Lipase Catalyzed Synthesis of Enantiopure Precursors and Derivatives for β -Blockers Practolol, Pindolol and Carteolol

: Sustainable methods for producing drugs have been developed. Chlorohydrins as building blocks for several β -blockers have been synthesized in high enantiomeric purity by chemo-enzymatic methods. The yield of the chlorohydrins increased by the use of catalytic amount of base. The reason for this was found to be the reduced formation of the dimeric by-products compared to the use of higher concentration of the base. An overall reduction of reagents and reaction time was also obtained compared to our previously reported data of similar compounds. The enantiomers of the chlorohydrin building blocks were obtained by kinetic resolution of the racemate in transesteriﬁcation reactions catalyzed by Candida antarctica Lipase B (CALB). Optical rotations conﬁrmed the absolute conﬁguration of the enantiopure drugs. The β -blocker ( S )-practolol (( S )- N -(4-(2-hydroxy-3-(isopropylamino)propoxy)phenyl)acetamide) was synthesized with 96% enantiomeric excess ( ee ) from the chlorohydrin ( R )- N -(4-(3-chloro-2 hydrox-ypropoxy)phenyl)acetamide, which was produced in 97% ee and with 27% yield. Racemic building block 1-((1 H -indol-4-yl)oxy)-3-chloropropan-2-ol for the β -blocker pindolol was produced in 53% yield and ( R )-1-((1 H -indol-4-yl)oxy)-3-chloropropan-2-ol was produced in 92% ee . The chlorohydrin 7-(3-chloro-2-hydroxypropoxy)-3,4-dihydroquinolin-2(1 H )-one, a building block for a derivative of carteolol was produced in 77% yield. ( R )-7-(3-Chloro-2-hydroxypropoxy)-3,4-dihydroquinolin-2(1 H )-one was obtained in 96% ee . The S -enantiomer of this carteolol derivative was produced in 97% ee in 87% yield. Racemic building block 5-(3-chloro-2-hydroxypropoxy)-3,4-dihydroquinolin-2(1 H )-one, building block for the drug carteolol, was also produced in 53% yield, with 96% ee of the R -chlorohydrin ( R )-5-(3-chloro-2-hydroxypropoxy)-3,4-dihydroquinolin-2(1 H )-one. ( S )-Carteolol was produced in 96% ee with low yield, which easily can be improved. the from the reactions of 1 with vinyl butanoate as acyl donor to produce butanoates


Introduction
Chiral compounds with one or several stereogenic centers consist of pairs of enantiomers. Many drugs on the market today have one or several stereogenic centers, βblockers normally have one stereogenic center, and then consist of two enantiomers. The enantiomers may have the same effect on the patient, or the enantiomers may have different effects, or worse, one enantiomer may have several unwanted side effects. FDA considers the "wrong" enantiomer as an impurity and demands for pure enantiomers as the active pharmaceutical ingredient (API) in the marketed drugs, not racemates. The demand for enantiomerically pure drugs has increased year-by-year since the 1990s when FDA demanded manufacturers to evaluate the pharmacokinetics of a single enantiomer or mixture of enantiomers in a chiral drug. Quantitative assays for individual enantiomers should be developed for studies in in vivo samples early in drug development. It is postulated that the lower the effective dose of a drug, the greater the difference in the pharmacological Mulik et al. reported in 2016 a four-step synthesis of (S)-practolol in 100% ee with the use of Pseudomonas cepacia sol-gel AK lipase as enantioselective catalyst. However, they report that the produced ester from the transesterification reaction is hydrolyzed and aminated to give the enantiopure building block with S-configuration [14]. According to previous reports, it is the slower reacting enantiomer, (R)-N-(4-(3-chloro-2hydroxypropoxy)phenyl)acetamide, which is aminated to give (S)-practolol. They also claim that the configuration of the chlorohydrin is inverted in the amination step, which will not be the case since the amine is not attacking the stereocenter, but the primary carbon with the chloro atom. Ader and Schneider reported in 1992 enzymatic kinetic resolution of racemic N-(4-(3-chloro-2-hydroxypropoxy)phenyl)acetamide with Pseudomonas sp. to give (R)-N-(4-(3-chloro-2-hydroxypropoxy)phenyl)acetamide which was subsequently aminated to yield (S)-practolol in 30% yield and >99% ee [15]. However, no optical rotation values of the building blocks or the drug have been reported. In order to develop greener and more sustainable processes for drugs with secondary alcohol side chains, we wanted to include the synthesis of practolol despite the regulations of the drug.
Pindolol, 1-(1H-Indol-4-yloxy)-3-(isopropylamino)-2-propanol, was first released for clinical use in USA in 1982. It is a non-selective β-blocker, used in the treatment of high blood pressure, chest pain and irregular heartbeat. Pindolol has its substituents on the aromatic ring in the orthoand meta-position, which is common for non-selective β-blockers. Selective β-blockers usually have a substituent in the para-position on the aromatic ring [4]. Pindolol is also a partial agonist and will therefore slow the resting heart rate less than other β-blockers like atenolol or metoprolol [16]. Pindolol is usually sold under the brand names Visken (Sandoz) or Barbloc (Alpha) and is often used to treat high blood pressure during pregnancy because it does not affect the fetal heart function or blood flow. Although pindolol is a non-selective β-blocker, other uses for the drug have been reported. It has been tested in the treatment of fibromyalgia and related fatigue diseases, as well as in the treatment of depression in combination with selective serotonin reuptake inhibitors [17,18]. Pindolol is a rapidly absorbed drug and after oral ingestion it can be detected in the blood after 30 min. In patients with normal renal function, pindolol has a plasma half-life of three to four hours. The drug is also lipophilic and enters the central nervous system rapidly. Reported side effects include unwanted lowering of heart function or changes in the respiratory system. These side effects are related to its β-adrenergic blocking activity; other side effects have also been reported, such as dizziness, vivid dreams, feeling of weakness or fatigue, muscle cramps, as well as nausea [16]. Precursors of β-blocker pindolol were synthesized by biocatalysis in 2017 by Lima et al. They performed hydrolysis of 2-acetoxy-1-(1H-indol-4-yloxy)-3-chloropropane using lipase from Pseudomonas fluorescens which yielded (2S)-1-(1H-indol-4-yloxy)-3-chloro-2-propanol in 96% ee and (2R)-2-acetoxy-1-(1Hindol-4-yloxy)-3-chloropropane in 97% ee, which was hydrolysed giving 97% ee of the R-chlorohydrin for the synthesis of (S)-pindolol with retention of ee [17]. However, we have some doubts about the stereochemistry in this report which will be discussed.
Carteolol is another β-adrenergic antagonist (β-blocker) manufactured mostly with racemic API and administered as eye-drops for reduction of aqueous production in the eye (glaucoma) [19]. In these patients, an elevated intraocular pressure (IOP) leads to damage to the optic nerve, reducing the visual field gradually until the patient is completely blind. It is the second most common cause of irreversible blindness after age-related macular degeneration in western Europe. In 2010, 2.1 million people worldwide went irreversibly blind because of glaucoma [20]. In 2019, the majority of Norwegian glaucoma patients (68%) were treated with β-blockers betaxolol or timolol, either as single drugs or in combination with other drugs such as prostaglandin analogues or carboanhydrase inhibitors [21].
With the aim of sustainable production of enantiopure β-blockers, we have performed several synthetic strategies with lower amounts of reactants and shorter reaction times than previously reported. The general mechanism of base catalyzed deprotonation of phenolic protons with subsequent nucleophilic attachment of epichlorohydrin has been studied with different bases and different concentrations of epichlorohydrin. Lipase B from Candida antarctica has been shown to catalyze reactions of similar compounds with high ee of both product and remaining starting material (hydrolysis and transesterification reactions) [22,23].

Results
Chlorohydrin building blocks (R)-1a-4a for the synthesis of enantiomers of the βblockers practolol ((S)-1c), pindolol and derivatives of carteolol ((S)-3c-4c) have been synthesized in 92-97% ee by chemo-enzymatic methods (Scheme 1). The highest yields of the racemic chlorohydrins were obtained with 0.3-1 equivalents of base in the deprotonation step of the starting materials 1-4, 2 equivalents of epichlorohydrin, 12-26 h reaction time and 30 • C reaction temperature. The intermediate epoxides 1e-4e were protonated with acetic acid and then opened with lithium chloride. Recently, we reduced the amount of acetic acid from 10 to 5 equivalents giving the same yields of the chlorohydrins. Kinetic resolutions of the racemic halohydrins were performed in different solvents with lipase B from Candida antarctica and vinyl butanoate as the acyl donor. Amination of the R-chlorohydrins (R)-1a and (R)-3a-4a gave the S-β-blockers with preserved or increased ee. Due to the low ee (92%), the amination step of (R)-2a was not performed. Previously, we have published the synthesis of the building block for (S)-atenolol in >98% ee by a similar protocol [7]. With the aim of sustainable production of enantiopure β-blockers, we have performed several synthetic strategies with lower amounts of reactants and shorter reaction times than previously reported. The general mechanism of base catalyzed deprotonation of phenolic protons with subsequent nucleophilic attachment of epichlorohydrin has been studied with different bases and different concentrations of epichlorohydrin. Lipase B from Candida antarctica has been shown to catalyze reactions of similar compounds with high ee of both product and remaining starting material (hydrolysis and transesterification reactions) [22,23].

Results
Chlorohydrin building blocks (R)-1a-4a for the synthesis of enantiomers of the βblockers practolol ((S)-1c), pindolol and derivatives of carteolol ((S)-3c-4c) have been synthesized in 92-97% ee by chemo-enzymatic methods (Scheme 1). The highest yields of the racemic chlorohydrins were obtained with 0.3-1 equivalents of base in the deprotonation step of the starting materials 1-4, 2 equivalents of epichlorohydrin, 12-26 h reaction time and 30 °C reaction temperature. The intermediate epoxides 1e-4e were protonated with acetic acid and then opened with lithium chloride. Recently, we reduced the amount of acetic acid from 10 to 5 equivalents giving the same yields of the chlorohydrins. Kinetic resolutions of the racemic halohydrins were performed in different solvents with lipase B from Candida antarctica and vinyl butanoate as the acyl donor. Amination of the R-chlorohydrins (R)-1a and (R)-3a-4a gave the S-β-blockers with preserved or increased ee. Due to the low ee (92%), the amination step of (R)-2a was not performed. Previously, we have published the synthesis of the building block for (S)-atenolol in > 98% ee by a similar protocol [7]. Analysis of the reaction mixtures from the syntheses of 1a-4a on LC-MS showed that the most abundant by-products in these reactions were the dimers 1d-4d (Scheme 1) of the deprotonated starting materials 1-4. In order to ensure full conversion of the starting materials and to avoid the formation of the dimer by-products in the syntheses, concentrations of base and 2-(chloromethyl)oxirane (epichlorohydrin), reaction time and temperature have been varied. When high concentration of base was used, the intramolecular cyclization of the anions of 1a-4a was observed to boost by increased reaction time; otherwise, the acid and lithium chloride were added immediately after full conversion of the starting materials. Investigations of reaction conditions in synthesis of the racemic practolol precursor 1a are shown in Table 1. When 0.5-1.0 equivalents of sodium hydroxide dissolved in water was used in the reaction of 1 with epichlorohydrin with a reaction temperature of 80 °C for 24 h, only the dimer N,N-(((2-hydroxypropane-1,3-diyl)bis(oxy))bis(4,1-phenylene))diacetamide (1d) was obtained in addition to a small fraction of the epoxide 1e (Table 1). Dimer 1d was characterized by NMR-, MS-and IR-analyses. The chemical shifts for 1d were assigned using 1 H-NMR-, 13 C-NMR-, COSY-, HSQC-and Scheme 1. Building blocks (R)-1a-4a synthesized in 92-97% ee for use in synthesis of the S-enantiomers of the β-blockers practolol, pindolol and derivatives of carteolol ((S)-1c-4c).
Analysis of the reaction mixtures from the syntheses of 1a-4a on LC-MS showed that the most abundant by-products in these reactions were the dimers 1d-4d (Scheme 1) of the deprotonated starting materials 1-4. In order to ensure full conversion of the starting materials and to avoid the formation of the dimer by-products in the syntheses, concentrations of base and 2-(chloromethyl)oxirane (epichlorohydrin), reaction time and temperature have been varied. When high concentration of base was used, the intramolecular cyclization of the anions of 1a-4a was observed to boost by increased reaction time; otherwise, the acid and lithium chloride were added immediately after full conversion of the starting materials. Investigations of reaction conditions in synthesis of the racemic practolol precursor 1a are shown in Table 1. When 0.5-1.0 equivalents of sodium hydroxide dissolved in water was used in the reaction of 1 with epichlorohydrin with a reaction temperature of 80 • C for 24 h, only the dimer N,N-(((2-hydroxypropane-1,3-diyl)bis(oxy))-bis(4,1-phenylene))diacetamide (1d) was obtained in addition to a small fraction of the epoxide 1e (Table 1). Dimer 1d was characterized by NMR-, MS-and IR-analyses. The chemical shifts for 1d were assigned using 1 H-NMR-, 13 C-NMR-, COSY-, HSQC-and HMBC-spectra. The analyses were performed on a 600 MHz NMR instrument from Bruker, Germany, with deuterated dimethyl sulfoxide as solvent. Table 1. A reaction temperature of 80 • C and 0.5-1.0 eq of NaOH dissolved in water favored the formation of the dimer 1d in the reaction of 1 with 2 eq of epichlorohydrin (Scheme 1).

Base
The strategy for avoiding the formation of the dimers 1d-4d and also to generate a high ratio of halohydrin to epoxide was to use only 0.3 equivalents of base. The chlorohydrins 1a-4a were synthesized in 59-77% yield.
The starting material 5-(3-chloro-2-hydroxypropoxy)-3,4-dihydroquinolin-2(1H)-one (4) for the synthesis of carteolol is quite expensive, so we wanted to investigate similar reactions of 7-(3-chloro-2-hydroxypropoxy)-3,4-dihydroquinolin-2(1H)-one (3) as a model substrate. Table 2 entry 3 shows that the highest relative rate of the formation of halohydrin 1a over the formation of the epoxide 1e is obtained with one eq of base. Table 2. Synthesis of N-(4-(oxiran-2-ylmethoxy)phenyl)acetamide, 1a, and N-(4-(3-chloro-2-hydroxypropoxy)phenyl)acetamide, 1e, from paracetamol, 1, with epichlorohydrin and NaOH at room temperature The table shows reaction conditions, equivalents and starting amounts of base (mmol), and the ratio of 1a and 1e in the product mixture, calculated from HPLC chromatograms on Eclipse XDB-C18-column and gradient program (  We saw the same trend in the synthesis of 2a/2e. However, in the synthesis of 3a/3e and 4a/4e, 0.3 eq of base gave the highest yield. When catalytic amounts of base are used, the anions of the halohydrins formed will likely deprotonate a water molecule which regenerates the base for new deprotonations of the starting materials. A plausible mechanism for regeneration of the base in these reactions is shown for the reaction of 7-(3-chloro-2hydroxypropoxy)-3,4-dihydroquinolin-2(1H)-one (3, Scheme 2). After deprotonation of phenol 3 by the base, 3 Anion can react with the least substituted epoxide-carbon in epichlorohydrin (path a) forming 3a Anion . 7-(Oxiran-2-ylmethoxy)-3,4-dihydroquinolin-2(1H)-one, 3e, is formed by a Williamson ether synthesis reaction between 3 and epichlorohydrin, as shown in path b. An internal cyclization of 3a Anion also forms epoxide 3e (path c), while protonation of 3a Anion yields chlorohydrin 3a and regenerates the base allowing for the use of catalytic amounts of base (<1 equivalent).
Addition of lithium chloride and acetic acid before any work-up of all the reactions forming 1a-4a gave higher yields than when the reactions were stopped after the nucleophilic attack of epichlorohydrin in the first step.
Kinetic resolutions of 1a-4a have been performed in different solvents catalyzed by CALB with moderate to high E-values (calculated by E&K Calculator, 2.1b0 PPC) [24], giving moderate to high ee-values of the chiral building blocks. (Figure 1). Optical rotation values of (R)-1a, (S)-1a, (S)-1b, (R)-3a, (S)-3b, (S)-3c, (R)-4a and (S)-4b have not been reported previously, and the determination of optical rotation and predictions of absolute configuration of these pure enantiomers should be of interest to both academia and industry. Compounds with >96% ee are regarded as "enantiopure" by pharmaceutical means. Due to the relatively low ee of (R)-2a in our hands (92% ee) we did not proceed with the synthesis of the pindolol enantiomer; however, we would have expected to retain the ee from the amination of (R)-2a giving (S)-pindolol of 92% ee.
Kinetic resolutions of 1a-4a have been performed in different solvents catalyzed b CALB with moderate to high E-values (calculated by E&K Calculator, 2.1b0 PPC) [24], gi ing moderate to high ee-values of the chiral building blocks. (Figure 1). Optical rotatio values of (R)-1a, (S)-1a, (S)-1b, (R)-3a, (S)-3b, (S)-3c, (R)-4a and (S)-4b have not been r ported previously, and the determination of optical rotation and predictions of absolu configuration of these pure enantiomers should be of interest to both academia and i dustry. Compounds with > 96% ee are regarded as "enantiopure" by pharmaceutic means. Due to the relatively low ee of (R)-2a in our hands (92% ee) we did not procee ported previously, and the determination of optical rotation and predictions of absolute configuration of these pure enantiomers should be of interest to both academia and industry. Compounds with > 96% ee are regarded as "enantiopure" by pharmaceutical means. Due to the relatively low ee of (R)-2a in our hands (92% ee) we did not proceed with the synthesis of the pindolol enantiomer; however, we would have expected to retain the ee from the amination of (R)-2a giving (S)-pindolol of 92% ee. In the amination reactions of (R)-1a and (R)-3a, the ee was retained and (S)-practolol ((S)-1c) and (S)-semi-carteolol ((S)-3c) were produced in 96 and 97% ee, respectively. (R)-4a was obtained in 96% ee, and (S)-4c was obtained in 96% ee from amination of (R)-4a (Table 3). Reaction times of the kinetic resolutions varied by the amounts of lipase added; however, 12 h should be a proper reaction time. Danilewicz and Kemp (1973) reported that the optical rotation of (R)-practolol was +4.3 • at 25 • C [25]. We conclude that the negative rotation of (S)-practolol determined by us (-3.9 • ) accounts for the S-configuration.  We noticed that other research groups have reported data for synthesis of enantiopure practolol and pindolol [14,17]. Reaction time for the enzymatic hydrolyses of the acetate of 1-(1H-indol-4-yloxy)-3-chloro-2-propanol reported by Lima et al. varied from 12 to 25 h and the authors reported an E-value of 30 in the hydrolytic kinetic resolution of the racemic acetylester of 2a using Novozym ® 435 [17]. By using lipase from Pseudomonas fluorescens in acetylation of racemic 2a, an E-value of 11 was obtained, with an ee s of the chlorohydrin 2a of 72% ee s and 69% ee p for 51% conversion, 24 h reaction time at 40°C. However, there is a misunderstanding of stereochemistry in Lima s report. The authors report that hydrolysis of the racemic acetate and transesterification reaction of the chlorohydrin 2a give enantiomers with opposite stereochemistry. The product and the remaining alcohol will have the same configuration in hydrolysis of the acetyl ester and transesterification of the chlorohydrin 2a. Hydrolysis of the ester enantiomer from the hydrolysis of racemic ester should not be the R-acetate of 2a, but the S-ester. We claim that they have produced (S)-2a in 97% ee instead of (R)-2a in their hydrolysis of the acetic ester of 1-((1H-indol-4-yl)oxy)-3-chloropropan-2-ol. The authors must have used the S-halohydrin in the synthesis and would then have achieved (R)-pindolol. The optical rotation value is not reported. In our project we used CALB from SyncoZymes as catalyst and obtained an E-value of 66 in the esterification of 2a. At 53% conversion (24 h), ee s and ee p values were 92% and 81%, respectively. The optimization of this synthesis is underway. We encourage researchers to report both optical rotation values and determination of absolute configuration of enantiomers.

Materials and Methods
All chemicals used in this project are commercially available, of analytical grade and were purchased from Sigma-Aldrich Norway (Oslo, Norway) or vwr Norway (Oslo, Norway). HPLC grade of 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).

Chromatographic Analyses
All analyses were performed on an Agilent HPLC 1100 (Santa Clara, CA, USA). Manual injector (Rheodyne 77245i/Agilent 10 µL loop) and a variable wavelength detector (VWD) set to 254 nm were used.
A New Brunswick G24 Environmental Incubator Shaker (New Brunswick co. Inc., Edison, NJ, USA) was used for enzymatic reactions.

Analyses
NMR-analyses were recorded on a Bruker 400 MHz Avance III HD instrument equipped with a 5 mm SmartProbe Z-gradient probe operating at 400 MHz for 1 H and 100 MHz for 13 C, respectively, or on a Bruker 600 MHz Avance III HD instrument equipped with a 5 mm cryogenic CP-TCI Z-gradient probe operating at 600 MHz for 1 H and 150 MHz for 13 C (Bruker, Rheinstetten, Germany). Chemical shifts are in ppm relative to TMS and coupling constants are in hertz (Hz). 1 H and 13 C NMR spectra can be found from Supplementary Materials. Infrared spectroscopy was performed at a Nexus FT-IR instrument (Madison, WI, USA). Exact masses were analyzed with a Synapt G2-S Q-TOF mass spectrometer from Waters TM (Waters Norway, Oslo, Norway). Ionization of samples was done with an ASAP probe (APCI), and calculation of exact masses and spectra processing were performed with Waters TM Software (Masslynxs V4.1 SCN871). IR and MS spectra details can be found from Supplementary Materials.

Optical Rotations
Optical rotations were measured on an Anton Paar (MCP 5100) polarimeter with a 2.5centimeter-long cell (Dipl. Ing. Houm AS, Oslo, Norway). The analyses were performed at 20-23 • C and the samples were dissolved in different solvents, c = 1.0g/100 mL, if not stated otherwise. Wavelength of the light was 589 nm (D).

Conclusions
The S-enantiomers of practolol, carteolol and a carteolol derivative were produced with ee's of > 96% from the enantiopure chlorohydrins from the CALB-catalyzed kinetic resolutions with preservation of the ee. The remaining enantiomer (in the hereby reported cases, the S-esters) may be converted to the wanted enantiomer by dynamic kinetic resolution, which we do not report here. The syntheses of the chlorohydrins 1a-4a have been optimized with reduced amount of base, lowering of reaction temperature and lowering of reaction time compared to previously reported methods, giving moderate to high total yields. With the use of a catalytic amount of base and shorter reaction time, the formation of by-products was reduced. We propose a mechanism for the regeneration of the base in these reactions. We struggled to reproduce the synthesis of the enantiopure chlorohydrin as precursor for enantiopure carteolol, however, we have now obtained 96% ee, which we are in the process of improving. Absolute configurations of the produced enantiomers were determined based on both optical rotation values and comparison of CALB preference for one stereoisomer of similar secondary alcohols. Determination of absolute configuration is of utmost importance when reporting data of enantiopure compounds. We report here several enantiopure compounds which have not been reported previously.

Data Availability Statement:
The data presented in this study are available online.