Biotechnol Lett DOI 10.1007/s10529-014-1519-0

ORIGINAL RESEARCH PAPER

Impact of the configuration of a chiral, activating carrier on the enantioselectivity of entrapped lipase from Candida rugosa in cyclohexane Jan Tobis • Joerg C. Tiller

Received: 14 February 2014 / Accepted: 20 March 2014 Ó Springer Science+Business Media Dordrecht 2014

Abstract Lipase from Candida rugosa was loaded into an amphiphilic polymer co-network (APCN) composed of the chiral poly[(R)-N-(1-hydroxybutan2-yl) acrylamide] [P-(R)-HBA] and P-(S)-HBA, respectively, linked by poly(dimethylsiloxane). The nanophase-separated amphiphilic morphology affords a 38,000-fold activation of the enzyme in the esterification of 1-phenylethanol with vinyl acetate. Further, the enantioselectivity of the entrapped lipase was influenced by the configuration of the chiral, hydrophilic polymer matrix. While the APCN with the (S)configuration of the APCN affords 5.4 faster conversion of the (R)-phenylethanol compared to the respective (S)-enantiomer, the (R)-APCN allows an only a 2.8 faster conversion of the (R)-enantiomer of the alcohol. Permeation-experiments reveal that the enantioselectivity of the reaction is at least partially caused by specific interactions between the substrates and the APCN.

J. Tobis Freiburg Material Research Center and Institute for Macromolecular Chemistry, University of Freiburg, Stefan-Meier-Str. 21, 79104 Freiburg, Germany J. C. Tiller (&) Department of Bio- and Chemical Engineering, TU Dortmund, Emil-Figge-Str. 66, 44227 Dortmund, Germany e-mail: [email protected]

Keywords Amphiphilic polymer co-networks  Biocatalysis  Enantioselectivity  Immobilization  Lipase  Non-aqueous media  Organic solvents

Introduction Making enzymes more active as well as more regioand enantioselective in non-aqueous media is the key issue of successfully using these biocatalysts for the organic synthesis of complex chemical compounds (Carrea and Riva 2000). Generally, enzymes can be activated in organic solvents by using additives such as structure-preserving salts (Khmelnitsky et al. 1994), crown ethers (Broos et al. 1995) or adsorbed surfactants (Paradkar and Dordick 1994). Further, they can be activated by immobilizing them in polymeric carriers (Wang et al. 1997) and in a silica sol–gel matrix (Reetz 1997). We have discovered previously that enzymes become particularly active in organic solvents by entrapping them in amphiphilic polymer co-networks (APCN) (Bruns and Tiller 2005). By varying the composition to make such networks swell in fluorophilic solvents, they also activate enzymes in supercritical CO2 (Bruns and Tiller 2006; Bruns et al. 2008). The enantioselectivity of enzymes in organic solvents can be greatly influenced by protein engineering methods such as directed evolution (Reetz et al. 2001). By analogy to that, medium engineering

123

Biotechnol Lett

evidenced a strong impact of the solvent on the stereoselectivity of enzymes in non-aqueous media (Fitzpatrick and Klibanov 1991; Natalia et al. 2012). Also the method of how an enzyme is immobilized onto a matrix can have a great impact on the enantioselectivity of the captured biocatalyst (Godoy et al. 2013). Although chiral supports, such as chitosan (Mallin et al. 2013), have been used for improved enantioselectivity, the direct influence of the chiral configuration of the support matrix on an entrapped biocatalyst has not been considered yet. In this study, we combine the activation capability of an APCN as enzyme carrier with a possibly enantioselective directing chiral hydrophilic phase.

Materials and methods Materials Methacryloxymethyl terminated poly(dimethylsiloxane) [PDMS(MA)2] was kindly provided by Wacker Chemie AG (Munich, Germany) and was purified via dialysis in methanol for 4 days with a cellulose membrane (MWCO 1,000 g/mol) (Mn = 4560 g/mol, Mw = 6180, Mw/ Mn = 1.36 (GPC in chloroform, polystyrene as standard). The photoinitiator, Irgacure 651 (2,2-dimethoxy1,2-diphenylethanone), was a gift from Ciba Specialty Chemicals (Basel, Switzerland). The racemic 1-phenylethanol, vinyl acetate, acryloyl chloride, ethyl acetate, cyclohexane, and chlorotrimethylsilane were purchased from Sigma-Aldrich and were used without further purification. (R)-2-Aminobutan-1-ol (technical grade) was purchased from Fluka. (S)-2-Aminobutan-1-ol was purchased from Molekular (Shaftesbury, UK). Cyclohexane was used as obtained and contains 37 ppm water according to Karl-Fischer-titration. The glass slides were kindly provided by Marienfeld (Lauda-Koenigshofen, Germany). Adhesive poly(propylene)-tape (PP; Tesafilm, Tesapack 4024) was purchased from Tesa AG (Hamburg, Germany). Lipase from Candida rugosa (lyophilized powder, 2 U/mg) was purchased from Sigma-Aldrich. Polymer co-network synthesis The chiral acrylamides (R)-N-[1-(trimethylsilyloxy)butan-2-yl]acrylamide [(R)-TMSBA] and the

123

respective (S)-enantiomer were synthesized as previously reported (Tobis et al. 2010). A monomer mixture containing PDMS(MA)2, (R)TMSBA and (S)-TMSBA, respectively, the photoinitiator Irgacure 651 and ethyl acetate was homogenized and poured on a glass slide. A second glass slide coated with a PP adhesive tape and two additional stripes (each 50 lm thickness) as spacers was pressed on top. The monomer mixture was photopolymerized in a commercially-available UV reactor (Heraflash, Heraeus Kulzer, Germany) for 4 9 180 s. The glass slides were separated and the co-network was removed from the adhesive tape. The product film was subsequently incubated at room temperature twice in 20 ml 2-propanol/1 mM HCl (50:50 v/v) and once in the same volume of a 2-propanol/water (50:50 v/v) overnight. The networks were subsequently washed with tetrahydrofuran and water. Samples were then air dried on a PTFE sheet and subsequently vacuum dried. Optically clear films with an average thickness ranging from 110 to 135 lm were obtained. Polymer co-network characterization Swelling measurements were carried out by measuring the dimension of the dry and the swollen sample, respectively (swelling in water at room temperature for 24 h). The degree of swelling was calculated as S = V (swollen)/V (dry). The tapping-mode atomic force microscopy (AFM) was carried out with MultiMode AFM and nanoscope IIIa controller (Digital Instruments) at ambient conditions in phase mode and by using NCL-W (Tapping Mode) cantilevers (Nanosensors, Neuchatel, Switzerland). AFM images were recalculated regarding the corrected volume fractions (density corrected weight fractions) of the respective nanophases by using the image analysis software Pyphant (Freiburg Materials Research Center, Freiburg, Germany). Enzyme loading Samples of the APCNS (50 mg) were pre-swollen in aqueous phosphate buffer (0.1 M, pH 6.0) and then added to 1 ml of the same buffer containing 2 mg CRL/ml and shaken at 4 °C for 24 h. The sample was rinsed three times with 5 ml of the buffer. The moist sample was frozen in liquid N2 and lyophilized overnight. Protein loading was determined by

Biotechnol Lett

measuring the protein amount in the combined loading and washing solutions using a Lowry assay and subtracting this amount from the protein amount in the fresh loading solution. Permeability determination Permeation measurements were carried out by using a two-chamber diffusion cell (Tobis et al. 2011). The blocks were equipped with steering bars, had a volume of 3 ml each and a contact area of 0.78 cm2. A circular membrane, diam. 12 mm, was cut from an APCN sample. The thickness was measured with a micrometer screw and the APCNs were equilibrated at ambient temperature in cyclohexane. During the experiment the concentration of the tracer [5 mM (R)- and (S)-1-phenylethanol, respectively] is considered constant and the first fick law of diffusion for unlimited reservoirs is applicable. From the slope of the transported amount of substance DN across the area A divided by the term Dc/Dz against the time Dt the permeability of the molecule is obtained from Eq. 1. DN 1
¼ P  Dt A DDcz

ð1Þ

Equation 1 is the calculation of the permeability. DN, amount of transported substance; A, contact area; Dc, concentration difference between donor and stripping phase; Dz, membrane thickness; P, permeation constant; Dt, time interval. Catalytic activity CRL-loaded samples (50 mg) were added to 4 ml cyclohexane containing 2 mM of a racemic mixture of 1-phenylethanol, 40 mM vinyl acetate, and 1 mM undecane. The mixture was shaken at 25 °C. Altogether five samples of 50 ll each were taken over 29 h. The samples were characterized by GC using a chiral column (CP Chirasil, Varian) and N2 as carrier gas (1 ml/min). The injection volume was 1 ll and a temperature ramp of 4 °C/min was applied from 120 to 165 °C. The sample was analyzed with a FI detector using undecane as internal standard. The respective free enzyme was used as commercially available powder, suspended in the same reaction mixture (50 mg/ml) under vigorous stirring.

Results and discussion The objective of this work was to explore how the chiral phase of an APCN influences the enantioselectivity of an entrapped enzyme in an organic solvent. The concept of APCNs as activating carrier is based on their two interconnected orthogonal nanophases, which allow an enzyme to be entrapped in the hydrophilic region and the substrate to reach the enzyme via the hydrophobic phase (Erdodi and Kennedy 2006). After having prepared three APCNs with a chiral P-(R)-HBA phase in different compositions and one with 50 % (w/w) P-(S)-HBA, the networks were characterized by AFM to analyze the distribution of the polymer nanophases. The measured images were then recalculated adjusting contrast and off-set to obtain an image that contains the two polymer phases in maximal contrast and true volume ratios (Fig. 1). The morphology of the APCN consisting of 29 % (w/w) P-(R)-HBA and 71 % (w/w) PDMS is dominated by the soft PDMS matrix (15–30 nm). The brittle P-(R)-HBA forms slender branches of 3–10 nm diameter, which are interconnected to a considerable extent. Increasing the weight content of the hydrophilic phase to 50 % (w/w), resulted in a bicontinuous structure. P-(R)-HBA forms a finely branched phase of 12–35 nm diameter. The PDMS forms roundish domains of up to 50 nm diameter. The respective APCN with 50 % (w/w) P-(S)-HBA has a similar morphology (not shown). The APCN with 71 % (w/w) of P-(R)-HBA forms a morphology dominated by the brittle P-(R)-HBA containing partially connected and isolated domains of the PDMS. Given the above mentioned requirements, the 50 % (w/w) APCN should be the best performing system. In order to find a suitable model, we were looking for an enzyme-catalyzed reaction that exhibits a low enantioselectivity to judge the influence of the chiral carrier. Lipases from Candida antarctica, Aspergillus niger, Rhizopus arrhizus and C. rugosa were evaluated with respect to their enantioselectivity in catalyzing the esterification of racemic 1-phenylethanol with vinyl acetate in cyclohexane (Fig. 2). While the first three enzymes afforded more than 90 % of (R)-1-phenyl ethyl acetate in the product mixture, the lipase from C. rugosa (CRL) was the least selective [67 % of (R)-1-phenyl ethyl acetate] and was therefore used in further experiments.

123

Biotechnol Lett

Fig. 1 Morphology of P-(R)-HBA-l-PDMS with (left) 29 wt%, (middle) 50 wt%, and (right) 71 wt% P-(R)-HBA. The hard P-(R)-HBA phase appears red and the soft PDMS phase blue.

The morphologies were calculated from the respective phase mode AFM images processed with the image processing program Python

Fig. 2 Reaction scheme of the transesterification of (R,S)-1-phenylethanol and vinyl acetate

The loading of the co-networks with CRL was performed by immersing the pre-swollen samples in an aqueous buffer solution of the enzyme at 4 °C for 24 h, followed by washing with the buffer and subsequent lyophilization. The enzyme uptake was determined by subtracting the protein amount in the combined washing solutions from the protein amount in the starting lipase solution. The results are given in Table 1. The obtained enzyme loadings are 50 ± 0.7 nmol/gAPCN in the case of an APCN with a P-(R)-HBA content of 29 % (w/w); increasing to 54.8 ± 0.9 nmol/gAPCN and 54.5 ± 0.6 nmol/gAPCN for a P-(R)-HBA content of 50 % (w/w) P-(R)-HBA and 71 % (w/w) P-(R)-HBA, respectively. As shown previously, lipases are attracted by the hydrophobic interface of APCNs (Dech et al. 2011). Thus, the increase of the weight content of the hydrophobic PDMS from 29 to 50 % (w/w) results in an increase of the CRL loading in an APCN. A further increase of the weight content of PDMS to 71 % (w/w) increases the hydrophobic interface but lowers the swellability of the APCN in water (Table 1). The results indicate that the obtained enzyme concentrations within the networks are a function of the degree of swelling of the APCNs in water and the extent of the hydrophobic interface.

123

Table 1 Composition dependent loading of CRL in APCNs APCN % (w/w) PDMS

Swelling S in watera

CRL loading (nmol/gAPCN)

71b

1.42 ± 0.21

50.0 ± 0.7

50b

1.23 ± 0.06

54.8 ± 0.9

50c

1.25 ± 0.05

54.8 ± 0.6

29b

1.07 ± 0.04

54.5 ± 1

The loading was performed with 2.0 mg CRL/ml in 0.1 M phosphate buffer at pH 6.0 and shaken at 4 °C for 24 h. All experiments were performed in duplicate a

Swelling after 24 h at ambient temperature, S = V (swollen)/V (dry)

b

P-(R)-HBA-l-PDMS

c

P-(S)-HBA-l-PDMS

Since the APCN with 50 % (w/w) PDMS shows high CRL loading in combination with an optimal nanostructure, the biocatalytic activity of the CRL-loaded networks was explored for APCNs with 50 % (w/w) P-(R)HBA and P-(S)-HBA, respectively. To this end, the CRL-loaded APCN was immersed in a cyclohexane containing a mixture of racemic 1-phenylethanol and vinyl acetate. The product formation at 25 °C was monitored by periodical sampling and a subsequent GC analysis. Although both co-networks were loaded with

Biotechnol Lett

Fig. 3 Transesterification reaction kinetics of (R,S)-1-phenylethanol with vinyl acetate in cyclohexane catalyzed by CRL entrapped in left P-(R)-HBA-l-PDMS and right P-(S)-HBA-l-

PDMS. Both APCNs contain 50 % (w/w) PDMS and the reaction is carried out at 25 °C. Filled dot (R)-1-phenyl ethyl acetate, hollow dot (S)-1-phenyl ethyl acetate

CRL to the same extent, the overall activity of CRL is dependent on the chiral configuration of the hydrophilic phase (Fig. 3). While CRL in the P-(S)-HBA containing APCN shows a specific overall activity Zspez,R?S of 6 lmol/(lmol enzyme per h), the overall activity of the enzyme in the P-(R)-HBA is with 7.7 lmol/(lmol enzyme per h) more than 20 % higher. In comparison to the native CRL powder, the APCN activity of CRL is increased up to 38,500 times. This finding is in line with other previously reported activations of lipases upon immobilization (Adlercreutz 2013). Further, the immobilization of the lipase in the chiral network mainly preserved the natural preference for (R)-1-phenylethanol as substrate (Table 2). A closer look reveals an increase of the specific activity Zspez,R of the lipase from 4.4 to 6.5 lmol/(lmol enzyme per h) with the change of the configuration of the amphiphilic carrier from P-(R)HBA-l-PDMS to a P-(S)-HBA-l-PDMS. The specific activity Zspez,S on the other hand decreases from 1.6 to 1.2 lmol/(lmol enzyme per h) when changing the (R)APCN to the (S)-configuration. As the ratio of the specific activities Zspez,R/Zspez,S increases from 2.8 in the case of a P-(R)-HBA-l-PDMS to 5.4 utilizing the respective (S)-APCN, the enantioselectivity of the transesterification of (R,S)-1-phenylethanol and vinyl acetate in cyclohexane is a function of the configuration of the chiral hydrophilic matrix surrounding the enzyme. The influence of the chiral configuration of the enzyme matrix on the CRL might be caused by configuration-dependent interactions of the matrix and the substrate. In order to explore this, the permselectivity (ratio of permeabilities PR/PS) of the amphiphilic co-network was examined for the 1-phenylethanol

stereoisomers. The permeation of (R)-1-phenylethanol across an APCN with a composition of 50 % (w/w) P-(R)-HBA was measured with an initial permeability PR of 7.4 9 10-11 m2/s which increases to 8.8 9 10-11 m2/ s after 4.5 h. The permeability PS of (S)-1-phenylethanol through the same network of 6.8 9 10-11 m2/s remains constant within the course of the experiment. Thus, the P-(R)-HBA-l-PDMS amphiphilic co-network shows a permselectivity of PR/PS of 1.2 towards (R)-1-phenylethanol, i.e. the (S)-1-phenylethanol is somewhat stronger binding to the matrix. As shown previously, the P-(S)-HBA network acts vice versa (Tobis et al. 2011). These measurements also allow the Thiele moduli to be calculated for high stationary substrate concentration according to Schoenfeld et al. (2013). The calculated values are in a range of 0.02 and 0.05 indicating no intrinsic diffusion limitation of the bioconversion within the APCN. The comparison of the permeation experiment with the selectivity of the biotransformation within a chiral host reveals that the binding of the substrate to the matrix might be the reaction influencing parameter. It seems that stronger binding does lead to an increased esterification rate for the respective substrate. We hypothesize that stronger binding of a substrate to the surface of the non-swollen CRL containing the P-HBA phase results in a relatively higher concentration of this substrate, which is then preferentially converted by the nearby enzyme. Besides the substrate/matrix binding, there seems to be an impact on the selectivity of the APCN-entrapped CRL itself. Presuming that the stronger binding substrate is only converted faster due to higher concentration, the reaction rate of the other substrate is not affected. Thus,

123

Biotechnol Lett Table 2 Enzyme catalyzed transesterification of (R,S)-1-phenylethanol and vinyl acetate in cyclohexane at ambient temperature catalyzed by native and in amphiphilic co-networks immobilized CRL CRL matrix

Zspez,S Zspez,R [lmol/(lmol enzyme per h)]

Zspez,R?S

–

Zspez,R/Zspez,S

eeb (%)

1.3 9 10-4

0.7 9 10-4

2 9 10-4

2

33

a

4.4 ± 0.3

1.6 ± 0.1

6.0 ± 0.4

2.8

69

P-(S)-HBAa

6.5 ± 0.8

1.2 ± 0.2

7.7 ± 1

5.4

47

P-(R)-HBA

The reaction was carried out in cyclohexane solution containing 2 mM of (R,S)-1-phenylethanol and 40 mM of vinyl acetate at 25 °C under shaking a

50 % (w/w) in the APCN, CRL loading 58 nmol/gAPCN

b

Enantiomeric access of (R)-1-phenyl ethyl acetate in the product mixture after 29 h

the reaction rate of the (R)-substrate in the (R)-APCN (4.4 lmol/lmol enzyme per h) and that of the (S)substrate in the (S)-APCN (1.2 lmol/lmol enzyme per h) are representing the CRL activity in the APCN without chiral influence. Thus, the selectivity of the CRL entrapped in the APCN increases compared to the free powder from 2 to 3.7 favoring the conversion of the (R)1-phenylethanol.

Conclusion A lipase from C. rugosa was entrapped into the hydrophilic, chiral nanophase of an APCN consisting of a PDMS-based makro-cross-linker and a chiral poly[(R,S)-N-(1-hydroxybutan-2-yl) acrylamide]. The biocatalytic esterification of racemic (R,S)-1-phenylethanol revealed that the lipase entrapped the (S) configuration of the APCN affords a two-fold higher selectivity towards the conversion of the (R)-1-phenylethanol compared to the enzyme entrapped in the respective (R)-APCN. Moreover, the comparison with permeation experiments suggests an activation of the more tightly bonded enantiomer. We believe that this concept might be transferred to other enzymes and reactions with chiral substrates and/or products. The most important prerequisite for enantioselective influence of the carrier is that the chiral matrix forms selective diasteriomeric complexes with the substrate or the product.

References Adlercreutz P (2013) Immobilisation and application of lipases in organic media. Chem Soc Rev 42(15):6406–6436

123

Broos J, Kakodinskaya IK, Engbersen JFJ, Verboom W, Reinhoudt DN (1995) Large activation of serine proteases by pretreatment with crown ethers. J Chem Soc Chem Commun 2:255–256 Bruns N, Tiller JC (2005) Amphiphilic network as nanoreactor for enzymes in organic solvents. Nano Lett 5:45–48 Bruns N, Tiller JC (2006) Nanophasic amphiphilic co-networks with a fluorophilic phase. Macromolecules 39:4386–4394 Bruns N, Bannwarth W, Tiller JC (2008) Amphiphilic co-networks as activating carriers for the enhancement of enzymatic activity in supercritical CO2. Biotechnol Bioeng 101(1):19–26 Carrea G, Riva S (2000) Properties and synthetic applications of enzymes in organic solvents. Angew Chem Int Ed Engl 39:2226–2254 Dech S, Cramer T, Ladisch R, Bruns N, Tiller JC (2011) Solid– solid interface adsorption of proteins and enzymes in nanophase-separated amphiphilic co-networks. Biomacromolecules 12:1594–1601 Erdodi G, Kennedy JP (2006) Amphiphilic co-networks: definition, synthesis, applications. Prog Polym Sci 31:1–18 Fitzpatrick PA, Klibanov AM (1991) How can the solvent affect enzyme enantioselectivity. J Am Chem Soc 113:3166–3171 Godoy CA, Romero O, delas Rivas B, Mateo C, FernandezLorente G, Guisan JM, Palomo JM (2013) Changes on enantioselectivity of a genetically modified thermophilic lipase by site-directed oriented immobilization. J Mol Catal B 87:121–127 Khmelnitsky YL, Welch SH, Clark DS, Dordick JS (1994) Salts dramatically enhance activity of enzymes suspended in organic solvents. J Am Chem Soc 116:2647–2648 Mallin H, Menyes U, Vorhaben T, Hoehne M, Bornscheuer UT (2013) Immobilization of two (R)-amine transaminases on an optimized chitosan support for the enzymatic synthesis of optically pure amines. ChemCatChem 5:588–593 Natalia D, Greiner L, Leitner W, Ansorge-Schumacher MB (2012) Stability, activity, and selectivity of benzaldehyde lyase in supercritical fluids. J Supercrit Fluids 62:173–177 Paradkar VM, Dordick JS (1994) Aqueous-like activity of alpha-chymotrypsin dissolved in nearly anhydrous organic solvents. J Am Chem Soc 116:5009–5010 Reetz MT (1997) Entrapment of biocatalysts in hydrophobic sol–gel materials for use in organic chemistry. Adv Mater 9(12):943–954

Biotechnol Lett Reetz MT, Wilensek S, Zha D, Jaeger K-E (2001) Directed evolution of an enantioselective enzyme through combinatorial multiple-cassette mutagenesis. Angew Chem Int Ed Engl 40:3589–3591 Schoenfeld I, Dech S, Ryabenky B, Daniel B, Glowacki B, Ladisch R, Tiller JC (2013) Investigations on diffusion limitations of biocatalyzed reactions in amphiphilic polymer co-networks in organic solvents. Biotechnol Bioeng 110:2333–2342

Tobis J, Thomann Y, Tiller JC (2010) Synthesis and characterization of chiral and thermo responsive amphiphilic conetworks. Polymer 51:35–45 Tobis J, Boch L, Thomann Y, Tiller JC (2011) Amphiphilic polymer co-networks as chiral separation membranes. J Membr Sci 372:219–227 Wang P, Sergeeva MV, Lim L, Dordick JS (1997) Biocatalytic plastics as active and stable materials for biotransformations. Nat Biotechnol 15(8):789–793

123