Green Chemo-Enzymatic Protocols for the Synthesis of Enantiopure β -Blockers ( S )-Esmolol and ( S )-Penbutolol

: The β -blocker ( S )-esmolol, has been synthesized in 97% enantiomeric excess and 26% total yield in a four-step synthesis, with a transesteriﬁcation step of the racemic chlorohydrin methyl 3-(4-(3-chloro-2-hydroxypropoxy)phenyl)propanoate, catalysed by lipase B from Candida antarctica from Syncozymes, Shanghai, China. The β -blocker ( S )-penbutolol, has been synthesized in 99% enantiomeric excess and in 22% total yield. The transesteriﬁcation step of the racemic chlorohydrin 1-chloro-3-(2-cyclopentylphenoxy)propan-2-ol was catalyzed by the same lipase as used for the esmolol building block. We have used different bases for the deprotonation step of the starting phenols, and vinyl butanoate as the acyl donor in the transesteriﬁcation reactions. The reaction times for the kinetic resolution steps catalysed by the lipase varied from 23 to 48 h, and were run at 30–38 ◦ C. Speciﬁc rotation values conﬁrmed the absolute conﬁguration of the enantiopure drugs, however, an earlier report of the speciﬁc rotation value of ( S )-esmolol is not consistent with our measured speciﬁc rotation values, and we here claim that our data are correct. Compared to the previously reported syntheses of these two enantiopure drugs, we have replaced toluene or dichloromethane with acetonitrile, and replaced the ﬂammable acetyl chloride with lithium chloride. We have also reduced the amount of epichlorohydrin and bases, and identiﬁed dimeric byproducts in order to obtain higher yields. and ( S )- 10 ∙ HCl. The kinetic resolutions were catalysed by CALB from Syncozymes in dry acetonitrile. Specific rotations [𝛼] (cid:3005)(cid:3021) were determined at 20–23°C in different solvents with c = 1. For additional parameters, see Materials and Methods. The yields for each compound in the table are for each reaction step. The overall yield for ( S )-esmolol (( S )- 5 ) is 26%, and for ( S )-penbutolol hydrochloride (( S )- 10 ∙ HCl) the overall yield is 20%.


Introduction
We have previously developed efficient synthesis protocols for the syntheses of single enantiomers of several β-adrenergic receptor blockers and their chiral building blocks [1][2][3]. We present here the syntheses of enantiopure (S)-esmolol and (S)-penbutolol ( Figure 1), with focus on green chemistry principles in every reaction step. By the use of kinetic resolution, high enantiomeric excess of the corresponding secondary alcohols as building blocks for such compounds can be obtained. The only drawback with this method is that the enantiopure product can only be obtained in a 50% yield. Dynamic kinetic resolution can be used to improve the yields of the kinetic resolutions, which will lower the amount of waste produced, and thus give an even greener synthesis of the enantiopure drugs [4]. Esmolol is a hydrophilic β1-adrenergic receptor blocker, which has a rapid onset and a short duration of action [5,6]. The drug is administered intravenously, and is widely used in the treatment of hypertension, cardiac arrhythmia, and angina pectoris. The mos potent enantiomer (eutomer) of esmolol is (S)-esmolol [7]. The drug is manufactured a  We also give more in-depth information about these syntheses, which we have found to be missing in previous reports. This includes an accurate measurement of the specific rotation of (S)-esmolol and precursors of both (S)-esmolol and (S)-penbutolol, and also the characterisation of a dimeric by-product formed in the synthesis path to (S)-esmolol. To limit formation of by-products, knowledge of their identity is essential, which is why we also provide plausible mechanisms for the formation of these dimers.
Esmolol is a hydrophilic β 1 -adrenergic receptor blocker, which has a rapid onset and a short duration of action [5,6]. The drug is administered intravenously, and is widely used in the treatment of hypertension, cardiac arrhythmia, and angina pectoris. The most potent enantiomer (eutomer) of esmolol is (S)-esmolol [7]. The drug is manufactured as Brevibloc ® , with a racemic active pharmaceutical ingredient (API). Two methods for the synthesis of (S)-esmolol have been reported. Narsaiah and Kumar obtained the epoxide (S)-methyl 3-(4-(oxiran-2-ylmethoxy)phenyl)propanoate in 94% enantiomeric excess (ee) by kinetic resolution of the racemic epoxide, catalysed by Jacobsens catalyst. Subsequent amination gave (S)-esmolol, however, the authors do not report any ee value of the product [8]. Banoth and Banerjee have reported a non-enzymatic and a chemo-enzymatic route to (S)-esmolol. Commercially available (R)-epichlorohydrin was used as a starting material, however, by this method, (S)-esmolol was obtained only in 93% ee. In a kinetic resolution of the chlorohydrin methyl 3-(4-(3-chloro-2-hydroxypropoxy)phenyl)-propanoate with vinyl acetate in toluene, lipase from Pseudomonas cepacia was used as the catalyst. The chlorohydrin was obtained in 98% ee and is reported to have R-configuration. Amination of the chlorohydrin gave (S)-esmolol in 98% ee [7]. Substituting toluene for a safer and more sustainable solvent in the enzymatic step is desired for a greener synthesis. Acetonitrile has lower toxicity and environmental impact than toluene, and would be a preferred alternative [9], however, Banoth and Banerjee did not find acetonitrile to give high enantioselectivity when using lipase from Pseudomonas cepacia [7].
Penbutolol is a non-selective β-blocker used in the treatment of hypertension. The drug inhibits both β 1 -and β 2 -adrenergic receptors in the heart and in the kidneys. Betapressin ® is manufactured with the racemic API. Penbutolol sulphate is manufactured as Levatol ® with enantiopure (S)-penbutolol sulphate as a prodrug giving (S)-penbutolol when it enters the body. There are several methods reported for the synthesis of (S)-penbutolol, however, many of these methods use expensive enzymes and resolving agents, and suffer from low yields and low enantiomeric purity. As early as 1984, Hamaguchi et al. obtained (S)-penbutolol in 100 % ee with lipase Amano 3 as a catalyst in the hydrolysis of the racemic building block 3-(tert-butyl)-5-(hydroxymethyl)oxazolidin-2-one. However, this method has many steps, and although the lipase is efficient, it is no longer listed on the market [10]. In a kinetic resolution of the corresponding chlorohydrin using lipase from Pseudomonas sp., Ader et al. obtained (S)-penbutolol, however, only in 91% ee [11]. (S)-Penbutolol has also been obtained in 95% ee by the use of Sharpless asymmetric dihydroxylation. The drawback of this protocol is the use of toxic and expensive catalysts, and the ee of the (S)-penbutolol obtained is not optimal. The authors also report an efficient synthesis of the starting material 2-cyclopentylphenol, which today is available from Merck, but is quite expensive. [12]. Klunder et al. reported in 1989, the synthesis of (S)-penbutolol in 86% ee by the addition of enantiopure (2S)-glycidyl tosylate to a penbutolol precursor [13]. Kan et al. have reported a synthesis of (S)-penbutolol hydrochloride from racemic 5-acyloxymethyl-3alkyl-2-oxazolidinones resolved with lipases or microorganisms [14].

Synthesis of Racemic Chlorohydrin 3 (for Esmolol)
The chlorohydrin (R)-methyl 3-(4-(3-chloro-2-hydroxypropoxy)phenyl)-propanoate, (R)-3, which is a chiral building block in the synthesis of the β-blocker (S)-esmolol ((S)-5), has been synthesised in 97% ee (Scheme 1). A deprotonation of the commercial phenol methyl 3-(4-hydroxyphenyl)propanoate (1a) gave the corresponding alkoxide, which in reacting with epichlorohydrin, gave the chlorohydrin 3. A transesterification reaction of 3, catalysed by lipase B from Candida antarctica, gave the corresponding S-butanoate ester (S)-4 and the R-chlorohydrin (R)-3. From the reaction of (R)-3 with isopropylamine, (S)-esmolol was synthesised in 97% ee and 26 % overall yield. We have previously shown that the type and concentration of base used in the reaction of phenolic starting materials with epichlorohydrin strongly influences the ratio of epoxide vs. chlorohydrin, and also the ratio of by-products formed in these reactions [3]. In the synthesis of epoxide 2 and chlorohydrin 3, sodium hydroxide favors the formation of chlorohydrin 3 over the epoxide. However, full conversion of the starting material 1a was not obtained. We have previously used catalytic amounts of the base in the syntheses of similar compounds, but in the deprotonation of the phenolic starting material 1a, this did not give sufficient conversion. By the use of potassium carbonate, epoxide 2 was obtained in 68% yield and with the full conversion of 1a. Ring opening of epoxide 2 was performed by the protonation with acetic acid and opening with lithium chloride in acetonitrile to give chlorohydrin 3 in 96% yield for this step. We have managed to increase the yield from the previously reported 92% [7] in the epoxide ring opening, with the use of LiCl and acetic acid in acetonitrile instead of the highly flammable acetyl chloride and less preferable solvent dichloromethane, at the expense of slightly longer reaction times. reacting with epichlorohydrin, gave the chlorohydrin 3. A transesterification reaction of 3, catalysed by lipase B from Candida antarctica, gave the corresponding S-butanoate ester (S)-4 and the R-chlorohydrin (R)-3. From the reaction of (R)-3 with isopropylamine, (S)esmolol was synthesised in 97% ee and 26 % overall yield. We have previously shown that the type and concentration of base used in the reaction of phenolic starting materials with epichlorohydrin strongly influences the ratio of epoxide vs. chlorohydrin, and also the ratio of by-products formed in these reactions [3]. In the synthesis of epoxide 2 and chlorohydrin 3, sodium hydroxide favors the formation of chlorohydrin 3 over the epoxide. However, full conversion of the starting material 1a was not obtained. We have previously used catalytic amounts of the base in the syntheses of similar compounds, but in the deprotonation of the phenolic starting material 1a, this did not give sufficient conversion. By the use of potassium carbonate, epoxide 2 was obtained in 68% yield and with the full conversion of 1a. Ring opening of epoxide 2 was performed by the protonation with acetic acid and opening with lithium chloride in acetonitrile to give chlorohydrin 3 in 96% yield for this step. We have managed to increase the yield from the previously reported 92% [7] in the epoxide ring opening, with the use of LiCl and acetic acid in acetonitrile instead of the highly flammable acetyl chloride and less preferable solvent dichloromethane, at the expense of slightly longer reaction times.

Characterisation of by-Products in Synthesis of Esmolol Precursors
A dimeric by-product was observed in the reaction between deprotonated phenol 1a, epichlorohydrin, and potassium carbonate to form epoxide 2 (Schemes 1 and 2). LC-MS

Characterisation of by-Products in Synthesis of Esmolol Precursors
A dimeric by-product was observed in the reaction between deprotonated phenol 1a, epichlorohydrin, and potassium carbonate to form epoxide 2 (Schemes 1 and 2). LC-MS analysis gave a peak with m/z = 439.2, molecular formula C 23 H 28 O 7 Na, which corresponds to 3,3 -(((2-hydroxypropane-1,3-diyl)bis(oxy))bis(4,1-phenylene))dipropanoate (3d) with a molecular mass of 416.45 g/mol. By purification of the reaction mixture using flash chromatography in order to obtain pure epoxide 2, dimer 3d was isolated in 6% yield with Catalysts 2022, 12, 980 4 of 12 a purity of 99%. Characterisation of 3d was performed by 1 H-, 13 C-, H,H COSY-, HSQCand HMBC NMR spectroscopy, with deuterated chloroform as the solvent. 1 H-and 13 C-NMR spectra for dimer 3d are given in the Supplementary Materials, in addition to the achiral HPLC chromatogram. We have observed the same type of dimers in the synthesis of similar β-blocker precursors [3], and here we propose plausible mechanisms for the 3d dimer formation, see Scheme 2. We suggest that using a larger amount of epichlorohydrin could limit the amount of dimer formed, as the alkoxide 1a Anion would have better access to epichlorohydrin as opposed to the products 2 and 3, which could improve the yield of the first reaction step. However, we have managed to reduce the epichlorohydrin from 5 equivalents in similar syntheses reported by Bevinakatti and Banerji in 1992 [15] to 2 equivalents, also used in our previous report on the syntheses of several β-blockers [3].
Catalysts 2022, 12, x FOR PEER REVIEW 4 of 12 analysis gave a peak with m/z = 439.2, molecular formula C23H28O7Na, which corresponds to 3,3′-(((2-hydroxypropane-1,3-diyl)bis(oxy))bis(4,1-phenylene))dipropanoate (3d) with a molecular mass of 416.45 g/mol. By purification of the reaction mixture using flash chromatography in order to obtain pure epoxide 2, dimer 3d was isolated in 6% yield with a purity of 99%. Characterisation of 3d was performed by 1 H-, 13 C-, H,H COSY-, HSQC-and HMBC NMR spectroscopy, with deuterated chloroform as the solvent. 1 H-and 13 C-NMR spectra for dimer 3d are given in the Supplementary Materials, in addition to the achiral HPLC chromatogram. We have observed the same type of dimers in the synthesis of similar β-blocker precursors [3], and here we propose plausible mechanisms for the 3d dimer formation, see Scheme 2. We suggest that using a larger amount of epichlorohydrin could limit the amount of dimer formed, as the alkoxide 1aAnion would have better access to epichlorohydrin as opposed to the products 2 and 3, which could improve the yield of the first reaction step. However, we have managed to reduce the epichlorohydrin from 5 equivalents in similar syntheses reported by Bevinakatti and Banerji in 1992 [15] to 2 equivalents, also used in our previous report on the syntheses of several β-blockers [3].
Two suggested mechanistic pathways for the formation of dimer 3d. Reaction mechanism (a) shows a nucleophilic attack on chlorohydrin 3 by alkoxide 1aAnion, mechanism (b) shows a nucleophilic attack on epoxide 2 by alkoxide 1aAnion.

Lipase-Catalysed Kinetic Resolution of Chlorohydrins 3 and 8
Kinetic resolution of the chlorohydrins 3 and 8 was catalysed by lipase B from Candida antarctica (CALB) in dry acetonitrile with vinyl butanoate as the acyl donor. This gave E-values of 157 and 183, respectively (calculated by E&K Calculator, 2.1b0 PPC) [14] (Figures 2 and 3, repectively), and ee-values of 87% for S-ester (S)-4 and 97-99% of the R-chlorohydrins (R)-3 and (R)-8, as described above. The ee-values were retained upon the conversion of the enantiopure chlorohydrins to the respective drugs, see Table 1. The reaction times for the transesterification reactions for obtaining (R)-3 and (R)-8 were 23 and 48 h, respectively. The use of acetonitrile as the solvent in these kinetic resolutions makes the syntheses greener than the previous reports using toluene [7], and here we have shown that acetonitrile gives high selectivity for CALB in the synthesis of R-chlorohydrins (R)-3 and (R)-8.

Scheme 2.
Two suggested mechanistic pathways for the formation of dimer 3d. Reaction mechanism (a) shows a nucleophilic attack on chlorohydrin 3 by alkoxide 1a Anion , mechanism (b) shows a nucleophilic attack on epoxide 2 by alkoxide 1a Anion .

Lipase-Catalysed Kinetic Resolution of Chlorohydrins 3 and 8
Kinetic resolution of the chlorohydrins 3 and 8 was catalysed by lipase B from Candida antarctica (CALB) in dry acetonitrile with vinyl butanoate as the acyl donor. This gave E-values of 157 and 183, respectively (calculated by E&K Calculator, 2.1b0 PPC) [14] ( Figures 2 and 3, repectively), and ee-values of 87% for S-ester (S)-4 and 97-99% of the R-chlorohydrins (R)-3 and (R)-8, as described above. The ee-values were retained upon the conversion of the enantiopure chlorohydrins to the respective drugs, see Table 1. The reaction times for the transesterification reactions for obtaining (R)-3 and (R)-8 were 23 and 48 h, respectively. The use of acetonitrile as the solvent in these kinetic resolutions makes the syntheses greener than the previous reports using toluene [7], and here we have shown that acetonitrile gives high selectivity for CALB in the synthesis of R-chlorohydrins (R)-3 and (R)-8.  with vinyl butanoate as the acyl donor. Enantiomeric excess of the remaining substrate (ee S , red circles) and enantiomeric excess of the product ester (ee P , blue circles) is shown in percent plotted against conversion in percent. The red and the blue curves are generated from the experimental values of ee S and ee P , respectively. The E-value was calculated to be 157. E-values were calculated from E&K Calculator 2.1b0 PPC [16].

Figure 2.
Graphical representation of the kinetic resolution of chlorohydrin 3 with CALB in dry acetonitrile with vinyl butanoate as the acyl donor. Enantiomeric excess of the remaining substrate (eeS, red circles) and enantiomeric excess of the product ester (eeP, blue circles) is shown in percent plotted against conversion in percent. The red and the blue curves are generated from the experimental values of eeS and eeP, respectively. The E-value was calculated to be 157. E-values were calculated from E&K Calculator 2.1b0 PPC [16].

Synthesis of (S)-Esmolol ((S)-5)
The R-Chlorohydrin (R)-3 was converted to (S)-esmolol ((S)-5) in 92% yield by amination with isopropylamine in methanol. The ee was retained in the conversion. For the ee to be retained, it is important that the S-ester ((S)-4) and R-chlorohydrin (R)-3 are completely separated during the flash chromatography separation of the crude mixture from the enzymatic kinetic resolution step. If the separation was not complete, we observed a lowering of the ee. We suggest that this lowering of the ee could be a result of aminolysis of the S-ester, followed by amination of the resulting S-chlorohydrin to give the unwanted (R)-esmolol ((R)-5).

Specific Rotation of Pure Enantiomers
Specific rotation values for the enantiopure chlorohydrin (R)-3 or the corresponding ester (S)-4 have not been reported previously. The absolute configuration of compounds (R)-3, (S)-4, and (S)-5 shown in Table 1 was determined by the known enantioselectivity of CALB towards similar compounds, which has been previously reported [3,17]. In 2011, Narsaiah and Kumar reported a specific rotation for the S-enantiomer of esmolol of [α] 20 D = +4.50 (c 1, CHCl 3 ) [8], while we report [α] 20 D = -6.80 (c 1.04, CHCl 3 ) for the S-enantiomer of esmolol ((S)-5) in 97% ee. Narsaiah and Kumar did not report the enantiomeric excess of their (S)-esmolol, nor how the absolute configuration was determined. We claim that our measurements are correct, and that (S)-esmolol is levorotatory. The specific rotation value of (R)-8 has not been reported previously. The absolute configuration of compound (R)-8 was determined by the enantioselectivity of CALB, which we have reported previously [3,17]. The specific rotation of (S)-penbutolol as a free base has been reported by Phukan

Chemicals and Solvents
All chemicals used in this project are commercially available, of analytical grade and were purchased from either Sigma-Aldrich Norway AS (Oslo, Norway), or VWR International AS, Norway (Oslo, Norway). HPLC-grade solvents were used for the HPLC analyses. Dry MeCN was acquired from a solvent purifier, MBraun MD-SPS800 (München, Germany), and stored in a flask containing molecular sieves (4Å).

TLC Analyses and Column Chromatography
TLC analyses were performed on Merck silica 60 F 254 and detection with UV at λ = 254 nm. Flash chromatography was performed using silica gel from Sigma-Aldrich Norway AS (Oslo, Norway) (pore size 60 Å, 230-400 mesh particle size, 40-63 µm particle size).

Enzymes
Candida antarctica Lipase B (CALB) (activity ≥ 10,000 PLU/g, lot#20170315) immobilised on highly hydrophobic macro porous resin, and produced in fermentation with genetically modified Pichia pastoris, was a gift from SyncoZymes Co. Ltd. (Shanghai, China). The enzyme reactions were performed in a New Brunswick G24 Environmental Incubator Shaker from New Brunswick Co. (Edison, NJ, USA).

Liquid Chromatography-Mass Spectroscopy (LC-MS) of Esmolol Dimer 3d
LC-MS analysis of by-product 3d was performed on an AQUITY UPLC I

Mass Spectrometry Analysis of by-Product 3d
Exact mass of 3d was determined with a Synapt G2-S Q-TOF mass spectrometer from Waters TM (Waters Norway, Oslo, Norway). Ionization of the sample was performed with an ASAP probe (APCI), and the calculation of exact masses and spectra processing were performed with Waters TM Software (Masslynxs V4.1 SCN871). See Supplementary Materials for spectra.

Optical Rotation
Optical rotation values were performed with an Anton Paar MCP 5100 polarimeter from Dipl.Ing. Houm AS (Oslo, Norway), and a wavelength of 589 nm (D), for values, see single enantiomers for specific rotation values.

Absolute Configurations
The absolute configuration of (S)-esmolol ((S)-5) was determined by the enantioselectivity of CALB which we have reported previously [3,17]. We report here a specific rotation of (S)-esmolol ((S)-5), disputing the previously reported value [8]. Specific rotation values of (R)-3, (S)-4 and (R)-8 have not been reported previously and were determined by the enantioselectivity of CALB, which we have reported previously [3,17]. The absolute configuration of (S)-penbutolol (S)-10 and (S)-10·HCl were determined by comparing the specific rotation with previously reported data [12,14].