Lipase Catalysed Kinetic Resolution of Racemic 1,2-Diols Containing a Chiral Quaternary Center

Optically active 1,2-diols are valuable buildings blocks in organic synthesis. In the present paper, a set of racemic 1,2-diols with an ester functional group are prepared, starting from α-ketoesters in a three-step procedure with moderate yields. The racemic 1,2-diols, containing a chiral quaternary center in their structure, are subjected to selective acylation in order to perform their kinetic resolution catalysed by a set of commercially available lipases. Under optimized reaction conditions, good conversions and enantioselectivities are achieved by using the lipase PSL-C from Pseudomonas cepacia in tert-butyl methyl ether. This biocatalyst could be reused up to five times without losing its properties.


Introduction
Optically active 1,2-diols are valuable compounds as they can be transformed into several interesting molecules [1][2][3].Due to their importance as building blocks in organic synthesis, a number of synthetic methodologies have been developed for their preparation [4,5].One of the most widely applied methods is Sharpless asymmetric dihydroxylation, which involves the oxidation of alkenes to form 1,2-diols in presence of chiral catalysts [6,7].This methodology suffers some limitations including the relatively low activity and selectivity of the aromatic compounds and the substrate scope.For these reasons, different alternatives have been investigated.The use of biologically active systems as catalysts in organic reactions, including whole cells, cells free extracts, or purified enzymes, has emerged as a powerful tool for the preparation of high added value compounds under mild and environmentally friendly conditions [8][9][10][11].
The enzymes that have been used to catalyse the formation of chiral 1,2-diols belong to the oxidoreductases (alcohol dehydrogenases and dioxygenases) [12,13], and the hydrolases, including epoxide hydrolases [14,15] and lipases.This last group of enzymes (EC 3.1.1.3)have demonstrated their synthetic applicability, being the most used type of enzymes in industrial chemistry [16][17][18].Lipases are widely available, have broad substrate acceptance and are able to catalyse reactions not only in aqueous mediums, but also in organic solvents, which expands their synthetic repertoire.In addition, lipases usually display a high degree of chemo-, regio-and/or enantioselectivity in the processes that they catalyse.
Lipases have been used to prepare optically active 1,2-diols catalysing the kinetic resolution of racemic mixtures in acylation reactions, thus leading to chiral diols and esters that can be converted back to the starting diols via hydrolysis [19][20][21][22].In 2000, the kinetic resolution of racemic ethyl 2-benzyl-2,3-dihydroxypropanoate was described, a 1,2-diol containing an ester moiety precursor in the synthesis of (S)-etomoxir, which is a powerful hypoglucemic reagent [23].After testing different biocatalysts, the lipase PS from Amano was found to catalyse the formation of (S)-1,2-diol and (R)-acetate with high selectivity and activity.Given this result, we decided to synthesise a set of functionalised 1,2-diols with an ester group and perform their kinetic resolution in the presence of different lipases, with the aim of obtaining these valuable optically active compounds.

Results and Discussion
2.1.Preparation of the Racemic 1,2-Diols (±)-1-6d The racemic 1,2-diols were prepared in a three-step procedure starting from the corresponding α-ketoesters 1-6a, as indicated in Table 1.These compounds were treated with N-tert-butyl formaldehyde hydrazone in toluene at room temperature to yield the racemic azocompounds (±)-1-6b with high yields (75-92%) after 24 h.For almost all the ketoesters, the reaction occurred in the absence of any catalyst, but for ethyl benzoylformate (1a), the addition was accelerated in the presence of the Schreiner's thiourea (I) at 10 mol%.The resulting azocompounds are valuable synthons that can be transformed in different compounds [24][25][26].Their hydrolysis in a biphasic system Et 2 O/HCl (aq) yielded the corresponding hydroxyaldehydes (±)-1-6c after four hours via a tautomerization and hydrolysis process.Debt to their instability were directly reduced without purification to the racemic 1,2-diols (±)-1-6d by treatment with a mild reductant as tetrabutylammonium borohydride (NBu 4 BH 4 ) in CH 2 Cl 2 at room temperature over two hours.Depending on the substrate structure, the 1,2-diols were obtained with yields from 41 to 57%.Attempts to improve these yields by modifying certain reaction parameters as the hydrolysis conditions or the reducing agent were unsuccessful.
Molecules 2018, 23, x FOR PEER REVIEW 2 of 10 functionalised 1,2-diols with an ester group and perform their kinetic resolution in the presence of different lipases, with the aim of obtaining these valuable optically active compounds.

Preparation of the Racemic 1,2-Diols (±)-1-6d
The racemic 1,2-diols were prepared in a three-step procedure starting from the corresponding α-ketoesters 1-6a, as indicated in Table 1.These compounds were treated with N-tert-butyl formaldehyde hydrazone in toluene at room temperature to yield the racemic azocompounds (±)-1-6b with high yields (75-92%) after 24 h.For almost all the ketoesters, the reaction occurred in the absence of any catalyst, but for ethyl benzoylformate (1a), the addition was accelerated in the presence of the Schreiner's thiourea (I) at 10 mol%.The resulting azocompounds are valuable synthons that can be transformed in different compounds [24][25][26].Their hydrolysis in a biphasic system Et2O/HCl (aq) yielded the corresponding hydroxyaldehydes (±)-1-6c after four hours via a tautomerization and hydrolysis process.Debt to their instability were directly reduced without purification to the racemic 1,2-diols (±)-1-6d by treatment with a mild reductant as tetrabutylammonium borohydride (NBu4BH4) in CH2Cl2 at room temperature over two hours.Depending on the substrate structure, the 1,2-diols were obtained with yields from 41 to 57%.Attempts to improve these yields by modifying certain reaction parameters as the hydrolysis conditions or the reducing agent were unsuccessful.

Kinetic Resolution of Racemic Diols (±)-1-6d
Once the racemic 1,2-diols were synthesized, their kinetic resolution was performed.Our initial studies were performed using racemic ethyl 2,3-dihydroxy-2-phenylpropanoate (±)-1d as the model substrate.The selective acetylation of this 1,2-diol (0.15 M) in toluene at 30 °C was performed in presence of 3.0 equivalents of vinyl acetate to ensure an irreversible acylation process.The effects of

Kinetic Resolution of Racemic Diols (±)-1-6d
Once the racemic 1,2-diols were synthesized, their kinetic resolution was performed.Our initial studies were performed using racemic ethyl 2,3-dihydroxy-2-phenylpropanoate (±)-1d as the model substrate.The selective acetylation of this 1,2-diol (0.15 M) in toluene at 30 • C was performed in presence of 3.0 equivalents of vinyl acetate to ensure an irreversible acylation process.The effects of different biocatalysts were analysed by screening several commercially available lipases, as indicated in Table 2.For all the biocatalysts tested, the (R)-enantiomer of the 1,2-diol was acetylated, yielding (R)-1e, whereas the (S)-1,2-diol remained unaltered.The use of immobilized Candida antarctica lipase B (CalB) resulted in a reaction without selectivity (entry 1), achieving a 23% conversion after four hours.When the reaction was catalysed by its isozyme A (entry 2), a more selective process was observed (enantioselectivity, E = 12), demonstrating acylation slower than with CalB.As shown in entry 3, Pseudomonas cepacia lipase (PSL-C) seems to be the most suitable biocatalyst for this reaction, as a moderate E value was obtained in a process with a 32% conversion after 20 h, achieving (R)-1d with 86% ee.The use of lipases from Pseudomonas fluorescens (PSF), Burkholderia sp.(BSL), Rhizopus oryzae, and Aspergillus oryzae led to very low enantioselectivities (E < 10, entries 4-7), with conversions varying from 41% after 20 h with PSF, to 16% with the same duration using the lipase from Aspergillus oryzae.The kinetic resolution catalysed by porcine pancreatic lipase (PPL) in toluene afforded a selectivity value of 13, and a conversion close to 50% after 24 h, as shown in entry 8.The opposite lipase from Mucor miehei was not a suitable biocatalyst for this reaction, as only a 13% of (R)-1e was obtained after 24 h in a very low selective resolution (entry 9).different biocatalysts were analysed by screening several commercially available lipases, as indicated in Table 2.For all the biocatalysts tested, the (R)-enantiomer of the 1,2-diol was acetylated, yielding (R)-1e, whereas the (S)-1,2-diol remained unaltered.The use of immobilized Candida antarctica lipase B (CalB) resulted in a reaction without selectivity (entry 1), achieving a 23% conversion after four hours.When the reaction was catalysed by its isozyme A (entry 2), a more selective process was observed (enantioselectivity, E = 12), demonstrating acylation slower than with CalB.As shown in entry 3, Pseudomonas cepacia lipase (PSL-C) seems to be the most suitable biocatalyst for this reaction, as a moderate E value was obtained in a process with a 32% conversion after 20 h, achieving (S)-1d with 86% ee.The use of lipases from Pseudomonas fluorescens (PSF), Burkholderia sp.(BSL), Rhizopus oryzae, and Aspergillus oryzae led to very low enantioselectivities (E < 10, entries 4-7), with conversions varying from 41% after 20 h with PSF, to 16% with the same duration using the lipase from Aspergillus oryzae.The kinetic resolution catalysed by porcine pancreatic lipase (PPL) in toluene afforded a selectivity value of 13, and a conversion close to 50% after 24 h, as shown in entry 8.The opposite lipase from Mucor miehei was not a suitable biocatalyst for this reaction, as only a 13% of (R)-1e was obtained after 24 h in a very low selective resolution (entry 9).
]. e Reaction performed with isopropenyl acetate as acyl donor.f Reaction performed with ethyl acetate as acyl donor.
After selecting PSL-C as the best biocatalyst for the acetylation of racemic 1d, we analysed other parameters that can affect the activity and selectivity of the biocatalyst.Thus, different organic solvents were tested in the acylation reaction.As shown in entry 10, tert-butyl methyl ether (TBME) was the best solvent for this process, as a good selectivity value could be achieved (E = 41) in a reaction much faster than in toluene (34% conversion after eight hours with 91% ee for 1e).This solvent was also tested in the CalA-catalysed acylation, promoting an increase in both the enzyme activity and

Entry
Lipase Solvent T ( ]. e Reaction performed with isopropenyl acetate as acyl donor.f Reaction performed with ethyl acetate as acyl donor. After selecting PSL-C as the best biocatalyst for the acetylation of racemic 1d, we analysed other parameters that can affect the activity and selectivity of the biocatalyst.Thus, different organic solvents were tested in the acylation reaction.As shown in entry 10, tert-butyl methyl ether (TBME) was the best solvent for this process, as a good selectivity value could be achieved (E = 41) in a reaction much faster than in toluene (34% conversion after eight hours with 91% ee for 1e).This solvent was also tested in the CalA-catalysed acylation, promoting an increase in both the enzyme activity and selectivity (entry 11).However, the E value remained low.The reaction catalysed by PPL in TBME afforded the chiral acetate (R)-1e with 77% ee in a process with a 38% conversion after 12 h (entry 12).Other solvents analysed in the acetylation catalysed by PSL-C, such as 1,4-dioxane or THF, led to slower resolutions and especially for the latter, whereas the enantioselectivities were around 15.The use of diisopropyl ether (DIPE) afforded (S)-1d with 67% ee and (R)-1e with 83% ee in a process with a 45% conversion after 12 h (entry 15).
The effect of temperature was also analysed in this kinetic resolution, performing the PSL-C-catalysed reaction of (±)-1d at 15 • C, as shown in entry 16.Lowering the temperature had no effect on the enzyme selectivity, whereas, as expected, the enzyme activity dropped (c = 25% after 12 h).The use of a less reactive acyl donor, such as isopropenyl acetate (entry 17), led to a similar selectivity.Chiral acetate (R)-1e was obtained with a 36% conversion and 91% ee after 16 h.When ethyl acetate was used as acyl donor (entry 17), a slower kinetic resolution was achieved, as 48 h were required to obtain a 34% yield of (R)-1e with 91% ee (entry 18).
The recycling of the PSL-C was studied in the selective acetylation of (±)-1d with vinyl acetate in TBME at 30 • C.After 20 h, the biocatalyst was filtered, washed with TBME, and used again in a new reaction cycle.As shown in Figure 1, this biocatalyst could perform the selective acylation of the racemic diol for five cycles while maintaining its activity and selectivity.For the sixth reaction, a significant drop in the enantioselectivity of the process was observed (E = 27).For the seventh reaction, this drop was accompanied by an appreciable loss in enzymatic conversion.
Molecules 2018, 23, x FOR PEER REVIEW 4 of 10 solvent was also tested in the CalA-catalysed acylation, promoting an increase in both the enzyme activity and selectivity (entry 11).However, the E value remained low.The reaction catalysed by PPL in TBME afforded the chiral acetate (S)-1e with 77% ee in a process with a 38% conversion after 12 h (entry 12).Other solvents analysed in the acetylation catalysed by PSL-C, such as 1,4-dioxane or THF, led to slower resolutions and especially for the latter, whereas the enantioselectivities were around 15.The use of diisopropyl ether (DIPE) afforded (R)-1d with 67% ee and (S)-1e with 83% ee in a process with a 45% conversion after 12 h (entry 15).The effect of temperature was also analysed in this kinetic resolution, performing the PSL-C-catalysed reaction of (±)-1d at 15 °C, as shown in entry 16.Lowering the temperature had no effect on the enzyme selectivity, whereas, as expected, the enzyme activity dropped (c = 25% after 12 h).The use of a less reactive acyl donor, such as isopropenyl acetate (entry 17), led to a similar selectivity.Chiral acetate (S)-1e was obtained with a 36% conversion and 91% ee after 16 h.When ethyl acetate was used as acyl donor (entry 17), a slower kinetic resolution was achieved, as 48 h were required to obtain a 34% yield of (S)-1e with 91% ee (entry 18).
The recycling of the PSL-C was studied in the selective acetylation of (±)-1e with vinyl acetate in TBME at 30 °C.After 20 h, the biocatalyst was filtered, washed with TBME, and used again in a new reaction cycle.As shown in Figure 1, this biocatalyst could perform the selective acylation of the racemic diol for five cycles while maintaining its activity and selectivity.For the sixth reaction, a significant drop in the enantioselectivity of the process was observed (E = 27).For the seventh reaction, this drop was accompanied by an appreciable loss in enzymatic conversion.After the optimized conditions were set up for the kinetic resolution of racemic ethyl 2,3-dihydroxy-2-phenylpropanoate, using PSL-C and CalA in TBME at 30 °C and vinyl acetate as acyl donor, the scope of the reaction using different aromatic 1,2-diols was studied (Table 3).For all the substrates, (S)-acetates 2-6e were the obtained products.The use of PSL-C led to higher enatioselectivities in all the aromatic and the heteroaromatic substrates (2-5a), whereas CalA showed higher activity.Thus, the enzymatic acylation of racemic methyl 2,3-dihydroxy-2-phenylpropanoate (2d) catalysed by PSL-C occurred with the same selectivity as for the ethyl analogue 1d (E = 42, entry 1) and with a higher conversion, achieving a 41% conversion after 12 h.When the acylation was catalysed by CalA (entry 2) (S)-2e was obtained with 80% ee and a 42% conversion after eight hours (E = 16).The p-cyano derivative (±)-3c was a good substrate for both catalysts (entries 3 and 4).When using PSL-C, a 44% of (S)-3e with 90% ee was obtained after 12 h, in a resolution with a good enantioselectivity value (E = 40), whereas the resolution catalysed by CalA led to a 47% conversion after eight hours and a good selectivity (E = 30).The presence of an After the optimized conditions were set up for the kinetic resolution of racemic ethyl 2,3-dihydroxy-2-phenylpropanoate, using PSL-C and CalA in TBME at 30 • C and vinyl acetate as acyl donor, the scope of the reaction using different aromatic 1,2-diols was studied (Table 3).For all the aromatic substrates, (R)-acetates 2-5e were the obtained products.The use of PSL-C led to higher enatioselectivities in all the aromatic and the heteroaromatic substrates (2-5a), whereas CalA showed higher activity.Thus, the enzymatic acylation of racemic methyl 2,3-dihydroxy-2-phenylpropanoate (2d) catalysed by PSL-C occurred with the same selectivity as for the ethyl analogue 1d (E = 42, entry 1) and with a higher conversion, achieving a 41% conversion after 12 h.When the acylation was catalysed by CalA (entry 2) (R)-2e was obtained with 80% ee and a 42% conversion after eight hours (E = 16).The p-cyano derivative (±)-3c was a good substrate for both catalysts (entries 3 and 4).When using PSL-C, a 44% of (R)-3e with 90% ee was obtained after 12 h, in a resolution with a good enantioselectivity value (E = 40), whereas the resolution catalysed by CalA led to a 47% conversion after eight hours and a good selectivity (E = 30).The presence of an electron-donating group in the aromatic ring of the diol seemed to have a negative effect on both the activity (41% of (R)-4e after Molecules 2018, 23, 1585 5 of 10 24 h) and the selectivity (E = 32) of PSL-C (entry 3).This substrate was tested with CalA, but a low selectivity was observed (E = 11), in a kinetic resolution with a 36% conversion after 16 h, as shown in entry 6.A heteroaromatic 1,2-diol as (±)-5d was successfully resolved by PSL-C in MTBE.After 24 h, a 45% of (R)-5e with 88% ee was achieved in a process with good selectivity, as shown in entry 7. The use of CalA led to a 38% conversion after 16 h and moderate selectivity (E = 17, entry 8).h) and the selectivity (E = 32) of PSL-C (entry 3).This substrate was tested with CalA, but a low selectivity was observed (E = 11), in a kinetic resolution with a 36% conversion after 16 h, as shown in entry 6.A heteroaromatic 1,2-diol as (±)-5d was successfully resolved by PSL-C in MTBE.After 24 h, a 45% of (R)-5e with 88% ee was achieved in a process with good selectivity, as shown in entry 7.
The use of CalA led to a 38% conversion after 16 h and moderate selectivity (E = 17, entry 8).Regarding the aliphatic diol (±)-6d, in which the stereogenic center presents an aliphatic substituent, the enzymatic acylation in TBME at 30 °C using PSL-C afforded a very low enantioselectivity value (E = 7) in a very fast resolution, achieving a 45% of (+)-6e after 2 h.In view of this result, the isozyme A from Candida antarctica was tested, leading to a slower (c = 17% after 6 h), but slightly more selective process (E = 12) than with PSL-C (see entry 10).In order to improve the reaction selectivity, the PSL-C-catalysed resolutions were carried out using ethyl acetate as a less reactive acyl donor.After 6 h, a 15% of (+)-6e was obtained in a process with a low enantioselectivity (E = 10).Finally, the kinetic resolution in presence of vinyl acetate was conducted at 10 °C.After 4 h, a 17% of diol 6d was converted into the acetate (+)-6e with 86% ee, but the E value was only increased to 16 (entry 12), indicating that this substrate was not appropriate for the biocatalysed acylation.

Materials and Methods
Unless otherwise noted, analytical grade solvents and commercially available reagents were used without further purification.Formaldehyde tert-butyl hydrazone [27] and organocatalyst I [28] were synthesized according to the literature.Racemic azocompounds (±)-1-6b [24,26] and 1,2-diols (±)-1d and (±)-4,5d [29] exhibited the same physical and spectral properties as described in the bibliography and the nuclear magnetic resonance (NMR) data of the Regarding the aliphatic diol (±)-6d, in which the stereogenic center presents an aliphatic substituent, the enzymatic acylation in TBME at 30 • C using PSL-C afforded a very low enantioselectivity value (E = 7) in a very fast resolution, achieving a 45% of (+)-6e after 2 h.In view of this result, the isozyme A from Candida antarctica was tested, leading to a slower (c = 17% after 6 h), but slightly more selective process (E = 12) than with PSL-C (see entry 10).In order to improve the reaction selectivity, the PSL-C-catalysed resolutions were carried out using ethyl acetate as a less reactive acyl donor.After 6 h, a 15% of (+)-6e was obtained in a process with a low enantioselectivity (E = 10).Finally, the kinetic resolution in presence of vinyl acetate was conducted at 10 • C.After 4 h, a 17% of diol 6d was converted into the acetate (+)-6e with 86% ee, but the E value was only increased to 16 (entry 12), indicating that this substrate was not appropriate for the biocatalysed acylation.

General Procedure for the Synthesis of Racemic Azocompounds (±)-1-6b
The corresponding α-ketoester 1-6a (5.0 mmol) was dissolved at room temperature in toluene (8.0 mL) and N-tert-butyl formaldehyde hydrazone (10 mmol) was added.For the ethyl benzoylformate, catalyst I (0.5 mmol, 10 mol%) was added prior to the hydrazone.Reactions were stirred at room temperature for 24 h until consumption of the starting material (TLC).The solvent was eliminated under reduced pressure and the obtained crudes were purified by column chromatography using toluene/EtOAc mixtures as the eluent in order to obtain the corresponding racemic azocompounds (±)-1-6b with yields between 80 and 92%.

a
,2-diols (±)-1-6d in a three-step procedure starting from the commercially available α-ketoesters 1For reaction conditions, see Materials and Methods.b Reaction performed with 10 mol% of catalyst I.

a
For reaction conditions, see Materials and Methods.b Reaction performed with 10 mol% of catalyst I.

Figure 1 .
Figure1.Effect of the PSL-C recycling on the conversion (blue line) and on the enantioselectivity (red bar) in the catalysed kinetic resolution of (±)-1d in tert-butyl methyl ether (TBME) at 30 °C using vinyl acetate as he acyl donor.

Figure 1 .
Figure 1.Effect of the PSL-C recycling on the conversion (blue line) and on the enantioselectivity (red bar) in the catalysed kinetic resolution of (±)-1d in tert-butyl methyl ether (TBME) at 30 • C using vinyl acetate as he acyl donor.