Combination of a UPO‐Based Epoxidation With a Subsequent Ring‐Opening Reaction for the Synthesis of Amino Alcohols

Abstract

This study presents the design to aim for an atom‐efficient chemo‐enzymatic synthesis route towards aromatic amino alcohols, based on an unspecific peroxygenase‐catalysed oxyfunktionalisation of styrene and a highly atom‐efficient conversion of the resulting epoxide with nucleophiles and electrophiles, respectively. This synthesis strategy features a simple two‐step approach, and the practicality has been demonstrated at a semi‐preparative scale. In a first step, the unspecific peroxygenase oxyfunctionalises the substrate, forming an epoxide. Due to its properties, the latter can serve as a starting material for the conversion into a wide range of products, thereby enabling the production of amino alcohols that are otherwise often difficult to synthesise. The shown concept features a one‐pot two‐step approach, depending on the respective ring‐opening reagent. This method aims for a direct synthesis route for the pharmaceutical industry with good yields and high atom efficiency.

The synthesis of aromatic amino alcohols via a combination of an unspecific peroxygenase‐catalysed reaction, oxyfunctionalization, and addition of selected nucleophiles and electrophiles in a highly atom‐efficient manner was investigated.

1. Introduction

Approximately 40% of all pharmaceuticals and agrochemicals, like fertilisers or pesticides, contain at least one amine moiety in some form, and thus efficient synthesis strategies towards these valuable compounds are highly desired [1]. Herein, vicinal amino alcohols are especially relevant and typically require tailor‐made synthesis strategies due to the introduction or presence of the adjacent hydroxyl group [2, 3]. Important examples of the class of chemicals include the antitumor reagent taxol, ethambutol for the treatment of tuberculosis and HIV‐suppressing drugs such as elvitegravir or indinavir [4]. Interestingly, arylethanolamines and the corresponding derivatives are a very common structural element in hormones and related pharmaceuticals such as adrenaline, norephedrine and isoprenaline, to name three of the most prominent examples that contain this important vicinal amino alcohol pattern [5, 6]. Please note that not necessarily enantiopure compounds are desired, e.g. isoprenaline or propranolol are used as racemates in medicine. This can even be extended to N‐heterocyclic compounds based on morpholines and quinine that are also known in various use cases, e.g. as an effective malaria treatment [6]. Consequently, arylethanolamines were chosen as the major target compounds within this study and were synthesised from readily available styrene 1 via the corresponding epoxide 2, which itself is converted in a secondary step to the final arylethanolamines 4 or the corresponding derivatives 3.

The key reaction step is the formation of the epoxide 2, which is, as a result of the ring tension of the oxirane structure, quite reactive and therefore a versatile precursor that can be used to synthesise a wide range of different organic compounds [5, 7, 8, 9, 10, 11]. It is possible to open this heterocycle by using both nucleophiles and electrophiles, and, depending on the respective reagent and overall structure of the epoxide, various product branches can be made available [12]. In this context, the ring opening of epoxides has already been investigated in various ways [13]. This could be achieved with the aid of metal‐containing catalysts, but activation with the aid of an H‐bond donor such as ionic liquids or in the presence of phosphonium salts and silicon‐ or boron‐based catalysts also leads to yields of 80% to 90%. Because the ring‐opening reagent is integrated in the product, the atom efficiency has to be considered to be very high. In order to provide epoxides that can be converted to different arylethanolamines, a possible pathway is the epoxidation of unsaturated hydrocarbons, as it is carried out in this work.

Epoxidations of alkenes can usually be carried out using peroxycarboxylic acids (Prilezhaev reaction) [14, 15] or oxidising agents as well as catalysts containing heavy metals (Jacobsen epoxidation) [16]. The Prilezhaev reaction can be further extended by an enzymatic sub‐step using a lipase, in which the peroxycarboxylic acid is continuously replenished, as reported by Manea et al. [17] This approach is suitable not only for the conversion of simple unsaturated hydrocarbons, but also of allyl ethers or terpenes [18, 19]. In addition to lipases such as Novozym 435 [19, 20], which have been used by several authors, monooxygenases and haloperoxidases are also suitable for the synthesis of epoxides [21, 22, 23]. These enzyme groups require cofactors (e.g. FAD) to exert their catalytic effect, which is a minor disadvantage compared to other enzymes or chemical epoxidation. Nevertheless, all of the enzymes mentioned are capable of performing alkene epoxidation. The chemo‐enzymatic synthesis of amino alcohols via the ring opening of epoxides has also been investigated by various research groups and can also be achieved with the aid of halohydrin dehalogenases, among other methods [24, 25, 26]. This group of enzymes enables ring opening with the formation of substituted alcohols with a variety of halides and pseudohalides, such as azide, thiocyanate or nitrite, which are incorporated into the resulting product during the reaction [24]. In addition, halohydrin dehalogenases catalyse both the dehalogenation of a vicinal halohydrin to the corresponding epoxide and the subsequent ring opening, making them a very efficient option for this type of reaction. The use of a hydrolase to catalyse the hydrolytic decomposition of epoxides has also been investigated and, in combination with a halohydrin dehalogenase, represents an elegant synthetic route for various products [27]. In addition, various enzyme cascades are known that use transaminases to obtain amino alcohols [28]. These synthesis pathways do not involve an epoxide intermediate, but transfer an amine to the position of a ketone in various molecules and require an additional donor molecule for this purpose, which in turn is converted into a ketone. Using transaminases in enzyme cascades represents a reliable pathway in which the ketone to be converted can be generated with the help of transketolases or dehydrogenases [29, 30, 31].

Another efficient variant of performing the epoxidation in a green, non‐toxic, biocatalytic fashion is the usage of unspecific peroxygenases (UPOs, EC 1.11.2.1) [32, 33, 34, 35, 36, 37, 38, 39]. These originally fungal haem‐thiolate enzymes only require hydrogen peroxide to achieve enantioselective oxyfunctionalisations, such as hydroxylations of sp³‐hydrogen‐carbon bonds or the epoxidation of alkenes, and produce water as the only by‐product [32]. The UPO from Myceliophthora thermophila (MthUPO) was chosen for this study as it enables a selective reaction with high conversions of 82% towards the desired styrene oxide (2) [32, 40]. The continuous addition of the oxidant H₂O₂ was carried out in a previously established synthetic cascade with the amino acid oxidase from Hebeloma cylindrosporum (hcLAAO4) [40, 41, 42, 43]. This cascade provides a constant supply of H₂O₂ within the aqueous reaction system at specifically low concentrations, which otherwise would lead to an undesired deactivation of the applied MthUPO due to a Haber–Weiss reaction [44].

2. Results and Discussion

A particular focus was therefore placed on the ring‐opening reaction step towards the desired arylethanolamines after the initial enzymatic reaction. Due to the ring tension, the opening of the oxirane structure can be facilitated by a variety of nucleophiles and electrophiles, e.g. aminolysis is frequently applied and was investigated in detail [45, 46, 47, 48, 49, 50]. Here, substituted epoxides can be converted into amino alcohols utilising primary and secondary amines as nucleophiles with metal ions of group 1 and 2 (alkaline or alkaline earth metal ions) as catalysts. Novel developments also include enzyme‐catalysed reactions with halohydrin dehalogenases, which have proven to be promising [7, 51, 52].

This study specifically focuses on the use of rare earth element salts, as they allow for milder reaction conditions and the unavoidable presence of residual water from the initial enzyme‐catalysed reaction [53]. This also allows for avoiding classical non‐environmentally friendly options such as tributylphosphine, boron trifluoride and similar, rather toxic compounds for the ring‐opening reaction [54, 55, 56]. Rare earth metal salts not only allow for a weakening of the oxirane ring structure by coordinating to the lone‐pair electrons of the included oxygen, but were also soluble in different solvents as their respective triflate salts [48, 57]. Please note that the ring‐opening reaction leads to isomeric structures, depending on the direction of the nucleophilic/electrophilic attack (Figure 1). The reaction can be performed in water as well as organic solvents after an extraction step, depending on how strong the nucleophiles or electrophiles used are compared to water. In any case, small amounts of water were necessary for the ring‐opening reaction, as it is needed as a proton source for the formation of the corresponding alcohols. The nucleophilicity strength of the ring‐opening reagents used proved to be decisive for both the choice of solvent, in which the reaction took place, and the use of additional auxiliaries such as further catalysts (e.g. lanthanide salts) or acids.

FIGURE 1.

Chemoenzymatic reaction scheme from styrene (1) to amino alcohols (4). Examples of vicinal amino alcohols in pharmaceuticals.

A compilation of all products obtained and isolated in this work, together with a general reaction scheme, is shown in Figure 2 below.

FIGURE 2.

Overview of isolated yields of the chemo‐enzymatical conversion of styrene (1) using MthUPO as biocatalyst for the oxyfunctionalisation reaction to styrene oxide (2) and subsequent use of different nucleophiles/electrophiles for the ring opening. The yields shown refer solely to the respective ring‐opening reaction.

Thiocyanate, as the strongest nucleophile tested, allowed for a reaction directly in the aqueous reaction system as a one‐pot cascade reaction. In the case of weaker nucleophiles, water as a solvent, which itself acts as a weak nucleophile, proved to be an obstacle as it competes with the nucleophile used. These reactions were performed in polar aprotic solvents like acetonitrile or methanol instead, after extracting the epoxide 2 from the enzyme solution, and adding very small amounts of water to the solution as a proton source, representing a simple, efficient two‐step synthesis approach.

The formation of 3a proved to be the most favoured reaction in this study. This can be explained by the fact that the thiocyanate ion can be regarded as a very strong nucleophile and therefore easily enables the ring opening [57]. This was also possible in aqueous solution in parallel to the UPO‐catalysed reaction, because thiocyanate is not harmful towards the biocatalytic reaction at low concentrations. Nonetheless, the reaction can also be performed in acetonitrile for 5 h at 85°C. The reaction was catalysed by lanthanoid ions, which weaken the bonds of the ring by coordinating to the lone‐pair electrons of the oxygen. The highest observed yield was 92%. Interestingly, the presence of metal‐containing porphyrins, such as the haem centre of the UPO, can catalyse the regioselective conversion of oxiranes to α‐hydroxy thiocyanates, further promoting the ring opening to be also possible next to the enzyme reaction [58]. In addition, the larger ionic radius of sulphur causes a steric hindrance, which leads to the formation of only the α‐alcohol isomer (anti‐Markovnikov) [59].

Using cyanate ions as a nucleophile led to the formation of two isomers of a heterocyclic species (3b and 3c). Due to the smaller atomic radius of oxygen, in comparison to sulphur as shown above, the formation of two regioisomers is possible by coordination of the nucleophile on both sides of the epoxide. Due to the resonance behaviour of cyanate, only the more stable isocyanate will undergo the nucleophilic attack, resulting in two intermediate alcohol isomers. The rearrangement of the hydrogen of the hydroxyl group to the lone pair of the nitrogen and the formation of a new C—O bond subsequently leads to the creation of the heterocycle. These oxazolidinones represent basically protected amino alcohols, which can be converted to unprotected amino alcohols, see below.

The reaction could not be performed in water alone, but had to be carried out in an aprotic solvent and required small amounts of acid for catalytic reasons, due to the weaker nucleophilicity of cyanate compared to thiocyanate. Unfortunately, this addition of acid prevents a simultaneous ring‐opening reaction as the UPO‐catalysed epoxidation is significantly deactivated at pH below 4. Thus, the ring opening was performed in a mixture of acetonitrile and water after an extraction of 2 from the enzymatic reaction solution, yielding 61% of 3b and 31% of 3c after 8h of refluxing at 80°C.

Product 3d was formed under similar conditions, as ammonia and amines function as nucleophiles of middle or low strength. In too alkaline pH environments, the ring‐opening reaction was not possible in parallel to the enzymatic reaction, as lanthanides form poorly soluble hydroxides in an alkaline environment and lose their catalytic properties in the process. Therefore, the reaction was performed in acetonitrile under reflux at 80°C for 24 h and yielded 87%, nonetheless. For steric reasons, the nucleophilic attack of the relatively bulky amine was hindered for the benzylic position of the epoxide 2, so that again only the anti‐Markovnikov isomer, 2‐(diethylamino)‐1‐phenylethanol (3d), was formed.

As mentioned above, the presence of water alone may also yield the ring opening, but this needs to be specifically catalysed by small amounts of acid, leading to the formation of 1‐phenylethan‐1,2‐diol (3e) as the expected product. The reaction, which was performed over 8 h at 80°C, led to a number of different products, including the expected diol (28%). Among the products are acetophenone, polymers of unknown chain length, which can only be composed of styrene oxide monomers, and 1,4‐diphenylbutane‐1,4‐diol, which makes up the majority of the products. As described above, the MthUPO cannot withstand low pH values without being deactivated.

In addition, iodomethane was also tested in the chemo‐enzymatic reaction towards the synthesis of amino alcohols. The epoxide (2) from the enzymatic reaction was dissolved in acetonitrile and reacted in a first step with equimolar amounts of the electrophile, and in a second step with 25% ammonia solution under reflux at 75°C for 6 h. This resulted in a mixture of two isomers of an amino ether (53% 3f and 35% 3g), which in turn could be further converted to the corresponding amino alcohols by acidic ether cleavage (shown below). Just like products 3b and 3c, these amino ethers represent protected amino alcohols. The reaction was performed in an aprotic solvent, as the choice of the latter influences the strength of the nucleophile or electrophile, respectively. For the mechanism shown in Figure 3 to occur, it was necessary to use acetonitrile [60].

FIGURE 3.

Reaction scheme for the conversion of styrene (2) with iodomethane in a first, and ammonia in a second step, resulting in the formation of 3f and 3g.

The relatively weak nucleophile ammonia itself at first seemed not to be useful as a ring‐opening reagent, as it would lead to the precipitation of the lanthanide catalyst in an aqueous environment and could quench the enzymatic reaction by raising the pH value, rendering the ring‐opening reaction next to the oxyfunctionalisation impossible. Nonetheless, the aminolysis of epoxides is a commonly known process [49]. For these reasons, another reaction pathway was chosen to avoid these difficulties. After the oxyfunctionalisation step, 2 was extracted from the aqueous enzyme‐catalysed solution and isolated from residues of 1, followed by a separate ring‐opening reaction. Using methanol as a solvent, the opening of the epoxide ring was possible by refluxing the solution at 80°C for 6 h with an excess of ammonia, yielding two isomers of amino alcohol (4a and 4b) in a nearly quantitative manner. Due to the capability of the relatively small ammonia molecule to perform the nucleophilic attack on both sides of the epoxide ring (compare Figure 1), two different products are inevitably formed, with 4a being more favoured than 4b. It can be assumed that the ring‐opening step follows primarily an S _N 2 mechanism, which is more likely in weak protic polar solvents such as methanol, used in that case, or aprotic polar solvents like acetonitrile. This includes a nucleophile attack from the back side of the epoxide in order to form the respective intermediate. Due to the significant steric hindrance of the adjacent aromatic ring, an attack at the benzylic position is more unlikely and thus the terminal position of the oxirane ring is favoured. Consequently, the formation of 4a is more likely and leads to the shown excess of 77% with an amino group at the terminal carbon atom.

Finally, the deprotection of the mentioned amino alcohol derivatives was also investigated. The oxazolidinones 3b and 3c could be easily converted to the amino alcohols 4a and 4b via decarboxylation by dissolving the protected species in cyclopentyl methyl ether and then adding concentrated hydrochloric acid. After refluxing the reaction solution for 8 h at 85°C, a mixture of both isomers of amino alcohols was obtained, conserving the same ratio of isomers as before. In the case of 3f and 3g, the de‐protection reaction was performed as an acidic ether cleavage, using hydrochloric acid in cyclopentyl methyl ether and refluxing for 3 h at 70°C, also yielding the same ratio of isomers. In both cases, the conversion of the protected amino alcohols to the free amino alcohols was quantitative, rendering it an effective synthesis strategy.

Overall, there are many possible pathways for the synthesis of substituted alcohols as well as amino alcohols, utilising the ring opening of epoxides, compare Figure 4 Depending on the nucleophilicity/electrophilicity strength of the respective ring‐opening reagent used, there is a broad spectrum of products which can be made accessible. In addition, nucleophilicity affects the achievable yields of the individual products and the reaction conditions to be used, which should be considered for possible applications.

FIGURE 4.

De‐protection reaction of the protected amino alcohols to 4a and 4b.

3. Conclusion

In summary, one of the biggest advantages of the shown synthesis route lies in the high atom efficiency. In the enzyme reaction, the only by‐product is water, which is not considered environmentally harmful, representing a major benefit of this reaction compared to conventional syntheses. For the ring opening, it can be noted that every nucleophile (anion) is completely included in the resulting product. Since water is produced in the enzymatic reaction in one equivalent and is consumed as a proton source during the ring‐opening reaction, it can be neglected in the overall calculation. Only in the case of the electrophilic ring opening with iodomethane does the atom efficiency decrease, as both the iodide and a proton are not contained in the product. Nevertheless, most of the atoms of the starting compounds are also found in the product in this example. In the direct synthesis of 4a and 4b, atom efficiency can even be assumed to be 100%, as the ammonia used is completely converted into the amino alcohols obtained. However, atom efficiency naturally decreases with regard to the conversion of the protected amino alcohols (3b, 3c, 3f, 3g), as the usage of additional chemicals is required.

The shown method for the synthesis of building blocks, such as amino alcohols, can be transferred to other substrates. The substitution of the reverse side of the oxirane ring could enable the orientation of functional groups in certain patterns as a result of the ring‐opening reaction without any need for spacious ring‐opening reagents. Bulky substituents favour the formation of products in which larger groups are arranged further away from the former. Of course, the ring‐opening reaction of an epoxide with two identical substituents on both sides of the oxirane ring would lead to only one type of product. In the case of chiral compounds, the formation of racemates would inevitably be observed.

Compared to standard methods such as cascades with transaminases or halohydrin dehalogenases, the chemo‐enzymatic variant presented in this paper displays a number of differences. Each cascade containing a transaminase requires a variety of different substrates and cofactors (e.g. PLP) in order to transfer the amino group from the donor to the main substrate [28, 29]. In comparison, the approach presented in this paper is significantly less complicated; only an unsaturated substrate and H₂O₂ are required for the production of the epoxide, which can then be further converted. Halohydrin dehalogenases offer the advantage that only a single enzyme is required to generate and convert the epoxide intermediate, as it catalyses both steps [24]. However, the dehalogenation is an equilibrium reaction, and its orientation towards the substrate or product side is the decisive factor for the efficiency of the overall conversion. In the case of the UPO‐catalysed oxyfunctionalisation reaction, the equilibrium is clearly on the product side, as the formation of the epoxide is favoured according to Le Chatelier's principle. The continuous supply of H₂O₂ promotes its consumption. In addition, the by‐product water is a stable compound which, under the given conditions, does not tend to decompose the epoxide in order to react back to H₂O₂ [40]. Ring opening or the breakdown of epoxides can, of course, also occur enzymatically, for example, through an epoxide hydrolase [61]. However, these hydrolases only produce diols and are therefore unusable for the synthesis of amino alcohols. The epoxidation of alkenes can also be achieved with enzymes other than UPO, such as monooxygenases or haloperoxidases [21, 22]. However, similar to transaminases, these enzymes require cofactors and a more complex experimental setup, which is a minor disadvantage compared to the UPO‐catalysed approach.

It was possible to demonstrate the compatibility of a UPO‐catalysed oxyfunctionalisation reaction with a chemical ring‐opening reaction in a simple two‐step process with high atom efficiency. The resulting substituted alcohols can be used as precursors for a large number of further reactions and syntheses. The heterocycles 3b and 3c, which can be converted to amino alcohols as well, are also of interest as these oxazolidones are chiral auxiliaries, which are useful compounds for selective aldol reactions, alkylations or Diel–Alder reactions [62]. Furthermore, oxazolidones are effective antibiotics, making their synthesis interesting for a variety of applications in the pharmaceutical industry [63]. The further conversion of protected alcohols opens up possible pathways for the production of a multitude of compounds. Although the stereoselective production of the compounds shown or a one‐pot approach was not the goal of this work, the demonstration of the chemo‐enzymatic synthesis was very successful. Also, not only do the protected amino alcohols represent interesting precursors. The same applies to 3a, which can be used for the preparation of synthetically valuable tetrazoles, using azides for the conversion. Considering the wide range of possibilities, the value of the methods shown here can confidently be rated as high.

Furthermore, the amino alcohols can, of course, be substituted with various groups to either directly produce pharmaceutically active compounds, like isoprenaline or propranolol, or to obtain starting materials for conversion into more complex substances. A commonly known issue regarding unspecific peroxygenases is their inability to convert amines to a satisfactory extent. The approach of this work presents a good alternative, opening up new possibilities for the synthesis of this highly desired group of compounds.

4. Experimental Section

4.1. General Information

All chemicals as well as solvents used for cell expression, synthesis, processing and analytical purposes were purchased from commercial distributors of analytical grade, Merck (Darmstadt, Germany), Thermo Fisher Scientific (Waltham, MA, USA) and Carl Roth (Karlsruhe, Germany). In addition, SC Drop‐Out for the expression media for the production of yeast was purchased from Formedium Ltd (Kings Lynn, England).

4.2. Gas Chromatography (GC)

For gas chromatography measurements, a Nexis GC–2030 system from Shimadzu Europa GmbH (Duisburg, Germany) was used. The device is equipped with an AOC‐20i Plus Auto Injector and a SH‐5 column (length 30 m, ID 0.25, DF 0.25, temperature range −60°C to 330/350°C, P/N 221‐75 701−30). The stationary phase of the column is composed of 5% diphenyl polysiloxane and 95% dimethyl polysiloxane. Hydrogen is used as the main carrier gas with a flow rate of 36.0 ml/min with nitrogen as make‐up gas (24.0 ml/min) and compressed air (200.0 ml/min) for the flame ionisation detector. The temperature programme was driven from 100°C to 200°C with a heating rate of 10°C/min, followed by a holding step over 2 min and a second heating step up to 250°C with a heating rate of 25°C/min over a total time frame of sixteen minutes. Samples of enzymatic reaction products were extracted from the aqueous reaction medium using dichloromethane. The completeness of the performed ring‐opening reactions as well as the purity of the isolated products were also observed via GC, using dichloromethane as solvent. The relatively inert n‐decane was used as an internal standard, being added to the dichloromethane used for extraction beforehand.

4.3. Nuclear Magnetic Resonance (NMR)

¹H and ¹³C spectra were measured on an AV III Bruker‐BioSpin (¹H: 400.13 MHz; ¹³C: 100.62 MHz) or an Avance Neo Bruker BioSpin (¹H: 600.13 MHz; ¹³C: 150.9 MHz). The conversion and purity of all substrates were observed using chromatographic methods. The solutions were prepared respectively in CDCl₃ and DMSO‐d₆, and ¹H and ¹³C shifts were referenced to internal solvent resonances and reported in parts per million relative to TMS. The NMR data are contained in the Supporting Information.

4.4. Photometer (TECAN Plate Reader)

Photometric measurements were performed on a Spark Multimode Microplate Reader (Tecan Trading AG, Switzerland). A xenon flash lamp (spectral range from 200 to 1000 nm, optical density range from 0 to 4 OD, scan speed ≤5 s, wavelength accuracy <0.3 nm) was used for absorbance measurements. The measurements were performed in Microplates (96‐well, PS, F‐Bottom, clear) purchased from Greiner Bio‐One.

4.5. NBD Assay

The activity of the unspecific peroxygenase was measured using 5‐nitro‐1,3‐benzodioxole (NBD) as substrate, which is converted to 4‐Nitrocatechol that can be measured at a wavelength of 425 nm. The reactions were performed in the above mentioned well plates in a total volume of 200 µl per well, containing 0.5 mM NBD (stock solution of 10 mM, dissolved in acetonitrile). 10 µl of concentrated UPO solution (1.44 U/ml equals 0.072 U/ml in the final assay volume) and 10 µl hcLAAO4 or H₂O₂ solution, respectively, as well as 100 mM HEPES buffer (pH 7.0) were contained in every assay reaction. In case the LAAO was used, the same concentration of L‐alanine was added to the solution in order to supply the oxidase for the production of H₂O₂. Every well was filled up with ultrapure water in order to match a total volume of 200 µl. The measurements were performed over the duration of 5 minutes. The slope of each measurement was then used to determine the activity of the UPO under the given circumstances. All reactions were performed in triplicate.

4.6. H₂O₂ Assay

The activity of the LAAO used had also been determined, as the supply of H₂O₂ was decisive for the performance and life span of the unspecific peroxygenase. That goal was achieved using two kinds of photometric assay. First, there was a molybdate assay, utilising ammonium molybdate and potassium hydrogen phthalate to determine the concentration of H₂O₂ of a given solution, allowing conclusions to be drawn about the production of H₂O₂ using LAAO and thus about the activity of the enzyme. Measurements took place at a wavelength of 351 nm with a total volume of 200 µl per well. 60 µl colour solution (0.05 M NaOH, 0.4 M KI, 10⁻⁴ M ammonium molybdate) and 60 µl potassium hydrogen phthalate solution (0.5 M) were mixed. Adding 80 µl of H₂O₂ containing the sample solution (10–100 mM) started the reaction. The calibration took place using precise H₂O₂ standard solutions. The data gathered from these experiments was used to select the correct ratio of UPO and LAAO for the oxyfunctionalisation reaction.

The second option for the determination of the LAAO activity was the usage of a POX/o‐Dianisidin assay. The total volume of the assay solution per well contained 1.25 µl (0.2 mg/ml) o‐Dianisidin (stock solution of 32 mg/ml, solved in DMSO), 10 µl L‐alanine (100 mM in ultrapure water), 1.25 µl peroxidase solution (POX, 400 U/ml) and 10 µl sample and was filled up with TEA buffer (50 mM, pH 7.0). Measurements took place at a wavelength of 436 mM for twenty minutes at 30°C and allowed statements over the activity of the LAAO.

4.7. Performed Conversions

The completion of each reaction was tested using gas chromatography at several points in time. Each of the reactions listed was performed several times, and the times specified, for example, were checked and confirmed more than once. At first, the goal was to achieve every conversion in a one‐pot approach, but it became clear that in most cases, this was not possible for reasons mentioned below. We decided to aim for the highest possible atom efficiency and a proof of concept.

4.8. Styrene Oxide (2)

The epoxidation of styrene to styrene oxide was carried out in the same way as described in our previous publication [40]. The produced styrene oxide was extracted from the aqueous enzyme reaction solution using dichloromethane. For the separation of styrene and styrene oxide, column chromatography with silica gel as the stationary phase was used. The dichloromethane was carefully evaporated at low temperatures afterwards. 2 was used as a precursor for every other product in this work. As an epoxide, it possesses a certain ring tension, which leads to a reactivity that can be used to open the three‐membered oxirane ring with nucleophiles as well as electrophiles. The initial biocatalytic oxyfunctionalisation of styrene requires an aqueous reaction medium, containing 10% (v/v) acetonitrile to enhance reactant solubility. With the optimised reaction conditions of 200 µl MthUPO (1.44 U/ml) at 30°C, using 0.1 M HEPES buffer (pH 7.0) and 20 mM styrene, the most complete conversion to the epoxide was achieved.

4.9. 1‐phenyl‐2‐thiocyanatoethanol (3a)

The reaction, which can be performed in water as well as organic solvents, was performed in acetonitrile, in which equimolar amounts of styrene oxide (1.14 ml, 10 mmol) and NH₄SCN (0.76 g, 10 mmol) in ultrapure water (10% of the total volume) were mixed. 5 mol% (0.31 g) of Ytterbium(III)‐trifluormethansulfonate (Ytterbiumtriflate, Yb(OTf)₃) were added to the solution, which was then refluxed at 85°C and constantly stirred for 5 h. The at first colourless solution turned yellow. After finishing the reaction, what was observed using gas chromatography, the solution was filtered and dried in an evaporator. The yellow crystals were washed twice with water and dichloromethane and dried again. The reaction yielded 1.65 g (92.3%). Purity was checked using NMR as well as gas chromatography. The reaction was also successfully tested on a scale of up to 100 mmol/l.

In contrast to the ring‐opening reactions with other nucleophiles, only a single product was found. This substituted alcohol seems to be formed in exactly this pattern for reasons of steric hindrance. Sulphur has a larger atomic radius than oxygen or carbon, resulting in the formation of a product in which the aromatic ring and the thiocyanate group are arranged in such a way that they are as far apart from each other as possible.

4.10. 5‐phenyloxazolidin‐2‐one (3b) and 4‐phenyloxazolidin‐2‐one (3c)

As an analogue to thiocyanate, the smaller cyanate ion was also tested for the ring‐opening reaction. It quickly became clear that a reaction in water is not possible, as this itself acts as a weak nucleophile and competes with the nucleophile actually intended. Cyanate has a weaker nucleophilicity and requires, therefore, a different treatment than thiocyanate before.

The reaction was performed in 27 ml acetonitrile with 3 ml (10% (v/v)) of ultrapure water. Because of the weaker nucleophilicity of the cyanate, the addition of small amounts of acids was required. For this purpose, 0.3 ml (1% (v/v)) of concentrated hydrochloric acid (12 mol/l) was added to the solution using a syringe pump over the course of the reaction, in which equimolar amounts of 2 (1.14 ml, 10 mmol) and NaOCN (0.65 g, 10 mmol) were solved and stirred under reflux at 80°C for 8 h. Yb(OTf)₃ (0.31 g, 5 mol%) was used as a catalyst. After finishing the reaction, the reaction medium was dried, and the colourless solid was then washed twice with water and dichloromethane to get rid of unwanted salts. The reaction yielded 1.50 g (9.17 mmol, 91.7%). Therefore, 0.1 g (6.12 mmol, 61.2%) was 3b and 0.50 g (3.06 mmol, 30.6%) was 3c, suggesting that the former was preferred to be built in that reaction. The isomers were separated using column chromatography with dichloromethane/isopropyl alcohol (1:1) as solvent and silica gel as the stationary phase.

In regards to the used nucleophile, the addition of the acid was necessary for two reasons. For once, in order to make up for the lower nucleophilicity, the acid helps open the ring by further weakening the bond stability of the oxirane. The second reason for adding the acid was to convert the stable NaOCN to the more reactive cyanic acid. Noteworthy, in terms of thermodynamic stability, cyanic acid is less stable than isocyanic acid. For this reason, isocyanate ions are to be expected in solution. The charge will be located at the nitrogen site of the ion, influencing the kind of interaction with the epoxide. The nucleophilic attack on the epoxide can occur from two sites, resulting in different isomers. The rearrangement of the proton of the hydroxyl group to the lone pair of the nitrogen and the formation of a new C—O bond result in a five‐membered heterocycle (compare Figure 5).

FIGURE 5.

Reaction scheme for the formation of 3b and 3c.

As the oxazolidinones represent a kind of protected amino alcohol, they could be converted to the latter by means of acidic hydrolysis. For this purpose, 3b and 3c were each solved in cyclopentyl methyl ether, to which hydrochloric acid was added. Through refluxing at 85°C for 6 h a decarboxylation was achieved, and the corresponding amino alcohols (4a and 4b, compare below) were obtained. In both cases, the reaction was quantitative.

4.11. 2‐(diethylamino)‐1‐phenylethanol (3d)

The opening reaction using diethylamine had to be performed without any lanthanoid as a catalyst. The reason therefore, is that lanthanoids form very poorly soluble hydroxides in basic environments, which cancels out their catalytic properties. After the extraction, 2 was solved in acetonitrile. A successful reaction in an aqueous solution was not possible because of the relatively low nucleophilicity of the amine, which would interfere with the weak nucleophilic acting water. Equimolar amounts of 2 (10 ml, 87.4 mmol) and diethylamine (9.00 ml, 87.4 mmol) were brought together and refluxed under constant stirring at 80°C for a day. After that, the solution had changed from colourless to dark yellow. The acetonitrile was evaporated, and the yellow solid was washed with dichloromethane and water. Both phases were evaporated separately, and the product, which was contained in the dichloromethane phase, was isolated and analysed. The yield of 3d was 14.71 g (76.13 mmol, 87.1%), and only one isomer was found.

4.12. 1‐phenylethane‐1,2‐diol (3e)

The reaction of 2 with just water under the influence of acid was also tested. However, it turned out that the expected product 3d was only obtained with a yield of 0.38 g (2.75 mmol, 27.5%). The attempt can therefore be considered a failure. Nonetheless, the reaction was performed both in aqueous solution (30 ml) as well as in acetonitrile (30 ml), adding equimolar amounts of the epoxide (1.14 ml, 10 mmol) and 12 mol/l hydrochloric acid (0.83 ml, 10 mmol) to the reaction medium with 0.12 g (2 mol%) of Yb(OTf)₃ as catalyst. Leftovers of styrene oxide were found in the product mixture. Also, acetophenone and certain polymers, which can only be attributed to the polymerisation of the epoxide, were detected. Lastly, a definite main product of the reaction could not be isolated and named.

4.13. 2‐methoxy‐2‐phenylethanamine (3f) and 2‐methoxy‐1‐phenylethanamine (3g)

In contrast to the other reactions of this work, this conversion was performed with an electrophile. Iodomethane is known to be a strong electrophile, capable of not only iodising a variety of hydrocarbons and interacting with aliphatic compounds and aromatics as well as any kind of electron‐rich groups. For the reaction, 2 (1.14 ml, 10 mmol) was dissolved in 20 ml acetonitrile. In a first step, iodomethane (0.62 ml, 10 ml) in acetonitrile was added dropwise using a syringe. Secondly, 10 ml of ammonia solution (25%, 13.30 mol/l) was carefully added to the reaction medium. The mixture was stirred and refluxed at 75°C for 6 h; the solution turned yellow. After filtering and evaporating the solution, the yellow solid was washed with dichloromethane and water. A total yield of 1.33 g (8.78 mmol, 87.8%) could be isolated. 0.80 g (5.29 mmol, 52.9%) were 3f and 0.53 g (3.51 mmol, 35.1%) were 3g. The isomers were separated using column chromatography with dichloromethane/isopropyl alcohol (1:1) as solvent and silica gel as the stationary phase.

The amino ethers can be converted to the corresponding amino alcohols by means of acidic hydrolyses with quantitative yields.

4.14. 2‐amino‐1‐phenylethanol (4a) and 2‐amino‐2‐phenylethanol (4b)

Ammonia acts as a relatively weak nucleophile. For this reason, it was necessary to perform the ring‐opening reaction for the direct synthesis of 4a and 4b in a polar‐aprotic solvent different from water. Methanol was chosen for this purpose. 2 ml (15.9 mmol) of 2 was dissolved in 20 ml of methanol and then mixed with 20 ml of 25% ammonia solution. The use of lanthanoids as catalysts was not possible in this alkaline environment. The reaction solution was refluxed at 80°C under constant stirring for 6 h. Afterwards, the light‐yellow solution was evaporated, and the solid was washed with hydrochloric acid and ethyl acetate and then neutralised with sodium hydroxide in toluol. Finally, the products were isolated with a total yield of 2.16 g (15.7 mmol, 98.9%), from which 1.69 g (12.3 mmol, 77.2%) were 4a and 0.47 g (3.45 mmol, 21.7%) were 4b. The isomers were carefully separated using column chromatography with dichloromethane/isopropyl alcohol (1:1) as solvent and silica gel as the stationary phase. There were no leftovers of styrene oxide to be detected, but the gas chromatography showed small amounts of polymers with high boiling points. These polymers are most likely composed of styrene oxide monomers.

Supporting Information

Additional supporting information can be found online in the Supporting Information section. Supporting Fig. S4: Gas chromatogram of 3a after the extraction with dichloromethane. Supporting Fig. S5: Gas chromatogram of 3b and 3c after the extraction with dichloromethane. Supporting Fig. S6: Gas chromatogram of 3d after the extraction with dichloromethane. Supporting Fig. S7 : Gas chromatogram of the product mixture of the ring‐opening of 2 with hydrochloric acid and water. Supporting Fig. S8: Gas chromatogram of the product mixture of 3f and 3g. Supporting Fig. S9 : Gas chromatogram of the product mixture of 3f and 3g using a longer temperature programme. Supporting Fig. S10 :Gas chromatogram of 4a after processing. Supporting Fig. S11 : Gas chromatogram of 4b after processing.

Funding

This work was supported by the Deutsche Forschungsgemeinschaft (505185500, 450014604 and 505185500).

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supplementary Material

Data Availability Statement

The data that support the findings of this study are available in the supplementary material of this article.