stability of Carbamate

Carbamate

Isocyanate

tBuOH addition

Scheme 5. Two-Pot Process



The second stage of this reaction involved heating 26 to at least 80 °C to facilitate cyclization to the desired product 27. Unfortunately, the rate of ring opening by liberated chloride is competitive with cyclization. As a result, we observed high levels (10-30%) of the chloro side product 29. This degradant can be avoided by removal of the HCl-TEA salt through filtration. Solvent screening found that dioxane gave the best reactivity in formation of bis-adduct 26 and most favorable physical properties for precipitation of TEA-HCl. Upon complete formation of 26 the reaction mixture was filtered, and the cake was rinsed with dioxane. The filtrate was collected in a clean reactor and heated to 85 °C, driving cyclization to completion with minimal formation of the undesired chloride, 29 (1-3%). Opening of the cyclic sulfamate 27 with fluoride occurred readily upon treating it with a solution of TBAF in THF at elevated temperature. Alternative fluoride sources were also studied in this transformation but gave inferior results: inorganic fluorides were quite unreactive, even in the presence of phase transfer catalysts; tetra-n-butylammonium bifluoride in the presence of TEA gave clean conversion but at a substantially slower rate. Fluorination with TBAF initially forms the tetrabutylammonium salt of sulfamic acid 30, which is fully soluble in the reaction mixture. Acidification leads to sequential formation of the zwitterionic species 31, then to the loss of the sulfamate group to generate Bocprotected amine 32 and, finally, to Boc-deprotection of the nitrogen and formation of the free amine salt 1. By careful selection of reagents and conditions, one can choose to stop the process at any stage and isolate a desired intermediate (Scheme 6).


Page 10 of 28

Page 11 of 28 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Scheme 6. Stepwise deprotection

The primary process for material production telescoped the fluorination into the deprotection. Upon complete fluoride substitution, the reaction mixture was treated with a solution of toluenesulfonic acid. This removed both sulfamate and Boc groups and resulted in precipitation of deprotected amine 1 as a bis-tosylate salt. A secondgeneration process was later developed where 27 was treated with HCl and the zwitterion intermediate 31 was crystallized and isolated as an additional purge point to reject process impurities. After treating 31 with additional acid, the HCl salt of deprotected amine 1 was crystallized and isolated. 4. Manufacturing Process for Chiral Pyrrolidine. The manufacturing process accessed the C2-symmetric diol 10 through condensation of L-tartaric acid, an inexpensive feed stock, (Scheme 7) with benzylamine by heating to 120 °C in xylenes while azeotropically drying to form the C2-symmetric imide 33. Upon reaction completion, 33 was directly crystallized from the reaction mixture through controlled addition of MeOH and EtOAc at 60 °C. The crude 33 was suspended in MeOH to enhance the purity and then isolated by filtration in 82% yield. Scheme 7. Second Generation Route



Reduction of imide 33 to 10 was achieved using a variety of reducing agents gave complete conversion to the desired product; however, isolation from aqueous conditions proved difficult. We employed BH3•THF (4 eq) for this reduction to enable an anhydrous workup. Upon completion, the reaction was quenched into MeOH and heated to break borate complexes. Methyl borate and THF were azeotropically displaced by distillation of MeOH13 and then displaced with EtOAc. Diol 10 was granulated, filtered from EtOAc, and isolated in 92% yield. Conversion of 10 to the cyclic sulfamate 27 was achieved through the conditions for the two-pot process as described above. The C2-symmetric bis-sulfonamide adduct 26 cyclized to afford the sulfamate as a single enantiomer with no loss of chiral purity. Several work-up options have been developed to isolate 27. On 100 g scale, we have isolated 27 by crystallization from an EtOAc solution. Upon reaction completion the product was extracted into EtOAc and dioxane removed through aqueous washes, the ethyl acetate solution was concentrated and cooled to 0 °C resulting in product crystallization. This process afforded 27 in 61 – 63% yield and gave a beneficial purge of side products (99.7% LC purity). On a plant scale, formation of 27 was telescoped into the subsequent fluorination step to minimize product loss. Upon completion, dioxane was


Page 12 of 28

Page 13 of 28 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


displaced with EtOAc by vacuum distillation to enable the extractive workup. The watersoluble byproducts of the reaction were removed by aqueous washes, the organic layer was dried over MgSO4 and then solvent was exchanged to THF. This process has consistently delivered 60 – 66% yield based on assay of the final THF solution (85-87% LC purity). A third optimized crystallization process was later developed where water was added as an anti-solvent to the final dioxane reaction mixture to crystallize the crude product which was then cooled, granulated, and filtered. The crude cyclic sulfamate was resuspended in n-heptane, granulated and filtered to afford the desired product 27 in 70% yield with 99% LC purity. Multiple batches of 27 were processed on up to 50 kg scale and gave consistent results with yields ranging from 61 to 71%. Early batches were telescoped as a THF solution into the subsequent fluorination while later batches of 27 were isolated through crystallization from dioxane / water to enhance purity (98.5-98.7%) and simplify downstream processing. The final fluorination and sequential deprotection were conducted as a one-pot process (Scheme 8). The sulfamate 27 was taken up in THF and treated with TBAF (2.5 eq) at 60 °C. The solvent was exchanged with EtOAc and the resulting intermediate treated with ptoluenesulfonic acid at 65 °C to fully deprotect the amine.14 The resulting amine bistosylate salt crystallized from the reaction mixture upon cooling and was isolated by filtration.15 Three batches were processed on a 100 kg scale to afford 1 in 79-81% yield. Crystallization of the fluoro-pyrrolidine bis-tosylate 1 provided an excellent purge point for byproducts as well as two low level process related impurities where fluoride was replaced either by chloride or hydroxide.



Scheme 8. Fluorination and Deprotection

4. Alternative Isolation of Zwitterion An alternative process was designed to isolate the zwitterion intermediate 34 to gain an additional purge point for process impurities and to provide more flexibility in choice of the salt form for the deprotected amine 1. Scheme 9. Zwitterion Formation and Deprotection

The zwitterion 34 was formed by controlled addition of aqueous citric acid at 20 °C to the fluorinated intermediate which then crystallized directly from the reaction mixture.16 This material was isolated by filtration and proved quite stable both as a solid and while suspended in the reaction mixture. Solubility of the zwitterion was low both in organic and aqueous media. This allowed high recovery of the product and an efficient impurity purge of the non-zwitterionic species remained soluble in one of the liquid phases. With the zwitterion isolation developed, it became possible to forgo the isolation of 27 and carry its solution, after the necessary solvent displacement, to the fluoride opening step.


Page 14 of 28

Page 15 of 28 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


This switch in isolated intermediate increased the overall efficiency of the process, as isolation of cyclic sulfamate 27 incurred higher losses due to its high solubility in organic solvents. The isolated zwitterion 34 could be further processed by heating the IPA solution with HCl to remove both the sulfamate and Boc groups, and reveal the desired free amine. Deprotection was clean with both anhydrous and aqueous HCl. In the former case, the deprotected product uncontrollably precipitated as the bis-HCl salt, leading to small particle size and poor filtration rates. Adding water to the system led to slow crystallization of the salt and much improved solid properties. The bis-HCl salt also had the advantage of higher bulk density as well as higher amine activity compared to the bistosylate salt and was thus much more preferred from a material handling perspective.

CONCLUSION We have described a synthesis of trans-3-fluoro-4-aminopyrrolidine 1 with complete chiral retention starting from the inexpensive feed stock of L-tartaric acid. The route utilized a Burgess-type transformation through the in-situ formation of a Boc-Burgess reagent. We demonstrated the solution stability of the chlorosulfonyl carbamate 28 by IR monitoring which allowed for direct sulfonylation of the diol 10 to mitigate instability of the Burgess inner salt 25, demonstrating that this type of transformation can be processed on plant scale. The resulting cyclic sulfamate was ring opened with TBAF and deprotected under acidic conditions to afford the desired fluoro-pyrrolidine 1. This new asymmetric route reduced the overall step count from seven to four, eliminated the need



for chiral separation, and increased overall yield from 11% to 42%. These improvements resulted in increased throughput as well as reduced cost and lead time. This process also replaced all high energy reagents and intermediates with thermally stable materials and eliminated chlorinated solvents to provide safer and more robust processing. EXPERIMENTAL SECTION All starting materials, reagents and solvents were purchased from commercial suppliers and used without further purification. Lab scale reactions were conducted under an atmosphere of nitrogen in glass reactors fitted with an internal temperature probe and overhead stirring. Large scale reactions were conducted under contract at Wuxi AppTec Co manufacturing facility. Reaction completion was determined by reverse phase UPLC analysis on an Agilent1290 Infinity using an Acquity T3 column and monitored at 210 nm wavelength. Nuclear magnetic resonance spectra were collected on a Varian AS400. High-resolution mass spectroscopy data was collected from samples prepared at ~0.1 mg/ml in a diluent composed of 1:1 water/acetonitrile and analyzed by flow injection using a mobile phase of 1:1 acetonitrile/0.1% aqueous TFA at 0.3 ml/min on a Waters Synapt G2 mass spectrometer. Melting points were measured on a Buchi Melting Point B-545 while ramping temperature at 1 °C/min.

(3R,4R)-1-benzyl-3,4-dihydroxypyrrolidine-2,5-dione (33) To a 2000L reactor were charged xylenes (689 kg), L-(+)-tartatic acid (150 kg, 997 mol, 1.1 eq) and benzylamine (97.0 kg, 905 mol, LR) with stirring. The reaction mixture was heated to 120 °C and held for 40 hours and removed water in receiver in parallel. The


Page 16 of 28

Page 17 of 28 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


reaction mix was cooled to 60 °C, a sample was pulled and HPLC analysis indicated that less than 2% of the starting material remained (target

Enantiospecific Synthesis of (3R,4R)-1-benzyl-4-fluoropyrrolidin-3

Recommend Documents