Conservation of the Biocatalytic Activity of Whole Yeast Cells by Supported Sol – Gel Entrapment for Efficient Acyloin Condensation

In this study, an efficient and generally applicable 2nd generation sol – gel entrapment method was developed for immobilization of yeast cells. Cells of Lodderomyces elongisporus, Candida norvegica, Debaryomyces fabryi, Pichia carsonii strains in admixture with hollow silica microspheres support were immobilized in sol – gel matrix obtained from polycondensation of tetraethoxysilane. As biocatalysts in the selective acyloin condensation of benzaldehyde catalyzed by pyruvate decarboxylase of the yeast, the novel immobilized whole-cell preparations were compared to other states of the cells such as freshly harvested wet cell paste, lyophilized cells and sol – gel entrapped preparations without hollow silica microspheres support. Reusability and storability studies designated this novel 2nd generation sol – gel method as a promising alternative for solid formulation of whole-cells bypassing expensive and difficult downstream steps while providing easy-to-handle and stable biocatalysts with long-term preservation of the biocatalytic activity.


Introduction
The biocatalytic systems based on whole-cells are blooming recently, due to the continuously growing interest on synthetic biology, protein engineering and recombinant techniques providing novel cells with extremely widespread biocatalytic activity of various origin [1,2]. In addition, the more and more screening tools for novel enzymatic activities within cells opens up new possibilities to realize various valuable biocatalytic processes with wild type cells as well. Whole-cell biocatalysts have many advantages compared to the isolated enzymatic systems, especially the avoidance of expensive and time-consuming downstream processes including enzyme purification represents significant benefits. Moreover, whole-cells containing natural co-factors, co-enzymes and their natural regeneration systems offer ideal environment for works with cofactor requiring enzymes [3,4]. On the other hand, the whole-cells suffer a few drawbacks such as low tolerance of organic solvents, difficult co-factor regeneration and sometimes lower activities and side reactions [5].
Among whole-cells, yeasts represent one of the most traditional biocatalyst. While yeast cells can express diverse enzymatic activities, they are easy-to produce and handle, stable and commonly non-pathogenic. Their redox, C-C coupling or hydrolytic enzyme sets get significant attention in several syntheses of chiral intermediates or compounds [6][7][8][9][10]. Amongst the yeast catalyzed biotransformations, the carbon-carbon condensation is one of the most important reaction to produce optically pure pharmaceutically active ingredients (APIs) like (R)-phenylacetyl carbinol being the precursor of ephedrine and pseudoephedrine [11,12]. Pyruvate decarboxylase (PDC, EC 4.1.1.1) -as a valuable member of the natural enzyme toolbox of yeasts -can catalyze the non-oxidative decarboxylation of pyruvate to acetaldehyde. Within the same enzyme, the accumulated acetaldehyde can react with activated acetaldehyde forming acetoin (3-oxo-2-butanol). Some yeasts, such as various Saccharomyces, Zymomonas, Candida or Hansenula strains, are able to catalyze the carbon-carbon condensation of several aromatic aldehyde analogues with the activated acetaldehyde leading to valuable acyloins [13,14].
Immobilization of cells and enzymes has been researched simultaneously since many decades. Even today -due to the extremely expanding requirements of biotechnologyenzyme and cell immobilization is a rapidly developing area. The main aims of immobilization strategies are the isolation, compartmentalization and stabilization of the valuable cells, cell organelles or enzymes with high biocatalytic activity [14][15][16][17][18]. Although several immobilization methods and carrier systems for cells or enzymes are based on similar methods (binding reactions and media, carrier materials etc.), in many instances enzyme and cell immobilization require significantly different approaches.
Among cell immobilization techniques, adsorption on tissue-like or macroporous solid supports and entrapment in gel matrices such as alginate, polyacrylamide, carrageenan, chitosan or agar based matrices are used commonly [18][19][20][21][22][23]. In most of these cases, traditional organic materials are applied which are natural, biocompatible, commercially available and easy-to-handle. In many instances, however, the stability, durability, mechanical and chemical resistance of the resulted immobilized whole-cell preparations based on such traditional natural materials are not satisfactory for certain applications [24,25].
The sol -gel systems based on polycondensation of various alkoxysilane precursors can provide porous organo-silica network with fine-tunable properties for many smart application such as embedding several biological components from small biomolecules to cells with enhanced stability [23,[26][27][28], sensor or optical development [29][30][31]. Tetraethoxysilane (TEOS) was successfully used as precursor material for sol -gel immobilization of plant and bacterial cells [26,30,[32][33][34]. The resistance of yeast (Saccharomyces cerevisiae) cells entrapped in sol -gel matrices in different organic media was studied earlier indicating that the sol -gel network could protect the living cells against the lethal effect of organic solvents. A mixture of tetra-, tri-and diethoxysilane precursors was also applied for immobilization of baker yeast demonstrating increased stability and biocatalytic activity of the immobilized cells in sucrose degradation [35,36]. In our previous report, co-immobilization of two different types of whole-cells -recombinant Escherichia coli cells expressing ω-transaminase and Lodderomyces elongisporus cells with ketoreductase activity -was developed by their simultaneous entrapment in a sol -gel system [4].
In this study, the biocatalytic potential of yeast cells of four different strains (Candida norvegica, Debaryomyces fabryi, Lodderomyces elongisporus and Pichia carsonii) were examined in acyloin condensation of benzaldehyde (1). Biocatalytic ability of different forms of the yeast cells in the acyloin condensation catalyzed by the PDCs present in the different yeasts were compared involving freshly harvested cell pellets, lyophilized cells and two different immobilized forms of freshly harvested cells [1. entrapped in simple 1 st generation sol -gel matrices (nanoporous network from TEOS as precursor) and 2. entrapped in a 2 nd generation supported sol -gel system (nanoporous network from TEOS as precursor supported by hollow silica microspheres)] which has not been reported earlier (Fig. 1). The effect of co-factors such as thiamine pyrophosphate (TPP) and magnesium chloride and different buffer systems were also studied. The reusability and storability of immobilized wholecell biocatalysts were examined as well.  (HiMedia M425). Cultivation was carried out as described earlier [10] in a 10 L fermenter.
The morphology of the freshly harvested cells was described by Olympus BX-43 light microscope and the immobilized biocatalysts was investigated with a JEOL JSM-5500LV scanning electron microscope (SEM). For SEM analysis samples were covered with gold, then electron beam energy of 10 kV was used.

Production of freshly harvested yeast cells
After cultivation, cells were centrifuged and washed with phosphate buffer (100 mM, pH 7.0), then the buffer was decanted and the residue was stored at 4 °C until use.

Lyophilization of cells
Following cultivation, the cell lyophilization was carried out as described earlier [10].

Immobilization of cells by the 1 st generation sol -gel method
Silica sol was prepared by addition of tetraethoxysilane (TEOS, 0.72 mL) to a solution of distilled water (0.25 mL) containing 0.1 M HNO 3 (65 µL) followed by sonication of the resulted mixture for 5 min at RT (Emag Emmi 20HC Ultrasonic Bath, 45 kHz) and cooling at 4 °C for 24 h. In case of simple sol -gel entrapment, cell paste (0.5 g) was suspended in phosphate buffer (3 mL, 0.1 M, pH 7.5) than silica sol was mixed with the cell suspension and the resulted mixture was shaken (Technokartell Test Tube Shaker Model T3SK, 40 Hz) at RT for 5 min. After gelation at RT (occurred within 30 min), the resulted gel was aged at 4 °C for 48 h. The crude biocatalyst was washed with phosphate buffer (2 × 15 mL, 100 mM, pH 7.5), centrifuged and dried at room temperature (24 h), and stored at 4 °C.

Immobilization of cells by the 2 nd generation sol-gel method using hollow silica microspheres as support
In case of 2 nd generation sol -gel entrapment, a suspension was prepared by addition of MAT540 support (150 mg) to freshly harvested (0.5 g) yeast cells resuspended in phosphate buffer (3 mL, 0.1 M, pH 7.5). The supported cells were shaken (Technokartell Test Tube Shaker Model T3SK, 40 Hz) until became homogeneous suspension (cc. 5 min at RT). The following steps were the same as previously described in the case of simple sol -gel entrapment. Cell immobilization yield (Y IC , %) was over 99 % as determined by viable cells counting on dilution plates and Helber counting chambers as previously described [4].

Reusability of the sol -gel entrapped yeast cell biocatalyst
The reusability of Lodderomyces elongisporus yeast cells entrapped in 2 nd generation sol -gel systems (Le-SG-2) was investigated in acyloin condensation reactions as described in Subsection 2.2.6 repeated five times. After a reaction, the immobilized biocatalyst was isolated by filtration (glass filter G4) and washed with phosphate buffer (5 mL, 100 mM, pH 7.0) three times and dried to constant weight. The mass of the biocatalyst was measured to ensure the constant reaction conditions in every reaction cycle.

Storage stability of the sol -gel entrapped yeast cell biocatalyst
The long-term stability or storability of the Lodderomyces elongisporus yeast cells entrapped in 2 nd generation sol -gel systems (Le-SG-2) was investigated by biocatalytic activity determination in acyloin condensation of benzaldehyde (1) as described in Subsection 2.2.5. During the examined time period (6 month), biocatalyst was stored at 4 °C.

Calculations
Conversion of the substrate (c, %), enantiomeric excess (ee, %), molar percentage of the components (x, %) and molar ratio of acyloin (R)-2 and alcohol 3, specific biocatalyst activity (U B , U g −1 ) and cell content based specific activity (U C , U g −1 ) were calculated at 3 h by using Eqs.
where n S and n P are the molar amounts of substrate (S) and product(s) (P), where n and n * are the molar amounts of the major and the minor enantiomer, where U is rate of the substrate conversion [µmol min −1 ], t is reaction time, m B is the mass of the biocatalyst, where m C is the mass of the cells in the immobilized biocatalyst.

Results and discussion
Based on scientific reports and our previous knowledge on microbial screening of ketoreductase activities in various yeast strains [7,10,[37][38][39][40], four different yeast strains (Candida norvegica, Debaryomyces fabryi, Lodderomyces elongisporus and Pichia carsonii) were selected for investigating their biocatalytic activity in the acyloin condensation of benzaldehyde (1). First, two different forms of non-immobilized yeast cells -freshly harvested cell paste and lyophilized cells powder -were screened as biocatalysts with acyloin forming ability ( Table 2). While only ketoreductase activity was detectable with the freshly harvested cell paste without observable PDC activity, the lyophilized cell form of all the four investigated strains resulted in formation of the acyloin product (R)-2 with moderate to good conversion (c 1 = 17-100 %), acceptable to excellent acyloin ratio (R = 0.47-21, the molar ratio of acyloin (R)-2 and alcohol 3) and in high enantiomeric excess (ee (R)-2 ≥ 89 %). These difference in acyloin forming ability can be attributed to inactivation of the most active ketoreductases on lyophilization due to the significantly different water content of the cell paste and lyophilized cells. The lyophilized cells of Lodderomyces elongisporus proved to be the most efficient acyloin producing biocatalyst providing the highest quantity the acyloin product (R)-2 (x = 59 %) with high enantiopurity (ee = 96 %). Thus, yeast cells from Lodderomyces elongisporus were selected for further cell immobilization experiments involving two different sol -gel systems. The simple, 1 st generation sol -gel method was based only tetraethoxysilane (TEOS) as monomers to form a nanoporous network entrapping the cells. In case of the 2 nd generation sol -gel method, wholecells and a proper inert solid support were co-immobilized. Hollow silica microspheres with amino/vinyl grafted surface (M540) were previously described as efficient enzyme carrier [41,42], thus M540 was tried in 2 nd generation sol -gel systems as support material as well.
The two sol -gel entrapment methods with freshly harvested cell paste were investigated by testing the acyloin forming ability of the immobilized cell biocatalysts from benzaldehyde (1). Although the different forms of whole-cells exhibited comparable biocatalytic activity ( U B ) in consuming benzaldehyde (1), the specific biocatalytic activity ( U C ) related to the cell content of the biocatalysts differed significantly (Table 3), being almost twice as high of the 2 nd generation sol -gel preparation compared to the freshly harvested cell paste. In addition, fresh cell paste and 1 st generation, support-free sol -gel systems exhibited almost exclusively alcohol dehydrogenase activity, while the PDC activity leading to acyloin product with remarkable enantiopurity could be observed only with the lyophilized whole-cells and with the 2 nd generation supported / sol -gel entrapped biocatalyst.
The successful co-embedding of the Lodderomyces elongisporus yeast cells (typical cell size 3-10 µm, Fig. 2 A) and the hollow silica microspheres (mean particle size 10-20 µm, Fig. 2 B) within the sol -gel network can be seen on the scanning electron microscopy (SEM) image of the 2 nd generation sol -gel biocatalyst (Fig. 2 C), where compact particles can be observed with 50-70 µm diameter.
Thiamine pyrophosphate (TPP) and MgCl 2 are known co-factors for PDC [43,44], thus the effect of addition of TPP and MgCl 2 to different buffer systems was studied in the acyloin condensation of 1 catalyzed by the 2 nd generation sol -gel form of Lodderomyces elongisporus (Le-SG-2) (Fig. 3). Supplementation of the reaction with TPP and MgCl 2 had enhancing effect on the biocatalytic activity at each pH between 5.0-7.0 (Fig. 3 A). The pH influenced a R corresponds to molar ratio of acyloin (R)-2 and alcohol 3.; b x 2 is the amount of (R)-2 in the mixture, calculated from GC data; c No acyloin product (R)-2 could be detected.
the selectivity towards the PDC-catalyzed acyloin condensation (Fig. 3 B). Among the different buffer systems, phosphate buffer at pH 7.0 led to the highest quantity of the acyloin product (R)-2 (x = 70 %). Reusability of biocatalyst is a key issue, especially when a sustainable process based on biotransformation should be developed. The durability of the 2 nd generation sol -gel system-based Lodderomyces elongisporus biocatalyst (Le-SG-2) was investigated in five consecutive reactions for acyloin condensation of 1 (Fig. 4). Although with some fluctuation, the total biocatalytic activity ( U B ) and the PDC activity (based on x 2 ) of the Le-SG-2 biocatalyst were preserved after the five repeated cycles. This is a promising result, considering the severe sensitivity and recovery problems with the freshly harvested cell paste or the lyophilized cell powder.
The most promising Le-SG-2 biocatalyst was also investigated before and after use with scanning electron microscopy (SEM). SEM investigation revealed, that the characteristic morphology of the freshly prepared immobilized biocatalyst -solid irregular particles with 30-60 µm mean particle size -could be preserved in the reusability experiments (Fig. 5).
Long-term storability of biocatalysts under easily sustainable conditions -such as storage at room temperature or in a common refrigerator (4 < T < 8 °C) -is commonly a required issue. Thus, a test of the residual biocatalytic activity of the Lodderomyces elongisporus cells in acyloin formation from benzaldehyde 1 (residual U B , %, relative to the initial biocatalytic activity); used as wet cell paste, lyophilized cell powder and as the two sol -gel encapsulated (1 st and 2 nd generation) was performed after 6-month storage at 4 °C (Table 4).
While the non-immobilized wet cell paste lost all biocatalytic activity, the lyophilized cell powder and the 1 st generation sol -gel system could preserve mostly the alcohol dehydrogenase activity. Fortunately, the 2 nd generation sol -gel system-based Lodderomyces elongisporus biocatalyst could conserve the PDC activity quite efficiently.  a U B biocatalytic activity, b U C specific biocatalytic activity, c x 2 is the amount of (R)−2 in the mixture, calculated from GC data. d No acyloin product (R)−2 could be detected.

Conclusion
A novel, so-called 2 nd generation cell immobilization method was developed based on sol -gel technique enabling in situ co-immobilization of yeast whole-cells and hollow silica microspheres as solid support material. This immobilization method could convert Lodderomyces elongisporus yeast cells to stable immobilized biocatalyst with pyruvate decarboxylase activity capable of performing acyloin condensation of benzaldehyde 1.
In contrast to the freshly harvested cell paste or the lyophilized cell powder forms, the novel immobilized form was able to conserve the biocatalytic activity the immobilized yeast cells and provided excellent reusability (biocatalyst recovery in five consecutive reaction cycles) and storability (6-month storage at 4 °C).
In conclusion, combination of organosilane-based solgel network with hollow silica microspheres as proper support material can provide new generation hybrid wholecell biocatalysts as promising sustainable alternative to the equipment-and cost-demanding lyophilization.   in 2 nd generation sol -gel system (Le-SG-2) 91. 8 41 a U B biocatalytic activity, b x 2 is the amount of (R)-2 in the mixture, calculated from GC data.