Journal of Chromatography A, 1391 (2015) 72–79

Contents lists available at ScienceDirect

Journal of Chromatography A journal homepage: www.elsevier.com/locate/chroma

Enzymatic hydrolysis in an aqueous organic two-phase system using centrifugal partition chromatography夽 J. Krause, T. Oeldorf, G. Schembecker, J. Merz ∗ Laboratory of Plant and Process Design, Department of Biochemical and Chemical Engineering, TU Dortmund University, D-44227 Dortmund, Germany

a r t i c l e

i n f o

Article history: Received 30 October 2014 Received in revised form 25 February 2015 Accepted 26 February 2015 Available online 5 March 2015 Keywords: Centrifugal Partition Chromatography Countercurrent chromatography Multi-phase reaction system Biocatalysis Lipase Candida rugosa

a b s t r a c t Multi-phase reaction systems, mostly aqueous organic systems, are used in enzyme catalysis to convert hydrophobic substrates which are almost insoluble in aqueous media. In this study, a Centrifugal Partition Chromatograph is used as a compact device for enzymatic multi-phase reaction that combines efficient substrate supply to the aqueous phase and separation of both phases in one apparatus. A process design procedure to systematically select the aqueous and organic phase to achieve stable and efficient reaction rates and operation conditions in Centrifugal Partition Chromatography for efficient mixing and separation of the phases is presented. The procedure is applied to the hydrolysis of 4-nitrophenyl palmitate with a lipase derived from Candida rugosa. It was found that the hydrolysis rate of 4-nitrophenyl palmitate was two times higher in Centrifugal Partition Chromatography than in comparable stirred tank reactor experiments. © 2015 Elsevier B.V. All rights reserved.

1. Introduction Since most of the biocatalysts, referred as enzymes, are active in aqueous solutions only, the substrates to be converted are limited to those soluble in aqueous media. For the enzymatic conversion of hydrophobic substrates, approaches like the use of non-aqueous or multi-phase systems as reaction milieu are needed. In case of multiphase systems, mostly aqueous–organic systems, the hydrophobic substrate is dissolved in the organic phase and mixed with the aqueous phase containing the biocatalyst. Besides the conversion of hydrophobic substrates the main advantage of using multi-phase reaction systems is the ability of “in situ” product removal from the catalytically active aqueous phase to the organic phase. As a consequence, the reaction equilibrium can be shifted toward the product side and product inhibition of the enzyme can be eliminated [1]. However, multi-phase reaction systems lead to different design aspects compared to monophasic aqueous systems [2]. Efficient mixing and separation of the phases need to be realized and mass transport hindrances for the substrate to reach the enzyme and the product to partition in the organic phase need to be taken into account [3–5]. Another challenge of aqueous–organic reaction systems is the preservation of the enzymatic activity. Beside good

夽 Presented at the 8th International Conference on Countercurrent Chromatography – CCC 2014, 23–25 July 2014, Uxbridge, United Kingdom. ∗ Corresponding author: Tel.: +49 0231 755 4325; fax: +49 0231 755 2341. E-mail address: [email protected] (J. Merz). http://dx.doi.org/10.1016/j.chroma.2015.02.071 0021-9673/© 2015 Elsevier B.V. All rights reserved.

mixing behavior to increase the interfacial area for mass transport, the effect of the interfacial forces on the activity of the enzyme needs to be considered [6]. Additionally, organic solvents can harm enzymes and significantly decrease their catalytic activity, so solvent screening becomes necessary before setting up the reaction. Different approaches to enhance the use of multi-phase systems for enzymatic reactions are available like genetic engineering to stabilize the enzymes toward organic solvents and the process conditions applied. Also different reactor concepts to increase mixing and optimize separation of the phases are available [6,7]. These reactor concepts range from classical mixer-settler devices to membrane contactors, and all aim to increase interfacial area to minimize interfacial mass transfer hindrances and simultaneously contacting the aqueous catalytically active phase with fresh organic substrate [7–9]. The CPC can be an efficient mixer that creates a large interfacial area for an efficient distribution of components between the phases and allows settling of the phases in one device. Commonly, CPC is used for liquid–liquid chromatography [10]. In CPC one phase of any two-phase system is immobilized and used as stationary phase. The immobilization is achieved in a chamber system that is arranged around a rotary axis. By rotating the system a centrifugal field is generated which keeps the stationary phase in the chamber cascade. The second phase is pumped through the stationary one and therefore referred to as mobile phase. Depending on the density of the phases used, two main operation modes are defined. In case the lighter phase is used as mobile phase CPC is operated in ascending mode. Separation with the heavier phase as mobile

J. Krause et al. / J. Chromatogr. A 1391 (2015) 72–79

73

(pNPB) (≥98%), palmitic acid (PA) (≥99%), 4-morpholinepropanesulfonic acid (MOPS) (≥99.5%), 2-propanol (≥99.5%), nheptane (≥99%), cyclohexanone (≥99%), methyl tert-butyl ether (≥99.8%), 1-octanol (≥99%), diisopropyl ether (≥98.5%), methyl isobutyl ketone (≥98.5%), trichloroethylene (≥99.5%), ethyl acetate (≥99.8%) and the lipase from Candida rugosa (L1754) were purchased from Sigma–Aldrich Co. LLC, Germany.

Fig. 1. From left to right: basic scheme of enzymatic multi-phase biocatalysis; CPC chamber in ascending mode (organic mobile phase (darker) and aqueous, catalytically active phase stationary phase (light)); real hydrodynamic in a single CPC chamber.

phase is called descending mode [10,11]. Commonly, the CPC technique is used for the purification of natural extracts including the use of aqueous two-phase systems for the purification of proteins [12–15]. In this study the Centrifugal Partition Chromatography (CPC) is investigated as a new reactor concept for aqueous–organic enzyme catalysis. The stationary liquid phase can be used to immobilize the catalytic active enzyme in solution without handling solid materials or functionalizing surfaces as proposed for other multiphase reactor concepts [16,17]. The mobile organic phase is used to supply substrate and remove product simultaneously. The concept is presented in Fig. 1. Utilizing CPC and CCC for reactions is not new. Some initial studies were published in the late 90s. Berthod [18] published a study about a chemical reactions conducted in CCC in 1998. He found that this brought various advantages toward a comparable plug-flow reactor. Studies about enzymatic reactions in CPC were published from 1998 to 1999 by van Hollander, who described the reaction and separation of chiral amino acids in an aqueous two-phase system by pulsing substrate into the CPC. This led to a process that combined reaction and separation of products and non-converted substrates based on chromatographic mechanisms. He also provided a mathematical model to predict concentration profiles in the CPC [19,20]. However, the CPC concept presented is batch mode due to pulsing the substrate solution and, therefore, capacity is limited. Additionally, an aqueous two-phase system was used, not allowing the conversion of hydrophobic substrates as proposed in the multi-phase reaction approach discussed before. Hence, a detailed description of the CPC for aqueous organic enzyme reactions with continuous product removal is not discussed yet. In this study, a procedure for the selection of an appropriate aqueous organic two-phase system for the conversion of hydrophobic substrates in CPC is developed, following the approach presented by Schembecker and Tlatlik [21] who proposed to determine optimal operating windows for the reaction, separation and the apparatus separately and operating the integrated process in the overlapping zone of these different windows. The procedure is applied to the hydrolysis of 4-nitrophenyl palmitate with a lipase derived from Candida rugosa. In their natural inhabitant lipases catalyze the hydrolysis directly at interfaces and are very stable in contact with organic substances. They are the most used enzymes in organic synthesis [6,22,23]. Beside the process design procedure the CPC as apparatus for multi-phase reactions is evaluated. Key issues to be addressed are the long-term stability of the chosen lipase from Candida rugosa and a comparison of the CPC technique to a classical stirred tank reactor system. 2. Materials and methods

2.2. Model system A wild type lipase from Candida rugosa is used as model enzyme. For demonstration purposes the hydrolysis of 4-nitrophenyl palmitate (pNPP) to 4-nitrophenol (pNP) and palmitic acid (PA) is investigated. The substrate pNPP is almost insoluble in aqueous media and suitable for the investigation of a multi-phase reaction system and the product pNP is easily detectable due to its strong yellow color in aqueous solutions. 2.3. Analytical procedures 2.3.1. Enzymatic activity assay For fast determination of lipase activity a test based on 4nitrophenyl butyrate (pNPB) was adapted from Zhong [24]. The test was modified using 50 mM 4-morpholinepropanesulfonic acid (MOPS) buffered lipase stock (0.15 g/L) and substrate solution (0.75 mM) to ensure no significant pH drop due to production of PA in the measured time frame. Both solutions are mixed in 50:50 volume ratios and absorbance is measured at 410 nm to determine pNP production kinetics. Temperature was kept constant at 25 ◦ C. The enzymatic activity U is defined as 1 ␮mol pNP produced in 60 s/mg lipase at 25 ◦ C using the substrate pNPB. All reported activities were averaged from triplicates. 2.3.2. pNP quantification in organic solvents To measure the pNP concentration in organic solvents (mobile phase) a sample (1 ml) is taken from the organic phase and extracted for 10 min with 1 ml of buffered aqueous solution (50 mM MOPS; pH 7.2). Then the absorbance is measured at 410 nm in 500 ␮l of the aqueous phase. For calibration, spectrophotometric grade pNP was dissolved in the organic solvent at defined concentrations and then extracted for 10 min into the aqueous phase. The product concentration as function of the absorbance at 410 nm was then derived by measuring the absorbance of the extracted aqueous phase. 2.4. Experimental procedures 2.4.1. Effect of pH-value on enzyme activity The pH-value for optimum activity of lipase Candida rugosa is determined using pNPP as substrate in 50 mM MOPS buffer at different pH-values. In addition, 2-propanol is used to dissolve the product pNPP in small amounts in the aqueous phase. This method is adapted from Ref. [25] and does not influence the lipase activity. The substrate solution, containing 2-propanol, pNPP and 50 mM MOPS buffer, is adjusted to different pH values and mixed with a 50 mM MOPS buffer with the same pH but containing enzyme (0.15 g/L). The activity is measured using the spectrophotometric method as mentioned before but by using the reaction substrate pNPP. The pH value is varied between pH 6.8 and pH 7.6 (pKs value of MOPS = 7.2). The temperature is kept constant at 25 ◦ C for all experiments. All experiments were performed in triplicates.

2.1. Chemicals 4-Nitrophenol (pNP) (spectrophotometric grade), 4nitrophenyl palmitate (pNPP) (≥98%), 4-nitrophenyl butyrate

2.4.2. Effect of organic solvent on enzyme activity To screen the enzymatic activity in a two-phase system using different organic phases the lipase Candida rugosa is dissolved in

74

J. Krause et al. / J. Chromatogr. A 1391 (2015) 72–79

buffered aqueous solution (50 mM MOPS, pH 7.2) at a concentration of 0.2 g/L. This enzyme solution is mixed with different organic solvents (volume ratio 50:50) at 160 RPM on a PTR-60, GrantInstruments Ltd. (Shepreth, United Kingdom) shaker. The activity is measured after 30 min, 1 h, 2, 4, 6, 8 and 24 h. For every solvent and time investigated a single experiment is prepared. As reference the activity development of the lipase (0.2 g/L) in 50 mM MOPS buffer at pH 7.2 is measured. After sampling the aqueous phase the pNPBactivity assay is used as described above. Due to the high amount of solvents screened some experiments are conducted in triplicates only to confirm reliability. 2.4.3. Measurement of partition coefficients The partition of the components is measured using shake flask experiments. The components are solved in a two-phase system and mixed at 160 RPM on a PTR-60, Grant-Instruments Ltd. (Shepreth, United Kingdom) for 1 h. Then the phases are analyzed to quantify the amounts in every phase by using the HPLC method published by Alarcón [26], however using an injection volume of 20 ␮l instead of 5 ␮l. Based on these results the partition coefficients KD are calculated using Eq. (1). KD =

corg caq

(1)

2.4.4. CPC set-up In this work a FCPC® by Kromaton (Annonay, France) is used. The original rotor, commercially available from Kromaton is replaced with a single disk rotor. Chamber and channel dimensions of the single disk are the same as of the original FCPC® rotor, but instead of 1320 chambers and a rotor volume of 200 ml (original rotor) the single disk rotor has a volume of 10 ml and is segmented in 66 chambers only. The CPC device is equipped with an image processing system taking videos of the hydrodynamic inside the chambers. A LED is used to illuminate the disk from one side of the rotor disk. The camera (not shown) and the LED are triggered in that way, that the same chambers are recorded once per rotation. The instrumental setup is described in detail by Adelmann and Schwienheer [15,27]. For temperature control of the CPC rotor and the outlet stream two thermostats are used. Two solvent delivery pumps with 50 ml pump heads (Varian PrepStar SD-1) are used for pumping mobile and stationary phase into the CPC. 2.4.5. Preparation of CPC For the reaction experiments in CPC the CPC-rotor is prepared as follows. The CPC is filled with the organic phase at 800 RPM using a mobile phase volume flow of 10 ml/min for at least 5 times the rotor volume in descending mode. After this the rotor is filled with the aqueous phase without enzyme at the same operating conditions in ascending mode, followed by another filling cycle with organic phase in descending mode. After the rotor is completely filled with organic phase the aqueous phase containing the enzyme is pumped into the rotor for at least 10 times the rotor volume to ensure that the rotor contains enzyme rich aqueous phase only. Then the organic phase without substrate is pumped into the rotor using the ascending mode to equilibrate the system and to ensure steady state operation. 2.4.6. Reaction in CPC After preparing the CPC and setting steady state operation for the operation conditions (rotational speed and volume flow) chosen, the reaction is started by pumping the mobile phase with substrate into the CPC. Additionally, the rotor exit is connected to the feed vessel with a volume of 75 ml (mobile phase with substrate) to recycle the mobile phase. This operation mode can be

used to adjust the residence time and enables extended reaction times and activity observation as well as easy sampling. The pNP concentration in the feed vessel was measured every 30 min. 2.4.7. Stirred tank reactor experiments To evaluate the CPC experiments the hydrolysis was set up in a multi-phase stirred tank reactor experiment in laboratory scale. The volume ratio of stationary aqueous to mobile organic phase was derived from the CPC experiments to achieve an adequate comparability. All operating times, concentration and masses were chosen equal to the CPC-experiments. The temperature was kept constant using a double jacket reactor and an external cooling device. 3. Results and discussion Phase system selection procedures for efficient separation processes in CPC are well established, like the HemWAT or Arizona systems [12,28,29]. For enzymatic reaction systems in CPC these phase selection approaches must be adapted due to various reasons. On the one hand, the partition coefficients of the components involved (substrate, product, enzyme, etc.) must be considered separately [12]. For example, the enzyme must be immobilized in the aqueous stationary phase. Hence, a partition coefficient of 0 (see Eq. (1)) is intended, as otherwise the enzyme would partition to the mobile phase and with it leaves the CPC column. On the other hand the organic solvents cannot be selected considering the partition coefficients only. Due to the fact that the solvent could inactivate or even destroy the enzyme a solvent screening is mandatory. Additionally, the partitioning of the other reaction components and additives like enzyme co-factors or regeneration systems to provide co-substrates or energy carriers like adenosine triphosphate (ATP) need to be adjusted carefully to enable a viable reaction system with simultaneous product removal. Taking these facts into account a new three-step procedure for designing an efficient reaction system using CPC has been developed (see Fig. 2). First step of the procedure, shown in Fig. 2, is the optimization of the aqueous phase by selecting the buffer type, buffer concentration and pH-value for the enzymatic reaction. Important questions for the aqueous phase optimization are whether co-factors, co-enzymes, co-substrates, energy carriers and/or regeneration systems are essential for the reaction and how these compounds can be brought into the system. In case of several reactions in the aqueous phase conditions need to be chosen which enhance the reaction system in total. For the model system investigated, the hydrolysis of pNPP, no co-substrates, co-factors or energy carriers are needed. Hence, the components to be considered are the substrate pNPP, the products pA and PNP and the lipase only. In step 1.b and 1.c an appropriate buffer type and concentration must be selected. In former works at our laboratory the 4-morpholinepropanesulfonic acid (MOPS) buffer in a concentration of 50 mM was taken for investigations (data not published). Referring to step 1.d the enzymatic activity was measured at different pH-values and plotted in Fig. 3. Lipase from Candida rugosa showed the highest activity at pH 7.2. Based on this all aqueous solutions used in this study are 50 mM MOPS buffer at pH 7.2. Step 2 of the procedure focusses on the selection of a suitable organic phase. First (step 2) literature is researched for already used solvents or multi-phase systems used for the reaction and enzyme. Additionally to be investigated constraints like the use of green solvents only or toxicity limitations can also considered to pre-select solvents. Next, the physicochemical properties of the solvents themselves are evaluated in terms of viscosity, because using CPC as reaction system the mobile, organic phase, is pumped through the whole CPC system. The pressure drop of the apparatus

J. Krause et al. / J. Chromatogr. A 1391 (2015) 72–79

75

Fig. 2. Procedure to determine media and operation conditions in reactive aqueous organic two-phase systems for Centrifugal Partition Chromatography.

is limited, and thus, highly viscous solvents are not recommended for the use in CPC [30]. With the pre-screened solvents activity preservation tests are carried out to select the solvents which maintain the enzymatic activity in small scale. Then the partition

Fig. 3. pH dependency of lipase from Candida rugosa activity in MOPS (50 mM) buffer.

behavior of all components involved is measured to detect the suitable organic phase. In this study, a pre-selection of suitable organic solvents was carried out based on literature and physiochemical properties of the solvents. Highly toxic substances were excluded to decrease the risk potential. Two-phase solubilizers are not considered because they equalize the partition properties of the components in the two-phase system, which is not favored for a CPC reaction system as mentioned before. From the remaining list, organic solvents with a viscosity