Theoretical and experimental study of enzyme kinetics in a microreactor system with surface-immobilized biocatalyst

Accepted Manuscript Theoretical and experimental study of enzyme kinetics in a microreactor system with surface-immobilized biocatalyst Nataša Miložič, Martin Lubej, Mitja Lakner, Polona Žnidaršič-Plazl, Igor Plazl PII: DOI: Reference:

S1385-8947(16)31782-X http://dx.doi.org/10.1016/j.cej.2016.12.030 CEJ 16195

To appear in:

Chemical Engineering Journal

Received Date: Revised Date: Accepted Date:

23 September 2016 7 December 2016 8 December 2016

Please cite this article as: N. Miložič, M. Lubej, M. Lakner, P. Žnidaršič-Plazl, I. Plazl, Theoretical and experimental study of enzyme kinetics in a microreactor system with surface-immobilized biocatalyst, Chemical Engineering Journal (2016), doi: http://dx.doi.org/10.1016/j.cej.2016.12.030

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Theoretical and experimental study of enzyme kinetics in a microreactor system with surface-immobilized biocatalyst Nataša Miložiča, Martin Lubeja, Mitja Laknerb, Polona Žnidaršič-Plazla, Igor Plazla, * a

Faculty of Chemistry and Chemical Technology, University of Ljubljana, Večna pot 113, SI-1001 Ljubljana, Slovenia

b

Faculty of Civil and Geodetic Engineering, University of Ljubljana, Jamova cesta 2, SI-1001 Ljubljana, Slovenia

ABSTRACT A mathematical model comprising transport phenomena and enzyme-catalyzed reaction performed on the inner walls of the continuously operated microreactor with surface-immobilized -transaminase

was

developed. Oriented enzyme immobilization enabling unhindered accessibility of enzyme active sites was obtained by using fusion protein N-SBM-ATA-wt consisting of selected ω-transaminase ATA-wt and the positively charged Zbasic2 tag, which established ionic interactions with silicon/glass microchannel surface. Enzyme-catalyzed transamination of (S)-(-)-α-methylbenzylamine and pyruvate to acetophenone and Lalanine was described by surface kinetics based on a ping-pong bi-bi mechanism. Reaction kinetic parameters were preliminarily defined in a batch system using various initial substrates concentrations and further applied in the surface reaction description. Based on the prevailing kinetic and convection/diffusion phenomena, the developed model could be reduced to the one-dimensional model which enabled immobilized enzyme concentration estimation and showed good agreeement with experimental data from the outlet of the microreactor at various flow rates and inlet substrates concentrations. Moreover, the model successfully predicted performance of two consecutively connected microreactors coated with N-SBM-ATA-wt and could be further used to design and optimize efficient and sustainable processes of chiral amine syntheses catalyzed with surface immobilized enzymes.

Keywords: microreactor; enzyme immobilization; surface-enzyme kinetics; transamination; mathematical modeling *

Corresponding author: [email protected]

1

GRAPHICAL ABSTRACT

HIGHLIGHTS


Continuous transamination in a microreactor with ω-transaminase was modeled.
Zbasic2-tagged enzyme was immobilized on the inner surface based on ionic interactions.
Surface reaction kinetics based on a ping-pong bi-bi mechanism was applied.
Based on time-scale analysis, a simplified 1D model was developed and verified.
Model enabled enzyme concentration estimation and microreactor performance forecast.

2

1. Introduction In the past two decades, microscale technology has emerged as the potent tool for a number of (bio)chemical applications, from analytics to process development and industrial production. Diminished reactor volumes and high surface-to-volume ratio result in substantially enhanced mass and heat transfer leading to improved safety and low waste production. Processes in microreactor systems are generally performed in continuous flow mode, where strict control of process parameters and reaction times is feasible due to the laminar fluid flow through microchannels. The use of microreactors enables process intensification and promotes integration both of process monitoring based on on-line or in-line analysis, as well as with downstream processing. Furthermore, numbering-up concept based on parallelizing multiple reactor units is accomplished to obtain large scale production systems, which eliminates the need for typically tedious scale-up used in conventional reactors, especially for the catalytic reactions [1-6]. Several recent review papers highlighted the opportunities and challenges of microscale teschnology also for the biocatalytic processes with isolated enzymes or whole-cells, both in analytics and in production of chemicals [2-6]. Biocatalyst immobilization is preferable and commonly employed in microreactor systems since it allows continuous mode of operation, often stabilyze the biocatalyst and facilitates product separation along with biocatalyst recovery enabling its reuse. Various immobilization methods used in conventional reactors could be employed also in microreactors, however, surface-immobilization exploring the big surface-tovolume ratio of microfluidic devices is often the technique of choice since it among others prevents the increase in backpressure, presenting a common problem of packed-bed or monolithbased microreactors [7-10]. On the other hand, biocatalyst immobilization can alter enzyme activity due to the conformational changes, steric hindrances, or non-oriented attachment of the enzyme to the support leading to the non-accessible active site, and can cause additional diffusional limitations of substrates and products to and from the biocatalyst active site, which all results in changes of reaction kinetics [7]. In order to prevent these effects, introduction of positively charged tags such as Zbasic2 (also termed as silica-binding module, SBM), which is a 7 kDa three-helix bundle obtained from the B domain of staphylococcal protein A with a high content of arginine, was proposed to induce ionic interactions with negatively charged silicon or glass microreactor walls, 3

leading to “smart” enzyme immobilization [7,11-13]. Recently, multiple Zbasic2 modules were introduced in a single sucrose phosphorylase molecule, which even increased the effectiveness of the immobilized enzyme by up to twofold [14]. In this study, oriented enzyme immobilization on the surface of the inner walls of a silicon/glass microchannel was enabled by using positively charged Zbasic2 tag fused with the selected ωtransaminase ATA-wt. ω-transaminases are pyridoxal 5’-phosphate (PLP)-dependent enzymes with broad substrate range catalyzing a reversible transfer of an amino group from an amine donor to a

prochiral ketone substrate acceptor molecule without the requirement of an adjacent carboxylic acid moiety obligatory with other transaminases. These enzymes are recently gaining attention as promising biocatalysts allowing for sustainable production of enantiomerically pure compounds that possess a chiral amine moiety and thereby are essential building blocks of intermediary metabolites and pharmaceutical drugs [15-17]. As a model reaction, enzymatic transamination of pyruvate with (S)-(-)-α-methylbenzylamine ((S)-α-MBA) yielding L-alanine (L-ALA) and acetophenone (ACP) was studied. Recently, a detailed study of reaction mechanism for this reaction using Chromobacterium violaceum ω-transaminase confirmed the ping-pong bi-bi enzyme mechanism composed of two half-reactions, shown on Figure 1 [18].

Figure 1 The reaction scheme of the transamination catalyzed by N-SBM-ATA-wt with cofactor PLP (E-PLP), where (S)-α-MBA and PYR are reversibly converted to ACP and L-ALA through the ping-pong bi bi catalytic cycle composed of two half-reactions including reversible conversion of PLP to PMP [18]. Denominations in parentheses depict the species terminology used in the model.

4

The reaction starts with the enzyme in its resting state (E-PLP), i.e. internal aldimine form, where the cofactor PLP is bound to an active-site lysine residue by means of a covalent Schiff-base linkage. Upon binding of an amine-containing substrate ((S)-α-MBA), a transaldimination reaction releases the lysine residue, and in the process the external aldimine between the substrate and the PLP is formed. The external aldimine undergoes a proton abstraction at the α-carbon atom by the catalytic lysine. The imine of the cofactor is then reprotonated and the internal rearrangements proceed through a quinonoid carbanion, yielding a ketimine. Hydrolysis of the ketimine intermediate yields free lysine and pyridoxamine 5’-phosphate (PMP), resulting in the PMP form of the enzyme (E:PMP) as well as in generation of the corresponding ketone product (ACP). The complete catalytic cycle includes the second half-reaction, which is the reversal of these steps with another ketone, an amino acceptor (PYR). The deamination of the PMP leads to the formation of an amine (L-ALA) and concomitant conversion (i.e. regeneration) to the PLP form of the enzyme [18]. The aim of this work was to efficiently describe and predict performance of the microreactor with surface-immobilized ω-transaminase comprising ping-pong bi bi mechanism. Typical approach to describe microreactors with reactions performed at the channel walls is to include reaction in the boundary conditions of a continuum-based macroscopic mass balance equation involving convective and diffusional mass transport, while reaction kinetic parameters incorporate changed enzyme activity due to the immobilization [19,20]. Here we avoid the introduction of reactive volume into the model and describe surface reaction kinetics by using surface biocatalyst concentration. This allowed us to use the same kinetic parameters obtained preliminary in a homogeneous system also in a heterogeneous system, which has according to our best knowledge not yet being performed. Mathematical model was developed for both, batch and continuous system, and a detailed numerical analysis was performed to verify the proposed models. Additionally, the time-scale analysis was applied and used to assess the limiting mechanisms occurring in the studied continuous bioprocess conducted in the developed microreactor with heterogeneous surface enzyme kinetics. Moreover, the model was validated with experimental data obtained in two consecutively-connected microreactors with immobilized enzyme.

5

2. Material and methods 2.1. Chemicals Chemicals (S)-(-)-α-methylbenzylamine ((S)-α-MBA), sodium pyruvate (PYR), L-alanine (LALA), acetophenone (ACP), pyridoxal 5’-phosphate monohydrate (PLP), sodium phosphate dibasic dihydrate and acetonitrile were purchased from Sigma–Aldrich (St. Louis, MO, USA). Sodium dihydrogen phosphate monohydrate and sodium hydroxide solution 50% were from Merck (Darmstadt, Germany), while 96% ethanol was from Kemika (Zagreb, Croatia). Deionized water was used throughout the experiments. 2.2. Biocatalyst The crude preparation of the N-SBM-ATA-wt was kindly provided by c-LEcta GmbH (Leipzig, Germany). The enzyme used is a modified (S)-selective ω-transaminase ATA-wt, which was discovered in a biodiversity screening within the strain collection from c-LEcta. For preparation of the Zbasic2-tagged enzyme, the Zbasic2 tag was fused to the N-terminus of the ATA-wt enzyme yielding N-SBM-ATA-wt as decribed in the literature [11-13]. 2.3. Microreactors For development of microreactor with immobilized N-SBM-ATA-wt, silicon/glass meander chips were used. The chips were kindly provided by iX-factory GmbH (Dortmund, Germany) and had smooth inner walls (surface roughness <100 nm), with the internal dimensions of the microchannel of 50 µm (h) x 400 µm (w) x 1620 mm (l). For the experimental work, the chips were used enclosed within a custom-made chip holders, which provided leak-free connections with VICI® Jour perfluoroalkoxy (PFA) tubes (VICI AG International, Schenkon, Switzerland) coming from high-pressure syringe pumps (PHD 4400 Syringe Pump Series) from Harvard Apparatus (Holliston, MA, USA). 2.4. Development of the microreactor system with surface-immobilized enzyme aminotransferase N-SBM-ATA-wt was immobilized on the inner surface of untreated silicon/glass meander chip, which was previously incubated for 1 h with 50 mM sodium phosphate buffer with pH 8. The enzyme solution in the concentration of 1 mg of crude enzyme preparation per mL, prepared in 6

the same buffer, was then introduced into the microchannel and left incubated for 1 h. Microchannel was afterwards thoroughly rinsed with 50 mM sodium phosphate buffer in order to remove any unbound proteins. Enzyme immobilization was performed at in a thermoregulated room at 23°C and at least three replicates. 2.5. Determination of immobilization efficiency The efficiency of enzyme immobilization was estimated based on measurements of protein concentration in the inlet and outlet solutions and by calculating the amount of the immobilized enzyme therefrom. Protein concentrations were determined spectrophotometrically by absorbance measurements at 205 and 280 nm using spectrophotometer Shimadzu UV-2600 from Shimadzu Corp. (Kyoto, Japan). 2.6. Biotransformation process catalyzed by ω-transaminase Biotransformations of (S)-α-MBA and PYR to acetophenone ACP and L-ALA, as well as transformation of ACP and L-ALA to (S)-α-MBA and PYR catalyzed by N-SBM-ATA-wt were always carried out at 30°C in 20 mM sodium phosphate buffer with pH 8, containing 0.1 mM of cofactor PLP. 2.6.1. Batch biotransformations with free ω-transaminase Batch experiments with free N-SBM-ATA-wt were carried out in 15 mL plastic centrifuge tubes in order to prevent adsorption of enzymes onto the glass surfaces. The total volume of the reaction mixture was 10 mL, while optimal mixing at 500 rpm was facilitated by micro magnetic stir-bar mixer. The initial substrates concentrations used were in the range from 10 to 100 mM and the final crude enzyme preparation concentration in the reaction mixture varied from 0.08 up to 0.32 kg m-3. Experiments were performed in at least three replicates, where samples were taken from the reaction mixture at predetermined time intervals and were 5-fold diluted in 0.1 M sodium hydroxide solution in order to deactivate the enzyme and stop the reaction. The assay for determination of the enzyme specific activity was carried out at 30°C using 40 mM initial equimolar concentrations of (S)-α-MBA and PYR and the crude enzyme preparation in concentration of 0.16 kg m-3 [21]. The specific activity was expressed as µmoles of ACP formed per mg of crude enzyme preparation per min (µmol mg-1 min-1 ; U mg-1) over the course of time

7

of the linear increase under described conditions. Experiments were performed in at least three replicates. 2.6.2. Biotransformations in microreactor system with surface-immobilized ω-transaminase Continuous biotransformation processes in the silicon/glass microreactor with surfaceimmobilized N-SBM-ATA-wt were carried out with the inlet substrates concentrations ranging from 10 to 40 mM and with the volumetric flow rates between 2 and 32 µL min-1. The reaction mixture from the outlet of the microchannel was 1-fold diluted in 0.1 M sodium hydroxide and further analysed with in-line HPLC analysis (Figure 2), which allowed for direct following of the substrate and product concentrations.

Figure 2 Schematic presentation of experimental set-up for a continuous biotransformation process in a microreactor system with surface-immobilized enzyme.

2.7. Determination of substrate and product concentrations Substrates and products concentrations were analysed by a Shimadzu Prominence modular HPLC system with SPD-M20A UV/Vis detector from Shimadzu Corp. (Kyoto, Japan) and Gemini® 3 µm NX-C18 110 Å, LC Column 150 x 4.6 mm from Phenomenex (Torrance, CA, USA). Concentrations of (S)-α-MBA and ACP were determined with acetonitrile: H2O mixture (1:1), with pH adjusted to 11 by sodium hydroxide solution, used as a mobile phase at flow rate of 1 mL min-1. The analysis, with the sample injection volume of 1 µl, was performed at 30 °C and by detection at 260 nm. Retention times of (S)-α-MBA and ACP were 2.6 and 3.4 min, respectively. 8

Before analysis, samples from batch experiments were 5-fold diluted in 0.1 M sodium hydroxide and centrifuged at 13000 rpm (14400 x g) for 5 min. For monitoring of continuous biotransformation process in a microreactor system, in-line HPLC analysis with integrated 1-fold dilution with 0.1 M sodium hydroxide solution was used which excluded the human factor and minimize experimental error throughout the entire process. 3. Mathematical model As stated in the introduction, enzymatic transamination reaction follows the ping pong bi-bi mechanism. Reaction is reversible and composed of two half-reactions (Figure 1), in which the cofactor PLP acts alternately as an acceptor and donor of the amino group, shuttling between the aldehydic (PLP) and amino (PMP) forms [18]. Assuming that the formations of products within the enzyme complexes are not limiting steps, the reaction scheme with corresponding kinetic constants could be represented as: k

2f  → ← k2 b

k

k

4f  → ← k4 b

1f  → ← k1b

k

3f  → ← k 3b

(1)

The reaction rate and equilibrium is then defined by eight reaction constants, four of them describing reactions in the forward direction (kf) and the other four in the backward direction (kb). At steady-state conditions, mass conservation equation for the substrate A in a microreactor system with surface-immobilized biocatalyst is:
 ,  ∙

 



=     +

   



(2)

with associated initial and boundary conditions:   =  ; − 



 ,

= ;

− 



 ,

= 

(3)

where CA0 is the inlet concentration of the substrate A, DA is diffusion coefficient of the substrate A in aqueous solution, and vz is the velocity of the fluid in the z-direction. Several chemical process phenomena may play a role in the characterization of microreactor performance, and characteristic time scales obtained through the time-scale analysis indicate 9

which phenomena exhibits the largest resistance to the overall process flow. Large characteristic time constant reflects large transport phenomena resistance. For analysis of microreactor performance, the impact evaluation of convective and diffusion mass transport as well as the rate of the surface enzyme-catalyzed reaction was conducted in terms of the following time-scale constants:  = "

!

#

%$=

&'

; (





/

*+((, = -∙. ; *+((, = -∙. ;  = 0 



12

(4)

where τmrt is the mean residence time, *+((, the diffusion time in x-direction, *+((, the diffusion time in y-direction, and τr the reaction time. VR is the microreactor volume, h and w microreactor height and width, respectively, while vavg is the average fluid velocity and f volumetric flow rate. For determination of the species diffusion coefficients, empirical correlations were employed [22, 23]. Considering that the diffusion times of substrates and products are sufficiently low compared to the reaction time, the system of differential equations in a three-dimensional form could be simplified to the one-dimensional (1D) form:


 

= − ∙ 3 = − ∙ 4

5∙6 ∙

7

( 5)

where 3 is the specific microreactor wall surface area, fluxA molar flux of the substrate A, while h and w are microreactor height and width, respectively. Conservation equations for batch and continuous biotransformation process written for each reaction specie are summarized in Table 1.

10

Table 1 The species conservation equations for batch and continuous biotransformation process. CA, CB, CP and CR are bulk volumetric molar concentrations (mol m-3) of (S)-α-MBA, PYR, ACP and L-ALA, respectively, while concentrations of enzyme with a cofactor PLP or PMP and their complexes with (S)-α-MBA and PYR are defined as volumetric concentrations (kg m-3) for the batch process (CE, CF, CAE and CBF, respectively), whereas for a continuous process, surface concentrations (kg m-2) of these species, namely C’E, C’F, C’AE and C’BF are applied.

Batch biotransformation process with free

Continuous biotransformation process in a microreactor with surface-

biocatalyst *

*

immobilized biocatalyst

[mol m-3s-1]

8  = :/; ∙ < − :/( ∙  ∙ < 89 8> = :?; ∙ >@ − :?( ∙ > ∙ @ 89 8C = :5( ∙ < − :5; ∙ C ∙ @ 89 8E = :-( ∙ >@ − :-; ∙ E ∙ < 89

8 < = :/( ∙  ∙ < + :5; ∙ C ∙ @ − :/; + :5(  ∙ < 89

8 >@ = :?( ∙ > ∙ @ + :-; ∙ E ∙ < − :?; + :-(  ∙ >@ 89

8 < = :/; ∙ < −:/( ∙  ∙ < + :-( ∙ >@ − :-; ∙ E ∙ < 89

8 @ = :5( ∙ < − :5; ∙ C ∙ @ + :?; ∙ >@ − :?( ∙ > ∙ @ 89

 [mol m s ] -2 -1

− 

8′ = :/; ∙ ′< − :/( ∙ ′  ∙ ′< 8 = 

8BA = :?; ∙ ′>@ − :?( ∙ > ∙ ′@ = > 8 8BD − D = :5( ∙ ′< − :5; ∙ C ∙ ′@ = C 8

− A

− F

8BF = :-( ∙ ′>@ − :-; ∙ E ∙ ′< = E 8 8′< =0 89 8′>@ =0 89 8 ′< =0 89 8 ′@ =0 89

Assuming the steady-state conditions as in a continuously operated microreactor with the enzyme immobilized onto the microchannel surface, it may be claimed that the surface molar fluxes of substrates and products are constant and that the concentrations of enzyme (in both forms) and enzyme complexes don’t change with time. By solving the system of equations listed above, solutions for fluxA, fluxB, fluxP and fluxR as a function of CA,CB, CP and CR bulk concentrations could be obtained. The functions of fluxes (i.e. 11

fluxA (CA, CB, CP, CR)) are then used as boundary conditions of species conservation equations inside the microchannel, for each specie. 3.2. Numerical methods The finite difference method (FDM), a numerical modelling method based on the discretization of continuous forms of model equations, was used to solve the system of partial differential equations. Additionally, the Monte Carlo method was applied as a tool to obtain the best-fit kinetic parameters. Numerical modelling procedure used herein is schematically presented in Figure 3.

Figure 3 Schematic representation of numerical modelling procedure used for the estimation and determination of the enzyme kinetic parameters and surface enzyme concentration.

4. Results and discussion 4.1. Batch biotransformation process with free ω-transaminase Based on initial rate measurement in batch experiments with free N-SBM-ATA-wt, a specific activity of 1.4 U mg-1of crude enzyme preparation was evaluated, which is lower than reported for ATA-wt without the Zbasic2 tag, which was above 4 U mg-1of crude enzyme preparation [21]. Further batch experiments with free N-SBM-ATA-wt at various substrates and enzyme concentrations in the forward and backward direction were used for determination of the kinetic constants applied in equations in Table 1 for batch biotransformations. In order to find the best 12

fit, a mathematical procedure based on the least-squares fitting was applied. Kinetic parameters were obtained by simultaneously fitting experimental data for (S)-α-MBA (substrate A), presented in Figure 4, as well as for the reversible reaction by following ACP depletion (data not shown) at various initial substrates (Figure 4 a) and enzyme (Figure 4 b) concentrations. Kinetic parameters obtained are gathered in Table 2.

13

Figure 4 Comparison of model predictions with experimental data obtained in a batch reactor system with free NSBM-ATA-wt. Experiments were performed at 30 °C in 20 mM sodium phosphate buffer with pH 8, containing 0.1 mM of cofactor PLP. a) 40 mM equimolar initial substrates concentration was used and the crude enzyme preparation concentrations varied between 0.08 and 0.32 kg m-3. b) 50 mM initial substrate (S)-α-MBA and 20, 40 and 100 mM PYR concentrations were used. The crude enzyme preparation concentration was 0.16 kg m-3. The points present the average of the experimental values and the brackets standard deviations, while the lines present the model.

Table 2 Kinetic parameter values estimated from experimental data obtained in a batch biotransformation processes with free N-SBM-ATA-wt. Parameter Value

k1f

k1b

k2f

[m kg s ]

-1

[s ]

-1

[s ]

2.77·10-3

1.74·10-2

0.612

3

-1

-1

k2b 3

-1

k3f -1

k3b

k4f

[m kg s ]

-1

[s ]

-1

[s ]

[m kg-1 s-1]

2.29·10-3

1.65

7.81·10-4

3.58·10-4

3

-1

[m kg s ]

-1

1.06

k4b 3

4.2. Immobilization of N-SBM-ATA-wt in a microreactor Exploiting the advantage of high wall channel surface area in comparison to reactor volume, for development of a microreactor system with surface-immobilized enzyme, a long meander silicon/glass microchannel was used (A/V = 450 cm-1). The surface of the silicon/glass microchannel in aqueous solutions at neutral pH due to deprotonation of the silanol groups, develops a negative surface charge [11-13]. The outcome of charge complementarity is a noncovalent and stable, yet readily reversible attachment of the enzyme onto the silicon/glass microchannel wall surface [7,11-13]. On the contrary, enzyme immobilization via covalent binding provides firm and stable attachment of the enzyme to the support and avoids the shedding and leakage of enzyme, while it often lowers the retained biocatalytic activity [24,25]. The amount of immobilized enzyme was estimated based on the measurements of protein concentration in inlet and outlet crude enzyme preparation aqueous solutions incubated in the reactor for 1 h. Approximately 28 µg of the protein was immobilized, which corresponds to µg cm-2 of microreactor wall surface and 0.86 µg of protein per µL. Since the immobilization method used is highly selective and despite other proteins being present in the crude enzyme preparation, it is mainly the N-SBM-ATA-wt that becomes immobilized. The enzyme load obtained was considerably higher than previously reported for the surface-immobilized ω14

transaminases and comparable to results obtained within a packed bed reactor [26,27]. As the silicon/glass microchannel surface was smooth, deposition of the immobilized enzyme was assumed to be homogeneous, with monolayer adsorption of only the Zbasic2 part of the fusion protein on the available microchannel surface. To the extent of our knowledge, this is the first report on surface immobilization of the Zbasic2-tagged ω-transaminase. 4.3. Continuous biotransformations within microreactor system with surface-immobilized N-SBM-ATA-wt Developed microreactor system with immobilized N-SBM-ATA-wt was further used for conducting continuous biotransformation process. Biotransformations were in microreactor performed shortly after the enzyme immobilization. Several inlet substrate concentrations were employed and the process was monitored at various residence times under steady-state conditions. The reaction mixture coming from the outlet of the microreactor was instantly analyzed by means of an in-line HPLC analysis, while prior analysis, integrated dilution was applied. Results are presented in Figure 5. Assuming that specific enzyme activity was not changed due to the oriented immobilization enabling efficient access to the enzyme active sites, the surface enzyme concentration in a microreactor was determined from kinetic parameters preliminary determined in the batch process, summarized in Table 2, using the conservation balances for continuous process described in Table 1. Based on the best-fit curve at the single inlet substrates concentration, shown in Figure 5, enzyme concentration immobilized on the microreactor inner walls of the reactor was estimated to be 3.8 µg cm-2. A discrepancy with the values obtained experimentally confirmed that N-SBM-ATA-wt was mostly immobilized within the microchannel, while unbound proteins were impurities in the preparation. Since reaction rates are proportional to enzyme concentration and kinetic constants, calculated enzyme concentration could compensate for more pure preparation obtained in the microreactor as compared to the batch system. This was verified by a very good agreement between simulations from proposed 1D model (Table 1) and experimental concentration profiles of the substrate A obtained at other tested inlet substrates concentrations (Figure 5).

15

Figure 5 Comparison of model predictions with experimental data obtained from continuous biotransformation process within microreactor with surface-immobilized ω-transaminase. Experiments were performed at 30 °C with 10, 20 and 40 mM equimolar inlet substrates (S)-α-MBA and PYR concentrations in 20 mM sodium phosphate buffer with pH 8, containing 0.1 mM of cofactor PLP. The points present the average of the experimental values, the brackets standard deviations, while the lines present the model simulations. Estimated concentration of the immobilized enzyme: CE,0 = 0.038 g m-2.

Furthermore, time-scale analysis was applied to characterize the continuous biotransformation process within a microreactor with surface-immobilized enzyme. As evident from Table 3, showing characteristic times and parameters used for their estimation, the diffusional mass transport in the x- and y-direction, τdiff-x and τdiff-y, revealed substantially lower values than reaction time, τr , indicating the reaction limited process. Diffusion coefficient values, determined by using the Wilke-Chang correlation [22,23], are sufficiently high to allow us the simplification of the conservation equations to the 1D form (Eq. 5). 16

Table 3 Characteristic times used in the time-scale analysis and parameters used for their estimation

Characteristic times

 =




H"I

*+((,

=

JF 

K5 = 4 ∙ 

*+((, =  =

M5 4 ∙ 

1 :-(

Estimated value 1-16.2 min

0.63 s

40 s

21.3 min

Parameters: VR= 3.24·10-8 m3, f= 3.3-5.3·10-11 m3 s-1, h = 5·10-5 m, w = 4·10-4 m, DA = 10-9 m2 s-1, k4f = 7.81·10-4 s-1

Additionally, the numerical solution of a fully developed 3D model with appropriate boundary conditions (Eq. 2 and 3) confirmed that the overall rate of the selected enzyme-catalyzed transamination in a microreactor is at given process conditions controlled by enzyme kinetics (Figure 6).

Figure 6 Concentration profile of the substrate A consumption, plotted across the width and length of the microchannel. Inlet substrate concentrations: CA,0 = CB,0 =40 mM.

17

4.4.

Continuous

biotransformation

process

within

two

consecutively-connected

microreactors with surface-immobilized N-SBM-ATA-wt Further experimental work involved microreactor system comprised of two silicon/glass microreactors connected consecutively. After the enzyme was introduced in the reactor and the unbound material was washed out, the amount of the immobilized enzyme within the microreactor system was estimated similarly as described previously for one microreactor. The developed system was then tested for conducting continuous biotransformation process, with 40 mM equimolar inlet substrate (S)-α-MBA and PYR concentration (Figure 7). Volumetric flow rates used varied from 2 to 16 µL min-1, providing residence times between 4.05 and 32.4 min. Biotransformation process was monitored with in-line HPLC analysis under steady-state conditions, where for each residence time samples were taken and analysed in at least three repetitions.

Figure 7 Comparison of model predictions with the experimental data obtained from a continuous biotransformation process within two consecutively-connected microreactors with surface-immobilized ω-transaminase. Experiments were performed at 30 °C with 40 mM equimolar inlet substrates (S)-α-MBA and PYR concentrations in 20 mM sodium phosphate buffer with pH 8, containing 0.1 mM of cofactor PLP. The points present the average of the experimental values, the brackets standard deviations, while the line presents the model.

18

Considering good agreement between predicted results from proposed mathematical model and acquired experimental data (Figure 7), biotransformation process conducted within two consecutively-connected microreactors served for model validation.

5. Conclusions A fusion protein N-SBM-ATA-wt consisting of selected ω-transaminase ATA-wt and the positively charged Zbasic2 tag enabling unhindered accessibility of enzyme active sites was successfully immobilized on a silicon/glass microchannel surface based on ionic interactions. Continuous enzyme-catalyzed transamination of (S)-(-)-α-methylbenzylamine and pyruvate to acetophenone and L-alanine and a reversible reaction were performed at several inlet substrates and enzyme concentrations for various residence times and the process was monitored with inline HPLC analysis. The important aspect of our work was to demonstrate the usefulness of time scale analysis in the (bio)chemical microreactor engineering. Based on that, a 3D model equation with appropriate boundary conditions describing convective flow in one direction and diffusional mass flow in two directions and the kinetics based on a ping-pong bi-bi mechanism at the inner walls of microchannel was simplified to the 1D model. The comparison of all included characteristic times for a given system clearly showed that the overall process is controlled by the surface enzyme-catalyzed kinetics. While the reaction kinetic parameters were preliminarily defined in a batch system, the immobilized enzyme concentration was estimated by simple solution of the system of first order differential equations for a given input concentration of substrates and various residence times, where the least squares fitting procedure was applied. Very good agreement between predicted results from proposed 1D model and experimental data from the outlet of the microreactor at various flow rates and inlet substrates concentrations. Moreover, the model successfully predicted performance of two consecutively connected microreactors, which can be considered as a successful validation.

19

Nomenclature

3

Specific microreactor wall surface area [m ]

A

(S)-α-MBA

B

PYR

E

E-PLP

C

Species volumetric concentrations [mol m ]

C0

Species initial or inlet volumetric concentration [mol m ]

C’

Species surface concentrations [mol m ]

DA

Diffusion coefficient for A [m s ]

f

Volumetric flow rate [m s ]

F

E:PMP

fluxA, fluxB, fluxP, fluxR, fluxAE, fluxBF, fluxE, fluxF

-1

-3

-3

-2

2

3

-1

-1

Reaction species molar fluxes [mol m-2 s-1]

h

Microreactor height [m]

k2b, k4b

Kinetic constants in the backward direction [m kg s ]

k1b, k3b

Kinetic constants in the backward direction [s ]

k1f, k3f

Kinetic constants in the forward direction [m kg s ]

k2f, k4f

Kinetic constants in the forward direction [s ]

l

Microreactor length [m]

P

ACP

R

L-ALA

vavg

Average fluid velocity [m s ]

vz

Average fluid velocity in z-direction [m s ]

VR

Microreactor volume [m ]

w

Microreactor width [m]

x

Coordinate in the direction of channel depth

y

Coordinate in the direction of channel width

z

Coordinate in the direction of channel length

3

-1 -1

-1

3

-1 -1

-1

-1

-1

3

20

Greek letters τdiff-x

diffusion time in x-direction

τdiff-y

diffusion time in y-direction

τmrt

mean residence time

τr

reaction time

ACP

Acetophenone

E-PLP

Enzyme with pyridoxal 5’-phosphate

E:PMP

Enzyme with pyridoxamine 5’-phosphate

L-ALA

L-Alanine

PFA

Perfluoroalkoxy

PLP

Pyridoxal 5’-phosphate

PMP

Pyridoxamine 5’-phosphate

PYR

Pyruvate

(S)-α-MBA

(S)-(-)-α-methylbenzylamine

Abbreviations

Acknowledgements The financial support of the Ministry of Education, Science and Sport of the Republic of Slovenia through Grant P2–0191 along with the financial support by the EC FP7 Project BIOINTENSE (Grant Agreement No. 312148) and COST Action CM1303 Systems Biocatalysis is gratefully acknowledged. The authors would like to thank c-LEcta GmbH (Leipzig, Germany) and iXfactory GmbH (Dortmund, Germany) for providing biocatalyst and microreactors, respectively.

21

References [1]

R.L. Hartman, J. P. McMullen, K. F. Jensen, Deciding Whether To Go with the Flow: Evaluating the Merits of Flow Reactors for Synthesis, Angew. Chem. Int. Ed. 50 (2011) 7502 – 7519

[2]

T. Wirth (Ed.), Microreactors in Organic Chemistry and Catalysis, second, completely revised and enlarged edition, Wiley-VCH, Weinheim, 2013.

[3]

R. Munirathinam, J. Huskens, W. Verboom, Supported Catalysis in Continuous-Flow Microreactors, Adv. Synth. Catal. 357 (2015) 1093–1123

[4]

P. Žnidaršič-Plazl, Enzymatic microreactors utilizing non-aqueous media, Chim. Oggi/Chem. Today. 32 (2014) 54–60.

[5]

R. Wohlgemuth, I. Plazl, P. Žnidaršič-Plazl, K. V. Gernaey, J.M. Woodley, Microscale technology and biocatalytic processes: Opportunities and challenges for synthesis, Trends Biotechnol. 33 (2015) 302–314.

[6]

Asanomi Y., Yamaguchi H., Miyazaki M., Maeda H., Enzyme immobilized microfluidic process reactors, Molecules, 16 (2011) 6041-6059

[7]

J.M. Bolivar, B. Nidetzky, Smart enzyme immobilization in microstructured reactors, Chim. Oggi/Chemistry Today. 31 (2013) 50–54.

[8]

G. Stojkovič, M. Krivec, A. Vesel, M. Marinšek, P. Žnidaršič-Plazl, Surface cell immobilization within perfluoroalkoxy microchannels. Appl. Surf. Sci. 320 (2014) 810–817

[9]

L. Hajba, A. Guttman, Continuous-flow biochemical reactors: Biocatalysis, bioconversion, and bioanalytical applications utilizing immobilized microfluidic enzyme reactors, J. Flow Chem. 6 (2016) 8–12.

[10]

E. Laurenti, A. dos Santos Vianna Jr., Enzymatic microreactors in biocatalysis: history, features, and future perspectives, Biocatalysis. 1 (2016) 148–165.

[11]

Wiesbauer, J.M. Bolivar, M. Mueller, M. Schiller, B. Nidetzky, Oriented Immobilization of Enzymes Made Fit for Applied Biocatalysis: Non-Covalent Attachment to Anionic Supports using Zbasic2 Module, ChemCatChem. 3 (2011) 1299–1303.

[12]

J.M. Bolivar, B. Nidetzky, Positively charged mini-protein Zbasic2 as a highly efficient silica binding module: opportunities for enzyme immobilization on unmodified silica supports, Langmuir. 28 (2012) 10040–10049. 22

[13]

J.M. Bolivar, M.A. Tribulato, Z. Petrasek, B. Nidetzky, Let the substrate flow, not the enzyme: Practical immobilization of D-amino acid oxidase in a glass microreactor for effective biocatalytic conversions, Biotechnol. Bioeng. 9999 (2016) 1–27.

[14]

D.Valikhani, J. M. Bolivar, M.Pfeiffer, B.Nidetzky, Multivalency Effects on the Immobilization of Sucrose Phosphorylase in Flow Microchannels and Their Use in the Development of a HighPerformance Biocatalytic Microreactor. ChemCatChem. (2016) doi:10.1002/cctc.201601019

[15]

M. Höhne, U.T. Bornscheuer, Biocatalytic routes to optically active amines, ChemCatChem. 1 (2009) 42–51.

[16]

D. Koszelewski, K. Tauber, K. Faber, W. Kroutil, ω-transaminases for the synthesis of nonracemic alpha-chiral primary amines. Trends Biotechnol. 28 (2010) 324–332.

[17]

L.H. Andrade, W. Kroutil, T.F. Jamison, Continuous flow synthesis of chiral amines in organic solvents: immobilization of E. coli cells containing both ω-transaminase and PLP, Org. Lett. 16 (2014) 6092–6095.

[18]

K. Engelmark Cassimjee, B. Manta, F. Himo, A quantum chemical study of the ω-transaminase reaction mechanism, Org. Biomol. Chem., 13 (2015) 8453-–8464

[19]

G. Stojkovič, I. Plazl, P. Žnidaršič-Plazl, L-Malic acid production within a microreactor with surface immobilised fumarase, Microfluid. Nanofluidics. 10 (2011) 627–635.

[20]

T. Van Daele, D. Fernandes del Pozo, D. Van Hauwermeiren, K. V. Gernaey, R. Wohlgemuth, I. Nopens, A generic model-based methodology for quantification of mass transfer limitations in microreactors, Chem. Eng. J. 300 (2016) 193–208.

[21]

T. Börner, C. Grey, P. Adlercreutz, Generic HPLC platform for automated enzyme reaction monitoring: Advancing the assay toolbox for transaminases and other PLP-dependent enzymes, Biotechnol. J. (2016).

[22]

C.R. Wilke, P. Chang, Correlation of diffusion coefficients in dilute solutions, AIChE J. 1 (1955) 264–270.

[23]

N. Miložič, M. Lubej, U. Novak, P. Žnidaršič-Plazl, I. Plazl, Evaluation of Diffusion Coefficient Determination using a Microfluidic Device, Chem. Biochem. Eng. Q. J. 28 (2014) 215–223.

[24]

L.-J. Zhang, Dong-Hao; Yuwen, Li-Xia; Peng, Parameters Affecting the Performance of Immobilized Enzyme, J. Chem. 2013 (2013) 1–7

23

[25]

W. Neto, M. Schürmann, L. Panella, A. Vogel, J.M. Woodley, Immobilisation of omegatransaminase for industrial application: Screening and characterisation of commercial ready to use enzyme carriers, J. Mol. Catal. B Enzym. 117 (2015) 54–61.

[26]

A. Abdul Halim, N. Szita, F. Baganz, Characterization and multi-step transketolase-ωtransaminase bioconversions in an immobilized enzyme microreactor (IEMR) with packed tube, J. Biotechnol. 168 (2013) 567–575.

[27]

S. Matosevic, G.J. Lye, F. Baganz, Immobilised enzyme microreactor for screening of multi-step bioconversions: characterisation of a de novo transketolase-ω-transaminase pathway to synthesise chiral amino alcohols., J. Biotechnol. 155 (2011) 320–9.

24