Overcoming the Challenges of Making a Single Enantiomer N-1

Subscriber access provided by ALBRIGHT COLLEGE

Full Paper

Overcoming the Challenges of Making a Single Enantiomer N-1 Substituted Tetrazole Prodrug Using a Tin-mediated Alkylation and Enzymatic Resolution Anne Akin, Mark Barilla, Thomas A Brandt, John Brennan, Kevin E Henegar, Steve Hoagland, Rajesh Kumar, Javier Magano, Emma L McInturff, Asaad Nematalla, David W. Piotrowski, Jared Van Haitsma, Liuqing Wei, Jun Xiao, and Shu Yu Org. Process Res. Dev., Just Accepted Manuscript • DOI: 10.1021/acs.oprd.9b00104 • Publication Date (Web): 09 May 2019 Downloaded from http://pubs.acs.org on May 9, 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

Overcoming the Challenges of Making a Single Enantiomer N-1 Substituted Tetrazole Prodrug Using a Tin-mediated Alkylation and Enzymatic Resolution Anne Akin,a Mark T. Barilla,a Thomas A. Brandt,a* John Brennan,a Kevin E. Henegar,a Steve Hoagland,a Rajesh Kumar,a Javier Magano,a Emma L. McInturff,a* Asaad Nematalla,a David W. Piotrowski,b Jared Van Haitsma,c Liuqing Wei,b Jun Xiao,b Shu Yu,a aChemical

Research and Development, bMedicinal Chemistry, cResearch Analytical, Pfizer Worldwide

Research and Development, Eastern Point Road, Groton, Connecticut 06340 U.S.A. [email protected]

ACS Paragon Plus Environment

1

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

Page 2 of 28

TABLE OF CONTENTS GRAPHIC 1. (Bu3Sn)2O 2. O

I

N N

N N NH N

Br O OEt 3. Enzymatic Resolution

N I

N N via I

N N

N N N SnBu 3 N

N N N N O O

BocN

O

Cl O

N

N 1. Suzuki Coupling 2. Deprotection

F

N

HN

Cl O HCl

F O

B

O N N

N N N N O O

O


2

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


ABSTRACT

The synthesis of proprotein convertase subtilisin/kexin type 9 (PCSK9) inhibitor 3 is described. This complex structure contains a tetrazole modified by a chiral hemiaminal carbonate prodrug. A regioselective tin-mediated alkylation was utilized to access the N-1 alkylated tetrazole isomer, and a highly selective enzymatic hydrolysis efficiently provided the desired prodrug enantiomer. A SuzukiMiyaura coupling was employed for the final fragment union, which was challenging due to base sensitivity of the prodrug. This route was enabled and used to manufacture multikilogram quantities of API 3 in an efficient manner.

KEYWORDS PCSK9 inhibitor, N-1 tetrazole alkylation, Enzymatic resolution, Tin-mediated alkylation, Suzuki Coupling


3


Page 4 of 28

INTRODUCTION The risk of atherosclerosis and coronary heart disease (CHD) is correlated with high levels of the serum LDL-cholesterol (LDL-C) biomarker. Statins such as Lipitor or Crestor are often frontline treatments when pharmaceutical intervention is required to lower LDL-C, but these agents do not always provide sufficient lowering of LDL-C. Inhibition of proprotein convertase subtilisin/kexin type 9 (PCSK9) has emerged as a promising new treatment for decreasing serum LDL-C levels and reducing risk of CHD. Examples of recently approved PCSK9 inhibitors include the injectable monoclonal antibody (mAb) drugs alirocumab1 and evolocumab.2 However, discovery of a small molecule PCSK9 inhibitor with the possibility to deliver it by the more convenient oral route has lagged behind the mAbs. Previous disclosures from the Pfizer laboratories have described small molecule protein translation inhibitors of PCSK9 such as PF-06446846 (1),3,

4

PF-07556769 (2)

5, 6

and PF-06815345 (3).6 Since

PCSK9 is synthesized primarily in the liver, the mode-of-action of the aforementioned inhibitors make them well-suited for application of a tissue-targeted delivery approach. Thus, the liver-targeted variants, 2 and 3, relied on prodrugs of parent zwitterionic tetrazoles to both enhance permeability and allow rapid carboxylesterase 1 catalyzed release and retention of the parent tetrazole in the liver. The identification and assessment of tetrazole prodrug 2 revealed a non-selective cleavage of the prodrug in both the gut and the liver. Subsequent optimization work indicated that not only the point of attachment to the tetrazole but also selection of a specific enantiomer of the prodrug were essential for the identification of a viable liver-targeted prodrug candidate 3. The synthetic challenges associated with the large-scale synthesis of 3 include the regio- and enantioselective formation of a tetrazole hemiaminal carbonate, its participation in the Suzuki-Miyaura coupling and the handling of the base-sensitive hemiaminal carbonates. Herein, we disclose a safe and scalable route directed toward kilogram quantities of 3 to support clinical studies.


4

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


N N

HN

N

Cl O

N N

HN

O

N

HN

Cl O

O

N

N N N N

N

N N

N N

O

F

N N N N O

N N

O PF-06446846 (1)

PF-07556769 (2)

O

PF-06815345 (3)

Figure 1. Small molecule PCSK9 inhibitors Initial synthesis of iodopyrazole tetrazole 6. A conserved approach for the manufacture of PCSK9 inhibitors 1, 2, and 3 installed a functionalized southern fragment by coupling to the amide containing northern fragment (4 or 5, Figure 2). For target 3, this required the bulk manufacture of prodrug modified tetrazole 6. A safe process for the manufacture of iodopyrazole 7 had been developed, allowing access to multi-kilogram quanitites.5 This material, while highly energetic (DSC onset of 201 °C, 1,020 J/g), is not friction or shock sensitive, allowing for safe handling and shipment under standard conditions. At the outset, the synthesis of N-1 substituted tetrazole 6 from parent tetrazole 7 by alkylation posed a significant challenge. The regioselective N-1 alkylation of 5-substituted tetrazoles is not well known, and the formation of the requisite hemiaminal ethyl carbonate in an enantioselective manner would make this task even more difficult.7, 8

N

N

HN

N

HN

N

Cl O

Cl O

4, X = Br 5, X = Bpin

F X I

F N N

N N N N O O

3

N N

I

N N N N O O

O

N N N N H

N N O

7

6


5


Page 6 of 28

Figure 2. General bond disconnection strategy Initially, alkylation of tetrazole 7 with 1-chloroethyl ethyl carbonate was performed in 9 volumes of DMF at 65 oC overnight, providing a 4:1 ratio of N-2:N-1 regioisomer products. Screening experiments with various solvents and bases resulted in a slight improvement to approximately 3:1 ratio of N-2:N-1 regioisomers. The alkylation of des-iodo-7 was also briefly explored, but the less sterically congested tetrazole did not provide any improvement in regioselectivity. For the first GMP synthesis of API, 6 was made as shown in Scheme 1, in which the alkylation was performed in 3 volumes MeCN at reflux to reach completion in 3-5 h. (Safety precaution: While room temperature storage and shipment is supported by TSU data, laboratory samples of carbonate 8 appeared to decompose over time, possibly due to moisture. Decomposition of 8 can generate three moles of gas per mole of 8, resulting in pressurization if stored in a sealed vessel.) Silica gel chromatography was used to separate the desired minor N-1 regioisomer 9 from the undesired major N-2 regioisomer 10, followed by preparative chiral chromatography for separation of the N-1 regioisomer enantiomers, resulting in a 12% isolated yield of enantiopure 6. Initially, 6 was isolated as a thick, viscous oil. Serendipitously, one particular sample solidified after being stored for several days on the bench, and this material was used as seed to develop a crystallization process. It was found that 6 could be readily crystallized from EtOH (1 mL/g) and allowed for the isolation of solids after chromatographic separation. Although the yield of 6 was low, the N-2 regioisomer and the undesired enantiomer of the N-1 regioisomer could be easily recovered by hydrolyzing the prodrug with aqueous NaOH. After pH adjustment, 7 could be isolated from aqueous mixtures via filtration.

Scheme 1. Synthesis of prodrug containing 6 for 1st GMP synthesis ACS Paragon Plus Environment

6

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


O I

N N

N N N N H

Cl

O

O 8 ACN, NEt3

I

N N

I

N N N N O O

Silica gel chromatography O

N N

O

25% yield

Chiral separation EtOH crystallization O

47% yield

9

9

7

I

N N N N O

N-1 regioisomer

N N

N N N N O O

O

6 12% overall yield

+ I

N N

N N N N

O

O O

10 N-2 regioisomer

Improving N-1 regioselectivity for the tetrazole alkylation. While not commercially available, the bromo and iodo analogs of carbonate 8 can be made from 8.9-11 Stability of the alkyl carbonate decreases in moving from Cl to Br to I. In fact, the iodo analog decomposed when stored for a few days at 0 oC. Despite this, improved N-1 selectivity was seen with 1-iodoethyl ethyl carbonate, giving an N1:N-2 ratio of 45:55 at -15 oC using N,N-diisopropylethylamine in MeCN. Selectivity deteriorated at higher temperatures and reactivity slowed at lower temperatures with no improvement in selectivity. In the literature, N-1 substituted tetrazoles are typically made from secondary amides, which, upon activation, react with azide to access the desired tetrazole. For tetrazole 9, attempts to access amide 12, the precursor to activated species thioamide 13 or imidoyl triflate 14, were unsuccessful, instead leading to dioxazinone cyclization products 15a or b (Scheme 2).12, 13 Other protocols (i.e. nitrilium salt reacting with azide or [2+3] cycloaddition of nitrile with an alkyl azide14, 15) that use more forcing conditions were ineffective due to lability of the hemiaminal carbonate ester functionality.

Scheme 2. Alternative approaches to N-1 substituted tetrazole 9. ACS Paragon Plus Environment

7


I

O

1. NaH, THF 2. Cl O 8 O

NH2

N N

I

I

O

O

N H

N N

11

O

S

Activation

O

12 unstable

I

N N

O N

I

O N N

15a

O

NaN3 or TMSN3

O

13

I

N N

N N N N O 9

O

O

or I

O

O

N H

N N

O

Page 8 of 28

O

X N

N N

O O

O

N 15b

O

14 X = OTf, Cl

O

Dioxazinones observed experimentally

Another common protocol for obtaining N-1 substituted tetrazoles is via “exhaustive alkylation,” in which an N-2 substituted tetrazole is generated using a relatively labile substituent, then reacted with a second electrophile to preferentially add to the N-1 site, forming a tetrazolium cation (17, Scheme 3). This is followed by submitting the per-alkylated tetrazole cation salt to conditions that cleave the more labile N-2 substituent to provide the desired N-1 substituted tetrazole.16 Typical electrophiles were employed

for

the

initial

“alkylation”

to

install

a

tri-isopropylsilyl

(TIPS),

[2-

(trimethylsilyl)ethoxy]methyl acetal (SEM), t-butyl, trityl, or para-methoxybenzyl (PMB) group. The TIPS substituted tetrazole was formed as the single N-2 regioisomer and upon alkylation with 1bromoethyl ethyl carbonate gave a 40:60 mixture of N-1:N-2 regioisomers. While encouraging, this result was not superior to the direct reaction of 7 with 1-iodoethyl ethyl carbonate, and the mixture of regioisomers suggests that the silyl group may be labile under the reaction conditions. The SEM substituted tetrazole was formed as a 3:1 mixture of N-1:N-2 regioisomers but did not appear to react in the subsequent alkylation step. The t-Bu and trityl substituted tetrazoles were each formed as a single N2 regioisomer, however, these substrates were unreactive in the second alkylation reaction. Protection with a PMB group resulted in a mixture of regioisomers and did undergo alkylation using 1-iodoethyl ethyl carbonate, however, upon heating, the PMB group remained intact. Scheme 3. Exhaustive alkylation strategy ACS Paragon Plus Environment

8

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


O I

N N N N H

N N

X

R N N N N

I Protection N N

O

O

R N N N N O

I

X= Cl, Br or I

N N

O 7

16

I

Deprotection

N N N N O

N N O

O

17

O

9

The tributyltin variant was more productive. Stannylated tetrazoles have been reported in the literature, and have been shown to participate in alkylation reactions with iodomethane to yield N-1 methylated tetrazole species.17, 18 Reaction of tetrazole 7 with bis-tributyltin oxide yielded stannylated tetrazole 18 (Scheme 4). Using flow NMR to evaluate this step, a ~0.2 ppm shift of the proton at the 3 position of the pyrazole was observed immediately upon mixing 7 and bis-tributyltin oxide. In practice, tetrazole 7 and 0.5 equiv bis-tributyltin oxide were mixed in a solvent (MTBE was chosen for convenience but other solvents such as MeCN, toluene, or EtOAc can be used) and stirred until 7 had completely dissolved (Note: 7 has low solubility in MTBE, so dissolution likely indicated that complexation had occurred). Scheme 4. Exhaustive tetrazole alkylation via stannane intermediate 18 Br I

N N

N N N N H

(Bu3Sn)2O MTBE

I

N N

SnBu3

N N N N

O 19 O

O

I

RT, 36 h N N

O 7

18

I

N N N N O

+ N N

N N N N

O

O O

Crystallize from MeCN

I

N N

N N N N O

O

9

O 10

O

9

80 : 20 N-1 : N-2 regiosiomer

A sample of the stannylated tetrazole 18, normally obtained as a viscous oil that would solidify to a waxy solid, was crystallized from MeCN. Single crystal X-ray data showed a multidimensional array with bridging tributyltin groups between the N-1 and N-3 positions (Figure 3 is a simplified representation of the crystal structure, showing three repeating subunits. Note that N-1 and N-4 positions are equivalent, N-2 and N-3 positions are equivalent.).


9


Page 10 of 28

Figure 3. X-ray crystal structure of 18. The tributyltin modified tetrazole 18 reacted with both 1-iodoethyl ethyl carbonate and 1-bromoethyl ethyl carbonate (19) at room temperature in the absence of base to give ~80:20, N-1:N-2 selectivity. 1Chloroethyl ethylcarbonate (8) reacted poorly with the stannylated tetrazole and gave low conversion. For optimization toward scale-up, 1-bromoethyl ethyl carbonate 19 was used due to superior reagent stability. (Safety precaution: Upon TSU evaluation at 60 °C for 24 h, 1-bromoethyl ethyl carbonate was found to generate significant pressure due to decomposition, and this decomposition event was accelerated in the presence of iron. A refrigerated sample was held for 6 months and showed no degradation by GC analysis. Based on this information, 19 was shipped and stored in glass bottles at 2-8 °C, and excess reagent was disposed of immediately.) Bromocarbonate 19 was added to the concentrated stannylated tetrazole 18 and the mixture was stirred at room temperature for 24 to 48 h. Reactions never showed complete consumption of starting material 7, reaching approximately 90% conversion. The reaction mixture was partitioned between MeCN and heptane to remove the majority of the non-polar Bu3SnBr reaction by-product (heptane layer), followed by addition of aqueous KF to the MeCN phase to precipitate any remaining tin species as the poorly soluble polymeric Bu3SnF. Product 9 was crystallized from MeCN and then recrystallized from MeCN to further reduce the tin levels. The overall yield was typically 60% and the isolated product only contained trace amounts (98% ee (Scheme 5). In early optimization using a mixture of N-1 and N-2 alkylated tetrazoles (9 and 10), the enzyme was found to be highly efficient at hydrolyzing both enantiomers of the N-2 regioisomer (complete in 5 to 6 h) as well as the undesired enantiomer of 9. Scheme 5. Enzymatic resolution with CAL-B I

N N

1. CAL-B, MTBE Phosphate Buffer 2. EtOH, H2O

N N N N O O 9

I

N N

I

N N N N O

O

O

N N

N N N N H

O

6

7

The focus then turned to reaction optimization for scale-up by varying organic co-solvent, temperature, dilution, enzyme loading, reaction time, and impact of reaction rate based on starting material lot variability (mainly residual tin content, because at this point the N-2 regioisomer was minimized due to the optimized alkylation and crystallization from MeCN). During optimization, a co-solvent screen identified MTBE and MeCN as suitable options. MTBE was selected for further development due to ease of product isolation from the biphasic reaction mixture. However, in the presence of MTBE, the performance of CAL-B enzyme was lower compared to the single phase reaction using MeCN. An immobilized CAL-B enzyme gave much higher performance in the MTBE/water biphasic reaction mixture. Further screening of different immobilized preparations led to Novozyme 435 as the optimal enzyme preparation. The final optimized process was run at 50 g/L using 40 wt% Novozyme 435 with 4 volumes MTBE co-solvent along with 16 volumes of pH 7 phosphate buffer at 42 oC for 60 to 65 h. The extended reaction time was required to achieve higher selectivity by complete hydrolysis of the undesired isomer because of significant slowdown of the reaction as the resolution neared completion due to the decreasing concentration of the enzyme substrate (the enantiomer of 6). After work-up and crystallization, yields of 42-45% were obtained on lab scale. Due to the excellent enzyme selectivity, enhanced chiral purity could be achieved by ACS Paragon Plus Environment

12

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


extending reaction time, as background hydrolysis of the desired enantiomer 6 was minimal. As an added benefit, the byproduct of the resolution, tetrazole 7, was extracted into the aqueous phase and could easily be recovered by adjustment to pH 1 by the addition of HCl and isolation by filtration. The recovered 7 did not contain any tin. The work-up of the enzymatic hydrolysis reaction consisted of separation of the aqueous and organic phases followed by removal of the immobilized enzyme from the organic phase by filtration. (See Figure 6 for crude reaction mixture prior to work-up.) The product was isolated by exchanging the solvent from MTBE to EtOH, and the EtOH concentrate was seeded with crystalline 6, cooled, and precipitated further by the addition of water as an anti-solvent. For the scale-up runs of the enzymatic hydrolysis, the chiral purity in the crude reaction mixtures ranged from 97.3-100% ee, and the isolated product contained 182-388 ppm residual tin. With the process improvements to give 80:20 N-1:N-2 regioselectivity in the tin-mediated tetrazole alkylation (65% yield) followed by the enzymatic hydrolytic resolution (42% yield), the overall yield of 6 was improved from 12% in the first GMP campaign to 27% in the second GMP campaign and all chromatographic operations were eliminated.

MTBE Enzyme Water

Figure 6. Picture of lab scale enzymatic resolution mixture ACS Paragon Plus Environment

13


Page 14 of 28

Synthesis of pinacol boronate 5. For the previous PCSK9 lead compound (2, Figure 1), the key coupling between the requisite bromobenzamide 4 could be accomplished via a Negishi coupling of the corresponding iodopyrazole and bromobenzamide 4.5 This same approach was evaluated using iodopyrazole 6. However, upon attempting halogen-metal exchange with i-PrMgCl or activated Zn metal insertion into the carbon-iodine bond, rapid reaction with the ethyl carbonate prodrug gave exclusively ethyl ester 20 (Scheme 6). Scheme 6. Attempted halogen-metal exchange of 6 I

N N

O

N N N N O O

O

iPrMgCl, THF -50 °C O

N N

6

N N N N H 20

The Negishi coupling was also attempted in the reverse direction. Quenching the reaction mixture after halogen-metal exchange of bromobenzamide 4 using either n-BuLi or i-PrMgCl resulted in the des-bromo product, but no desired product was formed when attempting the coupling reaction. Quenching a metallated sample into CD3OD did not show high deuterium incorporation by mass spectrometry, indicating that the lack of reactivity of metallated 4 is likely due to an internal proton quench. Direct activated Zn metal insertion into the ortho-fluoro aryl bromide was also unsuccessful. Since all attempts at a Negishi coupling had failed, the Suzuki conditions identified by medicinal chemistry for analog preparation were explored, requiring pinacol boronate 5 (Scheme 7).5 Amination of 2-bromo-3-chloropyridine (21) with commercially available (R)-1-Boc-3-aminopiperidine (22) via Buchwald-Hartwig amination followed by acylation with 5-fluoro-3-bromobenzoyl chloride (24) furnished the bromobenzamide 4. This was converted to pinacol boronate 5 using standard conditions. Scheme 7. Synthesis of pinacol boronate 5.


14

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

Organic Process Research & Development O

NH2

BocN

Br Cl

22 N

NaOtBu, Pd(OAc)2 Xantphos, PhMe

Cl Br

77% yield

21

N

N

24

F

N

THF, LiHMDS PhMe

Cl

B2(Pin)2, KOAc PdCl2(dppf) 2-MeTHF

Cl O

N

BocN

Cl O

N

BocN

NH

BocN

F

70% yield

68% yield F 23

O

Br

4

B

O

5

Suzuki reaction optimization. With the synthesis of coupling partners 5 and 6 complete, the final challenge was the optimization of the Suzuki coupling to make API precursor 25 (Scheme 8). Scheme 8. Suzuki coupling to give 25 and major byproducts observed during reaction optimization. N I

N

Cl O

N

BocN

N N

N N N N O

O O 6 Pd catalyst, base

F O

B

O

N N

BocN

N

Cl O

Cl O

N

BocN

Cl O

N

BocN

N F F

F N N

O 5

25

F

N N N N O

N N O

N N N N H

O Cl

N

26

BocN

NBoc

N

Cl O

N

F OH 28

27

The main challenges with this reaction were avoiding byproduct formation, including cleavage of the prodrug to give the unsubstituted tetrazole 26, homo-coupling of the boronate ester to form homodimer 27, and oxidation of the boronate ester to the phenol 28. Prodrug hydrolysis was always observed at low levels and was observed at higher levels in some reactions that were run for several hours, so it was important to ensure that the coupling was not held for an extended time. Base choice was also critical to avoid prodrug hydrolysis, and CsF was uniquely effective, while stronger bases (e.g. NaOH) resulted in more 26. Homodimer 27 was typically observed at 10-15% levels before reaction optimization and required some additional understanding to consistently control to low levels as is described below. Phenol 28 was likely the result of peroxide- or oxygen-mediated oxidation of boronate ester 5. These three impurities did not purge well during product isolation via crystallization so further optimization aimed to control the impurities was required. High-throughput screening was performed for the cross coupling, including six Pd0 catalysts and eighteen PdII catalysts. [1,1′-Bis(diphenylphosphino)ferrocene] dichloropalladium (II) (PdCl2(dppf)), [1,1′ACS Paragon Plus Environment

15


Bis(di-cyclohexylphosphino)ferrocene] dichloropalladium (II) (PdCl2(dcypf)),

Page 16 of 28

Bis(tricyclohexylphosphine)

palladium (0) (Pd(PCy3)2:) and dichloro[bis(dicyclohexylphosphinophenyl)ether] palladium (II) emerged

as best catalysts, each showing high conversion and minimal side product formation. Further screening of a range of solvents resulted in selection of PdCl2(dcypf) and Pd(PCy3)2, and optimization and initial scale-ups found best results with Pd(PCy3)2 (5 mol%), 1.2 equiv 6, CsF (3 equiv) in 1,4-dioxane in a 9:1 or 4:1 organic/aqueous ratio at 90 oC.19 In general, the Pd0 catalysts outperformed PdII pre-catalysts to minimize homodimer formation. However, even when the optimized conditions were employed for lab pilots, the amount of homodimer formed was variable. Further studies demonstrated that homodimer 27 formation was related to oxygen content of the reaction solution. Exposure of the catalyst to oxygen prior to addition to the reactor, incomplete degassing of the reaction mixture before catalyst addition, or exposure of the reaction mixture to atmospheric oxygen gave significant levels of homodimer. To better understand the relationship between the amount of homodimer formed relative to oxygen level, a headspace oxygen sensor was employed to monitor oxygen content and evaluate methods for effective degassing. Running the reaction under a nitrogen atmosphere that contained 5,000 ppm of oxygen produced only 0.8% of the homodimer, so this was adopted as a targeted upper limit. Sparging nitrogen directly into the reaction solution prior to catalyst addition made it possible to achieve the target oxygen level. Also, for lab pilots, it was critical to have the catalyst weighed out in a glovebox the same day it was planned to be used. Catalyst weighed in the glovebox and then stored sealed and refrigerated would turn from white to pale green after a few days, providing a clear visual indication that the catalyst had been exposed to oxygen. Once a consistent nitrogen purge protocol was adopted, the reaction ran reliably well with freshly subdivided catalyst. The first kilo-scale run used 20 volumes of 9:1 1,4-dioxane:water and 5 mol% catalyst loading. Before the second kilo-scale run, dilution and catalyst loading were evaluated. The reaction could be run successfully at 5 volumes of 9:1 1,4-dioxane:water and the catalyst loading lowered slightly to 4 mol% to maximize batch throughput. In laboratory experiments, further decrease of catalyst loading resulted in a longer time required to reach completion. Due to concerns of pro-drug ACS Paragon Plus Environment

16

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


hydrolysis to generate 26 and high confidence in purging residual Pd, the relatively high catalyst loading was deemed acceptable for this stage of development. For scale-up, some additional precautions were taken to ensure very low oxygen content. An oxygen sensor was placed in the vent line of the reactor and extensive nitrogen sparging ensured that dissolved oxygen was minimized. The Pd(PCy3)2 catalyst was ordered from the vendor in pre-subdivided ampoules so that on-site subdivision could be avoided and exposure to atmospheric oxygen could be minimized. During scale-up, the sealed ampoules were broken and the entire contents added to the tank via the hatch with a nitrogen sweep. Throughout the addition of catalyst and reagents, the maximum oxygen level observed was 1,000 ppm, but after purging with nitrogen, levels from 0 – 10 ppm were typically observed. The reaction scaled well, reaching >99% conversion. The first and second kilo-scale executions of the Suzuki coupling differed in isolation of crude product 25. For the first campaign, a robust crystallization process had not yet been developed to allow isolation from the crude reaction mixture. A method was developed for chromatographic purification and removal of residual palladium (crude PF-06842115 charged as a 50:50 heptane:acetone solution on 20 w/w% SiO2 with 70:30 heptane:acetone mobile phase). Once the material was pure, crystallization from MTBE was very facile, providing 25 as an MTBE solvate. To avoid chromatography in the second run, a solvent exchange from 1,4-dioxane to 2-MeTHF followed by heptane addition allowed for direct crystallization of product. The crystallization was challenging and required seeding. Initially isolated solids contained organic impurities and residual palladium and tin, so a rework was planned. Crude 25 was re-dissolved in 2-MeTHF, treated with SiliaMetS® Thiol, filtered, and then solvent exchanged to MTBE (99% purity in the crude reaction mixture. The main impurities were the parent zwitterion 29 formed upon prodrug cleavage, the N-2 t-butylated tetrazole 30 (from reaction of parent zwitterion with isobutylene/t-butyl cation), and transesterification with i-PrOH to give the iso-propyl carbonate analog 31 (Scheme 9). Each impurity was present at

Overcoming the Challenges of Making a Single Enantiomer N-1

Recommend Documents