Microbial Hydrolysis of Racemic β-Aryl-γ-ethylidene-γ-lactones and Antifeedant Activity of the Products against Alphitobius diaperinus Panzer

Hydrolysis of (±)-β-aryl-γ-ethylidene-γ-lactones by fungal strain Aspergillus ochraceus AM370 afforded (−)-(S)-γ-ethylidene-γ-lactones 2a–d and (+)-(R)-γ-ketoacids 3a–d. Enantiomeric purity of the unreacted lactones was strictly related to a size of an aryl substituent at C-4 of γ-lactone ring, with the highest ee (77%) obtained for the (−)-(S)-γ-ethylidene-γ-lactone possessing unsubstituted benzene ring (2a) and the lowest one (15%) determined for the (−)-(S)-γ-ethylidene-γ-lactone with bulky 1,3-benzodioxole system (2d). Lactones 2a–d, both racemic and enantiomerically enriched, as well as products of their hydrolysis showed varying degrees of feeding deterrent activity against lesser mealworm, Alphitobius diaperinus Panzer, which depended on the structure of the compound and the developmental stage of the lesser mealworm. In the case of adults, more active were γ-lactones 2a–d, compared with ketoacids 3a–d. Only in the case of lactone 2a was the effect of configuration of stereogenic center on the activity found. Particularly strong deterrents against this stage (T > 180) were racemic and (−)-(S)-γ-ethylidene-γ-lactone with p-methoxysubstituted phenyl ring (2c).


Introduction
One of the most common methods for the production of enantiopure compounds is the enzymatic kinetic resolution of racemic mixtures [1][2][3][4]. In this process, the yield of products is limited to a maximum of 50% due to the fact that one of enantiomers remains unreacted [3,4]. On the contrary, in dynamic kinetic resolution, the non-reacting enantiomer is racemized in situ, which allows to overcome the limitation of 50% conversion [5,6]. Despite these limitations, kinetic resolution is still a valuable tool leading to optically pure compounds. In enzymatic kinetic resolution, oxidoreductases [7] and hydrolases are mainly used [8]. Among hydrolases, lipases found a widest application due to a high substrate specificity, selectivity, stability, ease of immobilization, and no need of cofactors [9][10][11].
Structures of (+)-γ-ketoacids 3a-d were confirmed by spectroscopic data. The presence of a carboxyl group was proved by broad bands of O-H stretching vibrations at the range of 3519-2930 cm −1 for 3a, 3424-2926 cm −1 for 3b, 3367-2930 cm −1 for 3c, 3424-2360 for 3d and strong bands of C=O stretching vibrations at 1713, 1715, 1769, 1728 cm −1 for 3a-d, respectively. In the 13 C NMR spectra signals (δ = 177.6 for 3a, 177.2 for 3b, 177.6 for 3c and 177.2 ppm for 3d) undoubtedly indicated the presence of carbon atom from carboxylic group. Signals from C-4 for 3a-d were found at δ = 209.7, 209.7, 209.8 and 209.5, respectively and confirmed the presence of carbonyl group. In the 1 H NMR spectra characteristic shapes and chemical shifts of signals from protons of ethyl group (e.g., for ketoacid 3b triplet at 0.97 ppm and quartet at 2.42 ppm) proved the location of carbonyl group in these molecules at C-4. Formation of ketoacid is the result of keto-enol tautomerism and predomination of keto form over an enol tautomer which is a direct product of γ-ethylidene-γ-lactone hydrolysis (Scheme 1).
Biotransformation rate was strictly related to the size of aryl substituent. After 24 h for γ-lactone with unsubstituted benzene ring 2a ( Figure 1a) and methyl substituted benzene ring 2b (Figure 1b) conversion exceeded 25%. Significantly slower reaction rate was observed for lactone with methoxy substituted benzene ring (2c) ( Figure 1c) and 1,3-benzodioxole system (2d) (Figure 1d); for these substrates during first 24 h conversion reached only 13% and 9%, respectively. After three days of process conversion of lactones, 2a and 2b stopped at approximately 50%; at the same time, conversion of lactones 2c and 2d was only 28% and 21%, respectively. For these two substrates only after five days of biotransformation 51% conversion in the case of 2c and 48% in the case of 2d was observed and no further transformation was observed. In order to achieve maximal isolated yield of products with highest possible optical purity, biotransformations were stopped after three days in the case of lactones 2a,b and five days in the case of lactones 2c,d. In the 1 H NMR spectra characteristic shapes and chemical shifts of signals from protons of ethyl group (e.g., for ketoacid 3b triplet at 0.97 ppm and quartet at 2.42 ppm) proved the location of carbonyl group in these molecules at C-4. Formation of ketoacid is the result of keto-enol tautomerism and predomination of keto form over an enol tautomer which is a direct product of γ-ethylidene-γ-lactone hydrolysis (Scheme 1). Biotransformation rate was strictly related to the size of aryl substituent. After 24 h for γ-lactone with unsubstituted benzene ring 2a ( Figure 1a) and methyl substituted benzene ring 2b (Figure 1b) conversion exceeded 25%. Significantly slower reaction rate was observed for lactone with methoxy substituted benzene ring (2c) ( Figure 1c) and 1,3-benzodioxole system (2d) (Figure 1d); for these substrates during first 24 h conversion reached only 13% and 9%, respectively. After three days of process conversion of lactones, 2a and 2b stopped at approximately 50%; at the same time, conversion of lactones 2c and 2d was only 28% and 21%, respectively. For these two substrates only after five days of biotransformation 51% conversion in the case of 2c and 48% in the case of 2d was observed and no further transformation was observed. In order to achieve maximal isolated yield of products with highest possible optical purity, biotransformations were stopped after three days in the case of lactones 2a,b and five days in the case of lactones 2c,d. Enantiomeric excesses of unreacted isomers of lactones 2a-d determined at about 50% conversion are given in Table 1, which clearly shows a decrease of enantiomeric purity of slower reacting lactones along with the increase of size of an aryl substituent. The highest ee was achieved for γ-lactone possessing unsubstituted benzene ring (2a). Lactone 2b with methyl substituted benzene ring was obtained with 56% ee, the presence of methoxy group (compound 2c) resulted in lowering the optical purity of unreacted lactone. The lowest enantioselectivity of reaction was Enantiomeric excesses of unreacted isomers of lactones 2a-d determined at about 50% conversion are given in Table 1, which clearly shows a decrease of enantiomeric purity of slower reacting lactones along with the increase of size of an aryl substituent. The highest ee was achieved for γ-lactone possessing unsubstituted benzene ring (2a). Lactone 2b with methyl substituted benzene ring was obtained with 56% ee, the presence of methoxy group (compound 2c) resulted in lowering the optical purity of unreacted lactone. The lowest enantioselectivity of reaction was observed for the lactone with bulky 1,3-benzodioxole system 2d. The explanation of decreasing enantiomeric purity of lactones 2a-d is probably a steric hindrance which impedes an access of the substrate to the pocket of enzyme. Similar effect of 1,3-benzodioxole system on the effectiveness of the kinetic resolution was observed in our earlier studies over the lipase-catalyzed transesterification of (E)-4-(benzo[d] [1 ,3 ]dioxol-5 -yl)-but-3-en-2-ol [40]. Resolution of bulky substrates has been a remarkable challenge in organic chemistry. These limitations can be overcome by mutagenesis, the example is engineering of an epoxide hydrolase from Bacillus megaterium ECU1001 to improve resolution of α-naphtyl glycidyl ether, bulky precursor of (S)-propranolol-drug blocking β-adrenergic receptor [41]. Enantiopreference of Aspergillus ochraceus AM370 during hydrolysis of racemic γ-ethylidene-γ-lactones 2a-d was proved based on the determination of configuration of slowly reacting enantiomer of the γ-lactone with p-methylphenyl substituent (−)-2b. For that purpose, optically pure (−)-γ-ethylidene-γ-lactone 2b with S-configuration at C-4 was obtained from corresponding chiral δ-iodo-γ-lactone 1 with defined configuration of chiral centers [35] by dehydrohalogenation with DBU (Scheme 2). Negative specific rotation sign measured for this isomer and for enantiomer of 2b, and slower hydrolyzation by Aspergillus ochraceus AM370 (Scheme 1) proved the S configuration of the letter. Taking into consideration the structural similarity of all four lactones subjected to microbial hydrolysis, the S configuration was also assigned to the slower reacting (−)-enantiomers of γ-lactones 2a, 2c and 2d. Consequently, we could conclude that Aspergillus ochraceus AM370 prefers hydrolysis of (R-lactones 2a-d to corresponding (R)-ketoacids 3a-d. (Scheme 1). Unfortunately, despite using different GC and HPLC chiral columns and separation conditions, including resolution of methyl and ethyl ester derivatives, our attempts to separate enantiomers of 3a-d chromatographically failed and their enantiomeric compositions could not be determined. observed for the lactone with bulky 1,3-benzodioxole system 2d. The explanation of decreasing enantiomeric purity of lactones 2a-d is probably a steric hindrance which impedes an access of the substrate to the pocket of enzyme. Similar effect of 1,3-benzodioxole system on the effectiveness of the kinetic resolution was observed in our earlier studies over the lipase-catalyzed transesterification of (E)-4-(benzo[d][1′,3′]dioxol-5′-yl)-but-3-en-2-ol [40]. Resolution of bulky substrates has been a remarkable challenge in organic chemistry. These limitations can be overcome by mutagenesis, the example is engineering of an epoxide hydrolase from Bacillus megaterium ECU1001 to improve resolution of α-naphtyl glycidyl ether, bulky precursor of (S)-propranolol-drug blocking β-adrenergic receptor [41]. Enantiopreference of Aspergillus ochraceus AM370 during hydrolysis of racemic γ-ethylidene-γ-lactones 2a-d was proved based on the determination of configuration of slowly reacting enantiomer of the γ-lactone with p-methylphenyl substituent (−)-2b. For that purpose, optically pure (−)-γ-ethylidene-γ-lactone 2b with S-configuration at C-4 was obtained from corresponding chiral δ-iodo-γ-lactone 1 with defined configuration of chiral centers [35] by dehydrohalogenation with DBU (Scheme 2). Negative specific rotation sign measured for this isomer and for enantiomer of 2b, and slower hydrolyzation by Aspergillus ochraceus AM370 (Scheme 1) proved the S configuration of the letter. Taking into consideration the structural similarity of all four lactones subjected to microbial hydrolysis, the S configuration was also assigned to the slower reacting (−)-enantiomers of γ-lactones 2a, 2c and 2d. Consequently, we could conclude that Aspergillus ochraceus AM370 prefers hydrolysis of (R)-lactones 2a-d to corresponding (R)-ketoacids 3a-d. (Scheme 1). Unfortunately, despite using different GC and HPLC chiral columns and separation conditions, including resolution of methyl and ethyl ester derivatives, our attempts to separate enantiomers of 3a-d chromatographically failed and their enantiomeric compositions could not be determined. The obtained results showed the effect of aryl substituent at C-4 on the rate of hydrolysis of γ-lactone ring and enantiomeric purity of the unreacted lactone. In the earlier work, Enzelberger et al. [30] conducted hydrolysis of a series of γ-lactones catalyzed by lipase from Pseudomonas sp. This enzyme preferentially hydrolyzed (R)-enantiomers of γ-lactones which is coherent with our results and preference of Aspergillus ochraceus AM370 to the hydrolysis of (R)-γ-lactones 2a-d.
Commercially available enzymes were also used for the hydrolysis of β-methyl-γ-ethylidene The obtained results showed the effect of aryl substituent at C-4 on the rate of hydrolysis of γ-lactone ring and enantiomeric purity of the unreacted lactone. In the earlier work, Enzelberger et al. [30] conducted hydrolysis of a series of γ-lactones catalyzed by lipase from Pseudomonas sp. This enzyme preferentially hydrolyzed (R)-enantiomers of γ-lactones which is coherent with our results and preference of Aspergillus ochraceus AM370 to the hydrolysis of (R)-γ-lactones 2a-d.
Commercially available enzymes were also used for the hydrolysis of β-methyl-γ-ethylidene lactones by Fajkowska et al. [31]. Lipase PS (from Pseudomonas cepacia) showed moderate enantiopreference to (+)-isomers of γ-lactones leaving unreacted (−)-isomers with ee in the range of 18-60% depending on the reaction time (4-15 days). The change of enantiopreference was observed if the reaction was catalyzed by lipase from Rhizopus niveus. In this case (−)-isomers of unsaturated γ-lactones were hydrolyzed faster to the corresponding ketoacids.

Antifeedant Activity
In our previous studies we presented i.a. synthesis of racemic γ-ethylidene-γ-lactones with aryl substituents at β-position (2a-d) and evaluation on their antifeedant activity towards Tribolium confusum, Trogoderma granarium and Sitophilus granaries [33,34]. In this work, we tested both racemic γ-ethylidene-γ-lactones and products of their hydrolysis for the antifeedant activity against lesser mealworm Alphitobius diaperinus Panzer ( Table 2). The comparative analysis of racemic and optically active form of lactones was supposed to deliver the information about the possible relationship between configuration of these compounds and their antifeedant activity. On the other hand, for the first time we could also compare the antifeedant activity of γ-lactones 2a-d with ketoacids 3a-d, products of their metabolism in the culture of Aspergillus ochraceus AM370. Nevertheless, the influence of aryl substituent could also be determined.
Feeding deterrent activity of compounds studied depended on their structure and the developmental stage of the lesser mealworm. Among racemic lactones particularly high activity against adults (T = 185.52) was observed for lactone with methoxysubstituted benzene ring (2c) (Entry 3). Very good antifeedants, with T coefficient higher than 150 for both developmental stages, were lactones with methylsubstituted benzene ring (2b) and 1,3-benzodioxole system (2d) (Entry 2,4). The relatively weakest feeding deterrent among racemic lactones, especially against larvae, was γ-lactone with unsubstituted benzene ring (2a) (Entry 1). Thus, the racemic lactones containing additional substituents at the benzene ring are stronger feeding deterrents and this relationship is particularly pronounced in the case of larvae of A. diaperinus. Similar trend was observed among enantiomerically enriched γ-lactones (Entry 5-8). However, in this series of compounds the activity was higher against adults of A. diaperinus in comparison with larvae. Feeding of adults was reduced in highest extent by (−)-(S)-γ-ethylidene-γ-lactone with methoxy group (−)-2c (Entry 7). Interestingly, the high values of T coefficients for this lactone (T = 188.75) and its racemic form (Entry 3) was the most similar to azadirachtin, the most active antifeedant of natural origin. Among enantiomerically enriched lactones (−)-(S)-γ-ethylidene-γ-lactone with methoxy group exhibited also the highest activity against larvae of the lesser mealworm (T = 140.95). These results confirmed significant positive influence of this substituent on the deterrent activity towards the lesser mealworm which was also observed for the series of β-aryl-γ-lactones in the previous tests against storage pests [33,34]. The presence of 1,3-benzodioxole system caused only a slight decrease of activity but these differences were not statistically significant (Entry 8).
Only in the case of lactones with unsubstituted benzene ring (2a) a significant relationship between their configuration and antifeedant activity was observed. For adults, more active was racemic lactone 2a (Entry 1) than its (−)-(S)-enantiomer (Entry 5); for larvae this relationship was reversed. Therefore, it can be supposed that in this case (S)-enantiomer is more active against larvae and (R)-enantiomer is better antifeedant against adults of A. diaperinus. The comparison of the other three pairs of racemic lactones and their (−)-(S)-enantiomers indicates no significant differences in their activity; probably because of lower enantiomeric excesses of (−)-(S)-enantiomers 2c-d comparing to lactone (−)-(S)-2a. Ketoacids 3a-d differed significantly from lactones in terms of the deterrent activity and they were weaker antifeedants for adults than for larvae of the lesser mealworm. In the case of adults decrease in activity along with the increase of size of an aryl substituent was observed. Ketoacid (+)-(R)-3a with unsubstituted benzene ring strongly reduced feeding of adults, but only in the choice test. In the no choice tests the level of feeding in the presence of all ketoacids was at a similar level. Comparing to lactones, ketoacids 3a-d were definitely weaker antifeedants against adults, with the exception of compounds (+)-(R)-3a (Entry 9) whose activity was similar to (−)-(S)-γ-lactone 2a (Entry 5). In the case of larvae, ketoacid (+)-(R)-3a with unsubstituted benzene ring (2a) was more active than its parent lactone whereas the activities of ketoacids with substituted benzene ring (3b,d) (Entry 10-12) were lower than those of corresponding racemic substrates (±)-2b,d.

Screening Procedure
Fungal strains were cultivated in 250 mL Erlenmeyer flasks containing 50 mL of the culture medium (3% glucose and 1% peptobac, pH = 6) at room temperature from 3 to 7 days. Racemic (E)-5-ethylidene-4-phenyldihydrofuran-2-one (2a) (10 mg/1 mL of acetone) was added to 300 mL Erlenmeyer flask with grown fungal biomass. Biotransformation was carried out with continuous shaking (140 rpm) at room temperature. Samples were taken after 1, 2, 3, 4, and 5 days of transformation, products were extracted with methylene chloride (3 × 10 mL). After centrifugation (7000 rpm) organic layers were separated, combined, dried over anhydrous MgSO 4 and filtered. After evaporation of solvent in vacuo the crude mixture was analyzed by TLC and GC. In order to identify the metabolites of the tested microorganisms, fungal cultures without addition of the substrate 2a were cultivated in the same conditions. The stability of the model substrate 2a in the culture media was checked as well.

General Procedure for Multiplied Scale Biotransformation of Racemic γ-Lactones 2a-d and Isolation of the Products
Aspergillus ochraceus AM370 strains was cultivated in 16 Erlenmayer flasks (250 mL volume) containing 50 mL of biotransformation medium described in screening procedure (Section 3.5). To each of 16 Erlenmeyer flasks containing biomass of Aspergillus ochraceus AM370, 10 mg of substrate dissolved in 1 mL of acetone was added (total volume of biotransformation medium was 800 mL, total substrate amount 160 mg). All reaction conditions were the same as described in the screening procedure. The biotransformations were carried out with continuous shaking (140 rpm) at room temperature. The progress of transformations was followed by GC and TLC (chloroform/methanol, 15:1, v/v) for visualization of γ-lactones (−)-2a-d and (chloroform/methanol/formic acid, 10:1:0.5, v/v) for detection of (+)-(R)-γ-ketoacids 3a-d. After 3 days (in the case of lactones 2a,b) or 5 days (in the case of lactones 2c,d) biomass was centrifuged and the products were extracted from the liquid medium with methylene chloride (3 × 30 mL). Extracts were dried over anhydrous MgSO 4 and filtered. Solvent was evaporated in vacuo and the biotransformation products were separated by column chromatography using two different developing systems: (chloroform/methanol, 15:1, v/v) to elute (−)-(S)-γ-ethylidene-γ-lactones 2a-d, followed by (chloroform/methanol/formic acid, 10:1:0.5, v/v) to isolate (+)-(R)-γ-ketoacids 3a-d.

Antifeedant Activity Assays
Feeding deterrent activity of (±)-γ-ethylidene-γ-lactones 2a-d and products of their biotransformations (−)-2a-d and (+)-3a-d against Alphitobius diaperinus Panzer in the choice test and the no-choice test was evaluated. Oat flakes were used as the test food. For feeding assays acetone solutions of tested compounds at a concentration of 10 mg/10 mL −1 were prepared. A quantity of 1 mL of a test solution or acetone alone as control was applied to 1 g of flakes by micropipette. After evaporation of solvent (30 min of air drying), the flakes were weighted and placed in petri dishes (15 cm diameter) together with ten larvae (25-30 days old) and ten unsexed adults (7-10 days old). In order to obtain large number of insects at approximately same age, the Rice-Lambkin culture method for lesser mealworm was used.
In the choice tests (where insects could choose either control or treated food), control and treated flakes in petri dishes were separated by a thin glass capillary. In the no-choice test, insects were exposed to only one kind of food-treated or control. Dishes were kept in the rearing chamber at 29 ± 1 • C in the dark for three days. After this period, the remaining uneaten oat flakes were reweighed, and the average weight of food was calculated. The amount of eaten food was the basis for calculating the deterrence coefficients (relative R and absolute A). Three coefficients of deterrence were calculated: (1) Relative, R = [C − E/C + E] × 100 (in the choice test) (2) Absolute, A = [CC − EE/CC + EE] × 100 (in the no-choice test) (3) Total, T = A + R, where C and CC is the amount of food consumed from control discs; E and EE is the amount of consumed food from discs treated with tested compounds [13].
The value of total coefficients of deterrence (T) ranges from −200 to +200. The negative value indicates that a compound is a feeding attractant. The positive value indicates feeding deterrent activity, within the range 0-50 described as week, 50-100-medium, 100-150 good and 150-200-very good.

Statistical Analysis
The mean values of the deterrence coefficients were compared by means of one-way analysis of variance (ANOVA) followed by Tukey's tests at a level of P < 0.05.

Conclusions
Reported for the first time A. ochraceus-catalyzed hydrolysis of racemic lactones is a useful tool to obtain both enantiomerically enriched (S)-γ-ethylidene-γ-lactones (−)-2a-d and (+)-(R)-γ-ketoacids 3a-d. All eight compounds obtained by the biotransformation were not reported previously in optically active forms and three γ-ketoacids 3b,d were not described previously in the literature. Enantioselectivity of the transformation reaction is strictly related to a size of aryl substituent. Most of lactones 2a-d are stronger feeding deterrents for adults than for larvae of the lesser mealworm, whereas ketoacids 3a-d reduce feeding of larvae to a higher extent. The influence of a size of aryl substituent on antifeedant activity is also significant, especially the presence of p-methoxy group at the phenyl ring significantly increase the activity. No clear differences between racemic and enantiomerically enriched forms of lactones were observed. The exception was lactone 2a; in this case (S)-enantiomer is more active against larvae whereas (R)-enantiomer is more active against adults of the lesser mealworm. The racemic lactones with methylsubstituted benzene ring (2b) and 1,3-benzodioxole system (2d), due to their high activity against both developmental stages, can find practical applications for A. diaperinus control.
Funding: This research was supported financially by the European Union as a part of the European Social Fund, Human Capital Operational Programme