miniaturizing biocatalysis: enzyme-catalyzed reactions in an aqueous/organic segmented flow...

DOI: 10.1002/adsc.201100394

Miniaturizing Biocatalysis: Enzyme-Catalyzed Reactions in anAqueous/Organic Segmented Flow Capillary Microreactor

Rohan Karande,a Andreas Schmid,a,* and Katja Buehlera

a Laboratory of Chemical Biotechnology, Department of Biochemical and Chemical Engineering, Technische Universit�tDortmund, Emil-Figge-Strasse 66, 44227 Dortmund, GermanyFax: (+49)-231-755-7382; phone: (+ 49)-231-755-7380; e-mail : [email protected]

Received: May 16, 2011; Revised: August 5, 2011; Published online: September 5, 2011

Supporting information for this article is available on the WWW underhttp://dx.doi.org/10.1002/adcs.201100394.

Abstract: A segmented flow capillary microreactorwas used to perform the enzyme-catalyzed conver-sion of 1-heptaldehyde to 1-heptanol in a two liquid-liquid phase system. These reactor formats are estab-lished for chemical reactions but so far data describ-ing the relevant system parameters for enzymaticcatalysis are lacking. This work now addresses theimpact of important parameters such as capillary di-ameter, flow velocity, phase ratio, and enzyme aswell as substrate concentration on the performanceof the enzymatic reaction under segmented flow con-ditions. All key parameters governing reaction per-formance have been correlated in a novel operation-al window for an easy assessment of the varioussystem constraints. Such systems are characterizedby high productivities and easy phase separation fa-cilitating downstream processing. This work under-

scores the importance of segmented flow systems asa promising tool to perform multiphasic enzymaticcatalysis.Abbreviations/Nomenclature: Da: Damkçhlernumber; kcat : turnover number (s�1); eo: enzyme con-centration (mM); f : phase ratio; kL: mass transfercoefficient (m s�1); a: interfacial area per volume(m�1); CAe: equilibrium substrate concentration inthe aqueous phase (mM); CAL: substrate concentra-tion in the bulk aqueous phase (mM); rA: rate of re-action in the aqueous phase; mA: substrate masstransfer into the aqueous phase; STY: space timeyield.

Keywords: alcohol dehydrogenase; enzyme catalysis;microreactors; segmented flow; two-phase system

Introduction

The main objective of bioreactor design is to ensurehigh productivities and catalyst stabilities at low man-ufacturing and maintenance costs. In the last decades,considerable efforts have been made in this respecteither by design of new reactors or by optimization ofestablished systems. Especially, the concept of minia-turizing reaction formats has gained interest as pro-ductivities of chemical reactions are enhanced byorders of magnitude.[1] Accordingly, microreactorshave been successfully implemented in the chemicalindustry for the production of fine chemicals.[2]

Segmented flow capillary microreactors are espe-cially suited for reactions comprising multiple liquid-liquid phases due to the improved mass transfer ratescompared to the conventional batch systems.[3] A seg-mented flow is formed by the contact of immisciblefluids through a T- or Y-shaped mixer generating al-

ternate fluid segments.[4] The high mass transfer ratesare obtained from the enhanced surface area tovolume ratio between the segments, while the shear-ing motion generates recirculations or vortex motionsand enhances mixing within the segments.[5] Addition-ally, the power input necessary to obtain high masstransfer coefficients in the segmented flow is lowcompared to that of the conventional systems.[3]

Moreover, these types of microreactors are simple indesign, easy to construct and cheaper than conven-tional reactors.

Although well established for chemical reactions,segmented flow capillary microreactors are so far notused for enzymatic catalysis. One has to keep in mindthat only fast reactions limited by mass and heattransfer profit from the advantages of miniaturization.Contrary to chemical reactions, biocatalytic conver-sions are relatively slow and are often not hamperedby mass or heat transfer problems. Therefore just a

Adv. Synth. Catal. 2011, 353, 2511 – 2521 � 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim 2511

FULL PAPERS

limited number of multiphase enzymatic reactionswhich are depending on sufficient inter-phase trans-port will benefit from a microscale reaction system.[6]

However, high surface to volume ratios between thesegments might lead to biocatalyst inactivation. In aprevious study, the enzyme inactivation in the seg-mented flow was investigated and successfully circum-vented by addition of surfactant to the aqueousphase.[7] This work presents our continuing effort forthe development of segmented flow enzymatic micro-reactors (SFER).

The model enzyme used during this study was thethermophilic alcohol dehydrogenase (TADH), be-cause of its broad substrate spectrum and enantiospe-cificity,[8] which makes it a valuable catalyst for organ-ic synthesis. An additional water immiscible organicphase is necessary to overcome the low solubility ofthe non-polar substrates and their toxic impact on en-zymes.[9] It serves as a substrate reservoir and a prod-uct sink, and the efficient mass transfer of the differ-ent compounds between the organic storage and theaqueous reaction phase constrains the performance ofthe TADH reaction.

In the course of SFER development the impact ofthe mass transfer rates on the TADH reaction kinet-ics were pre-evaluated by using the Damkçhlernumber. Thus, the reduction of 1-heptaldeyde to 1-heptanol was chosen as a model reaction. Importantparameters governing the interplay between reactionrates and mass transfer rates were investigated and vi-sualized in an operational window for a systematicoptimization of the SFER performance (Figure 6 andFigure 7). An average productivity of10.4 gproduct Lorg

�1 h�1 (90 mM h�1) was obtained in thesegmented flow system using a 0.5 mm inner diametercapillary. Phase separation, as well as product isola-tion, was straightforward.

Results

Which Enzymatic Reaction Profits from aMicroreactor System?

The true benefit of micro-scale systems are excellentmass transfer rates accomplished by the enhanced sur-face area to volume ratio.[10] Therefore, reactions lim-ited by mass transfer rates benefit from the substan-tial transfer potential of micro-systems. Hence, know-ing the rate limits of the mass transfer and the enzy-matic reaction can be helpful to judge in suitabilityfor the micro-systems. To evaluate the TADH reac-tion rate limit, the substrate mass transfer resistancewas assumed to solely reside in the aqueous phase(based on the high substrate partition coefficients,Korg./aq.)

[11] .The substrate mass transfer rate throughthe organic phase into the aqueous phase is thus:

where kL is the mass transfer coefficient, a is the inter-facial area per volume, CAe is the equilibrium sub-strate concentration in the aqueous phase, CAL is theaqueous phase substrate concentration. For theenzyme reactions following Michaelis–Menten kinet-ics, the reaction rate is:

where Km is the Michelis constant. At steady state thesubstrate mass transfer rate is in balance with the re-action rate (mA = rA), hence Eq. (1) and Eq. (2) areequated as follows:

where f is the phase volume ratio (Vaq/Vorg). Thisequation is rearranged in terms of dimensionlessquantities by putting:

The right side term can be reduced by introducingthe dimensionless variable Damkçhler numbersecond (Da). Therefore Eq. (4) now becomes:

and

where, Vmax =kcat � e0, kcat is the turnover number ande0 is the enzyme concentration. This number parame-terizes the reaction kinetics and mass transfer rates,similar to the Hatta number or Thiele modulus.[12] Ahigh Damkçhler number represents a mass transfercontrolled regime where the reaction rate is fasterthan the mass transfer rate, whereas a low Damkçhlernumber indicates a reaction rate controlled regimesince the reaction rate is much slower than the masstransfer.

2512 asc.wiley-vch.de � 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim Adv. Synth. Catal. 2011, 353, 2511 – 2521

FULL PAPERS Rohan Karande et al.

http://asc.wiley-vch.de

Table 1 illustrates two reactions catalyzed byTADH with a different turnover number: reductionof rac-3-methylcyclohexanone to (1S,3S)-3-methylcy-clohexanol, and reduction of 1-heptaldeyde to 1-hep-tanol (Scheme 1). As the reaction rate equals themass transfer rate at a Da number of 1 (Da= 1), itwas possible to estimate the mass transfer coefficient(kLa) necessary in the reactor system to normalize thereaction speed. The respective mass transfer coeffi-cient (kLa) differs for both reactions: for the conver-sion of 1-heptaldehyde a kLa of 0.25 s�1 is obtained,whereas the conversion of 3-methylcyclohexanonegives a kLa of 0.0012 s�1 (Table 1).

This suggests a 200-fold higher mass transfer coeffi-cient for the conversion of 1-heptaldehyde as com-pared to the conversion of 3-methylcyclohexanone. Itindicates that the conversion of 1-heptaldehyde ismass transfer limited, if the reactor is not operated ata mass transfer coefficient equal to or above 0.25 s�1.Kashid and co-workers investigated mass transfer invarious aqueous organic segmented flow systems andcompared those to traditional extraction formats(e.g., agitated vessels). The reported mass transfer co-efficients for the segmented flow (kLa) have been inthe range of 0.1 to 2 s�1,[3] which were higher thenthose obtained for the conventional units. As our pre-dicted mass transfer coefficient based on the Da

number (Da= 1) for the conversion of 1-heptaldehydelies within this range (Table 1, kLa= 0.25 s�1), the seg-mented flow system is highly attractive to performthis enzymatic transformation. Based on this pre-eval-uation, 1-heptaldehyde was chosen as a model sub-strate for further investigation of the SFER.

Which Parameters Influence the Enzymatic ReactionPerformance in a SFER?

Several authors have identified important parametersaffecting mass transfer performance in segmentedflow reactors, such as phase ratio, capillary diameter,and flow velocity.[3,5,13] These parameters were studiedin detail to understand their impact on TADH perfor-mance.

Trade-Off between the Enzyme Concentration andthe Capillary Diameter

The impact of enzyme concentration on the reactionperformance in the segmented flow reactor was inves-tigated in 0.5 mm, 1 mm, and 2.15 mm inner diameterPTFE (polytetrafluoroethylene) capillaries. For allstudied capillaries, the STY increased linearly withthe enzyme concentration, until it reached a maxi-mum before levelling off (Figure 1). This points to a

Table 1. Pre-evaluation of the reactions based on the Dam-kçhler number.[a]

Reaction 1 Reaction 2

kcat ~10 s�1 ~100 s�1

Enzyme 0.025 mM 0.025 mMSubstrate[b] 200 mM 10 mMDamkçhler number 1 1Mass transfer coefficient (kLa) 0.00125 s�1 0.25 s�1

[a] Mass transfer coefficients were determined by assumingDamkçhler number=1 and phase ratio =1.

[b] Solubility limit of the substrate in the aqueous phase.

Scheme 1. Reactions catalyzed by TADH and used for pre-evaluation presented in Table 1.

Figure 1. Influence of enzyme concentration on the spacetime yield (STY) using capillaries of varying inner diameter(i.d.). Flow rate and capillary length were adapted toachieve 30 min residence time for each capillary. Aqueousand organic phase flows were kept equal and constantduring the experiment. 2.15 mm i.d.: length 4.9 m, total flow0.58 mL min�1 (0.26 cm s�1); 1 mm i.d.: length 11 m, totalflow 0.29 mL min�1 (0.61 cm s�1); 0.5 mm i.d: length 25 m,total flow 0.165 mL min�1 (1.40 cm s�1). For more informa-tion on enzyme stability (TADH/FDH) see Supporting In-formation.

Adv. Synth. Catal. 2011, 353, 2511 – 2521 � 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim asc.wiley-vch.de 2513

Miniaturizing Biocatalysis: Enzyme-Catalyzed Reactions


transition from a reaction rate limited regime to amass transfer limited regime. For 2.15 mm PTFE ca-pillary, no significant increase in the reaction rate wasseen at higher enzyme concentrations (2 U FDH mL�1

and 4 U TADH mL�1). However, the reaction ratewas further increased by reducing the capillary innerdiameter to 1 mm or 0.5 mm. The maximum STY of90 mM h�1 was achieved for the smallest capillary (i.d.0.5 mm), which illustrates an enhancement of masstransfer due to increase in the length of segments(Figure S1 in the Supporting Information) and flowvelocity with decreasing capillary diameter. For theefficient usage of enzymes, it is thus essential to deter-mine the optimal enzyme concentration for each ca-pillary unit.

The Impact of Flow Velocity on the ProductFormation Rate

Segmented flow forms internal circulations or forcedvortex within the segments which are developed fromthe shear forces implied by the capillary wall across asegment interface.[5] These circulations improvemixing within the segments and are dependent on theflow velocity, length of segments and inner diameterof the tubing. Out of these parameters, flow velocitywas varied in order to investigate the effect of higherinternal circulations on the reaction performancewhile other parameters were kept constant. At highflow velocities and short residence times (1.86 cm s�1,10 min), the space time yield was high (Figure 2).Lowering the flow velocity to 0.91 cm s�1 or0.61 cm s�1 (residence times of 20 or 30 min), in-creased the total product concentration at the ex-

pense of STY. This experiment shows the effects offlow velocity on the STY at varying residence times.

Impact of the Phase Ratio on Product Concentration

In segmented flow reactors, the ratio of aqueous toorganic phase (phase ratio) alters the segment sizesand is an important parameter influencing the masstransfer rates.[5] The phase ratio was varied by chang-ing the individual phase flow rate, while keeping thetotal flow rates constant to obtain a residence time of20 min (Figure 3). The product concentration in theorganic phase increased up to a phase ratio of 1.5 andthen remained constant. However, the maximalamount of product considering both phases wasformed at a phase ratio of 1. Though, the optimalphase ratio of 1.0 was based on the total amount ofproduct formed, a trade-off between the product con-centration and the amount is necessary to meet down-stream separation cost.

Trade-Off between Substrate Concentration andCapillary Diameter

The effect of varying substrate concentrations on theenzymatic reaction performance in the SFER was an-alyzed in 0.5, 1, and 2.15 mm inner diameter PTFEcapillaries. The maximum reaction rate was obtainedin the smallest capillary (0.5 mm i.d.) at a substrateconcentration of 100 mM 1-heptaldehyde in the or-ganic phase (Figure 4). A further increase of the sub-strate concentration was not beneficial as TADH was

Figure 2. Impact of flow velocity on the product formation.TADH= 8 UmL�1; 1 mm i.d. capillary 11 m long; total flowrate varied from 0.19 to 0.88 mL min�1.

Figure 3. The influence of phase ratio on the product forma-tion. TADH=4 U mL�1, 1 mm i.d. capillary 11 m long;aqueous flow 0.15 to 0.3 mL min�1 and organic flow 0.3 to0.15 mL min�1; residence time 20 min.




inhibited by substrate concentrations above 1 mM 1-heptaldehyde in the aqueous phase (100 mM in theorganic phase, partition coefficient Korg./aq. = 100). Incapillaries with a larger inner diameter a lower con-version rate was obtained and the optimal substrateconcentration shifted from 100 mM to 150 mM. Inter-estingly, similar amounts of product were convertedat a substrate concentration of 500 mM, irrespectiveof the capillary inner diameter. This indicates thatsmaller capillaries are only beneficial for the masstransfer rate at lower substrate concentrations.

Based on the above findings, different reactor set-ups were considered for maximizing product forma-tion. Utilizing small sized capillaries is not recom-mendable, because they have to be very long to ach-ieve a sufficient residence time for the reaction. Toachieve a residence time of 210 min, one would needa capillary of 175 m (0.5 mm i.d.), which leads toproblems with high backpressure and is thus not at-tractive. Simply changing the flow rate to prolong theresidence time is no option, as this will in turn changethe operating region of the system and lower theoverall reaction performance (Figure 2). Instead, dif-ferent sized capillaries have been combined(Figure 5), which reduced the overall capillary lengthto 41 m. The reaction was started in a 2.15 mm i.d. ca-pillary with the organic segments containing 300 mM1-heptaldehyde. With ongoing conversion to 1-hepta-nol the inner capillary diameter was reduced to 1 mmand finally to 0.5 mm, because at lower substrate con-centrations smaller capillaries support higher reaction

rates. In this combined system 85 mM of the product1-heptanol could be gained in 105 min (Table 2). Withthe decrease in capillary inner diameter the flow ve-locity and the surface area to volume ratio increases,which contributes to the enhancement of the masstransfer rates and results in a higher product concen-tration. However, for the longer residence time(210 min) single 2.15 capillary turned out to beneficialas TADH inactivation was more pronounced in thecoupled capillaries.

Figure 4. Influence of substrate concentration on the prod-uct formation and 1-heptaldehyde conversion using capilla-ries of varying inner diameter (i.d.). TADH=6 U mL�1; resi-dence time 30 min.; 2.15 mm i.d.: length 4.9 m, total flow0.58 mL min�1; 1 mm i.d.: length 11 m, total flow0.29 mL min�1; 0.5 mm i.d.: length 25 m, total flow165 mL min�1.

Figure 5. Experimental data shown in Figure 4 used to elab-orate the coupled capillaries SFER set-up. SFER combining2.15 mm, 1 mm and 0.5 mm capillaries in series.

Table 2. TADH-catalyzed conversion of 1-heptaldehyde indifferent SFER configurations.[a]

System Residencetime [min]

Product[mMorg.]

STY[mMorg h�1]

TADH[d]

cSFER[b] 105 85 48 50SFER[c] 105 64 36 88cSFER[b] 210 112 32 36SFER[c] 210 137 39 63

[a] General conditions: total flow105min =0.3 mL min�1; totalflow210min =0.165 mL min�1. Aqueous phase : NADH=0.5 mM, TADH= 15 UmL�1, ammonium formate=1 M,Tween 20=0.11 mg mL�1. Organic phase : 300 mM 1-hep-taldehyde.

[b] cSFER: coupled SFER combining different sized capilla-ries: total capillary length 40.9 m (2.15 mm i.d=4.9 m,1 mm i.d.= 11 m, 0.5 mm i.d.=25 m).

[c] SFER: standard set-up. Total capillary length 9 m(2.15 mm i.d. only).

[d] Recovered TADH activity; FDH activity was above 89%in all cases.




Product Isolation

Product isolation is crucial for the overall processdesign and very often neglected. Two phase systemstend to form stable emulsion in traditional stirredtank systems, which severely complicates downstreamprocessing and the recycling of the organic phase dueto difficulties in phase separation.[14] Applying twophase reaction systems to a SFER circumvents thisproblem, as no emulsions are formed and the phasesmay be easily separated based on their density differ-ence. The maximum product concentration obtainedin the organic phase of our SFER corresponds to 1.5to 2 wt% of the total organic phase. For the practicalapplication of 2-phase enzymatic reactions one of thekey steps is to isolate the product and to recycle theorganic phase. In preliminary experiments, fractionalbatch distillation of the organic phase at 7 mbar and180 8C recovered 410 mg of 1-heptanol (purity 97–98% based on GC analysis) as a clear liquid from thedistillate. The bottom phase of distillation containedmainly hexadecane (99.45% based on GC analysis),which could be recycled in the 2-phasic enzymatic re-actor. The isolated product yield was 40%, and fur-ther optimization to enhance the product yield will beof special interest in future works.

Discussion

The lacking information about the suitability of agiven reaction for SFER systems hampers their broadapplication in biocatalysis. This was demonstrated bystudies investigating the application of lyases and li-pases in microscale systems.[15] These reactions wereobviously not mass transfer limited and thus the out-come of the studies was that there was no benefitfrom applying those catalysts to a microscale system.The performance was comparable to a lab-scale batchsystem. The work presented here elaborates a simpleapproach to evaluate enzymatic reactions for applica-tion in a microreactor system. By analyzing the Dam-kçhler number [Eq. (6)], it is possible to predictwhether the enzymatic reaction would profit from thetransfer potential of the microsystem. Pompano andco-workers have used the Damkçhler number to pre-dict the outcome of autocatalytic reactions in seg-mented flow systems.[16]

Tuning Reaction, Fluidic and System Parameters toMaximize Product Formation Rates

In microreactor technology, the scale-up from labora-tory to production level is achieved by the numberingup or scale out technique.[17] Therefore, optimizingthe operational parameters for a single unit is the

most essential step to develop a large-scale system op-erating at maximal output. In this work, various pa-rameters governing the SFER performance were in-vestigated. For the optimal operation of SFER, it isnot only important to study the impact of single pa-rameters, but also to evaluate how these factors areinterconnected with each other and influence theoverall performance of the SFER. There are two ap-proaches to recognize the collective process parame-ters in two phase systems with an aim to understandand optimize a bioprocess. A mathematical approachrequiring numerical computations,[18] or graphical rep-resentation referred to as windows of operation.[19]

The later approach has been demonstrated and ap-plied to various examples in system and process de-velopment to identify key limitations using qualitativeor quantitative windows of operation.[19] In this study,the window of operation approach was extended to aternary diagram for the conceptual design of theSFER (Figure 6). The three variables that form theaxes of the ternary diagram are the reaction efficiencyor effectiveness factor (E), the Damkçhler number(Da) and the productivity (P).

The effectiveness factor is defined as the ratio ofthe actual reaction rate in the presence of mass trans-fer limitation to the maximum reaction rate possible(no mass transfer limitation). The second variable, theDamkçhler number [Eq. (6)], couples all of the im-portant parameters in the SFER and is used in the

Figure 6. Window of operation for qualitative representationof the Damkçhler number (Da) and effectiveness factor (E)plotted as a function of productivity (P). Each edge of thetriangle represents 100% of the respective variable, and thepercentage decreases as the variable moves towards the op-posite side of the triangle, where it is 0%. Lines drawn showthe impact of the individual parameter on Da, P and E. Theshaded area indicates the feasible operational area.




ternary diagram as a measure of an individual or col-lective parameter. For example, an increase inenzyme concentration, phase ratio and capillary diam-eter will lead to a rise in the Da number, whereas theincrease in flow velocity and substrate concentrationwill decrease it. Finally, the third variable on the ter-nary diagram is the productivity, which governs thesystem performance. These variables were used toplot a qualitative (Figure 6) and quantitative [Figure 7(A) and (B)] ternary diagram correlating the findingsof our experimental study. The feasible operationalregime is indicated by the shaded region and deter-mined by six factors as shown in Figure 6.

Out of these six factors, mass transfer limitation, re-action rate limitation and low productivity regimesresult from the very low values of the variable oppo-site of the respective factor. An enzyme inactivationregime occurs at a very high effectiveness factor com-bined with a low Da number and low productivity.This sounds contradictorily at first. A high E valuemeans the enzyme is operating at Vmax, but as the pro-ductivity is low, the amount of enzyme has to be lowas well, otherwise the productivity should be higher.The low Da number reflects a high mass transfer co-efficient, which could be obtained by high velocitiesand/or a high surface area in a segmented flowsystem. As seen in Karande et al. these conditionswill lead to enzyme inactivation.[7] This has also beenshown by Woodley and co-workers in a typical stirredtank batch reactor and was evaluated in a comparableoperating window.[19a] A technical system limitationregime occurs at a very high productivity, combined

with a low effectiveness factor and a low Da number.In this situation, the system operates at very highenzyme concentrations and necessitates high masstransfer rates to obtain high productivities, which isusually not possible due to the technical system con-straints (e.g., stirring speed in batch formats and flowvelocities in the segmented flow systems).

At a very high Da number, low effectiveness factor,and low productivity the enzyme is only used ineffi-ciently. In this regime, the system is either operated ata very high enzyme concentration or phase ratio, butis strongly limited by the mass transfer rate. There-fore, most enzyme molecules are unable to perform areaction because of the low amount of substrate. Toenhance the reactor performance, it is essential tomove towards the top variable (high productivity),but concurrently keeping the balance of the variableson right and left, and to operate within the shadedregion (Figure 6).

For the quantitative representation of the ternarydiagram the maximum productivity in the SFER wasset to 150 mM h�1 by extrapolating the data obtainedin Figure 1 (for further information see Supporting In-formation, Figure S4). To achieve the P-values themeasured productivities were then divided by thismaximum of 150 mMh�1. To determine the E-values,the observed reaction rate in an aqueous-organic 2-phase system was divided by the reaction rate in thesingle phase system. These two variables [effective-ness factor (E) and productivity (P)] were used toplot the experimental sample points in the ternarydiagram.

Figure 7. Quantitative representation of the experimental data in the ternary plot for the here investigated SFER set-up. (A)Impact of the enzyme concentration in 2.15 mm, 1 mm and 0.5 mm inner diameter capillary as a function of P, Da and E.(B) Impact of substrate concentration, flow velocity and phase ratio as a function of P, Da and E. The respective regionswere plotted based on the experimental data. 1.00 corresponds to 100%. Region A=mass transfer limited regime, RegionB= inefficient use of enzymes, Region C= low productivity regime; Region D = reaction limited regime; P: productivity, Da:Damkçhler number, E: effectiveness factor.




With increasing enzyme concentration the systemmoves towards a mass transfer limited regime(Figure 1). In the ternary diagram, this shift is repre-sented by dotted lines for 2.15 mm, 1 mm and 0.5 mminner diameter capillaries [Figure 7 (A)]. At lowenzyme concentrations, the system operates at verylow productivity (Region C) irrespective of the capil-lary inner diameters. With increasing enzyme concen-tration in the 2.15 mm i.d. capillary, the system opera-tion moves into region B where the enzymes are inef-ficiently used as the system is strongly limited by thesubstrate transfer from the organic into the aqueousphase. To overcome this situation, thinner capillarieshave been used to improve the mass transfer rate andthus the productivity. The capillary inner diameter isdirectly influencing the segment size which controlsthe surface area to volume ratio and therefore themass transfer rate (Figure S1 in the Supporting Infor-mation). However, at very high enzyme concentra-tions the system performance again moves towardsthe mass transfer limited regime (Region A).

Our experimental results have shown an increase inthe STY in parallel to the flow velocity (Figure 2).The flow velocity mainly attributes to the internal cir-culations within the segments, which influences themixing and enhances the mass transfer rates.[3] This iswell mirrored in the ternary diagram where thesystem operation moves from mass transfer limitedregime (Region A) into the feasible operational area.Changing the phase ratio has an impact on both themass transfer and the reaction rate. By increasing thephase ratio, the organic segments become shorter ascompared to the aqueous segments which results in ahigher surface area of the aqueous phase and an in-crease in the substrate transfer rate. However, thechange in the aqueous segment is further compound-ed by the rise in the enzyme to substrate ratio, whichin turn enhances the reaction rate and shifts the reac-tion towards a mass transfer limited regime. This iswell reflected in the ternary diagram, where the pro-ductivity improves with the phase ratio at the cost ofeffectiveness factor. At a certain phase ratio (phaseratio 1.5, Figure 3), the system operation moves to-wards mass transfer limited regime [Region A,Figure 7 (B)] with an increasing Da number.

The substrate concentration in the organic seg-ments has to be correlated to the respective capillarydiameter (Figure 4). The product formation rate ishigher in the smaller capillaries as compared to thebigger ones at low substrate concentrations(Figure 4).

At 1-heptaldehyde concentrations above 1 mM,TADH is inhibited by the substrate in the singlephase system (data not shown). In the segmentedflow a substrate concentration of 100 mM in the or-ganic phase corresponds to 1 mM in the aqueousphase at equilibrium (partition coefficient of 100),

which is reached only in the smallest capillaries be-cause of higher mass transfer rate. Therefore theproduct formation rate is higher in the smaller capilla-ries as compared to the bigger ones at 100 mM(Figure 4). In the ternary diagram [Figure 7 (B)]; thisenhancement is represented by lines A, B and C for2.15 mm, 1 mm and 0.5 mm capillaries, respectively.Productivity and effectiveness factor are improved inthe initial phase, whereas the Da number is lowered.However, in the later stage, the observed reactionrate is reduced due to substrate inhibition for the1 mm and 0.5 mm capillaries and the system operationfollows the same line but in the reverse direction.Here the productivity and effectiveness factor are re-duced, whereas the Da number is increased. As theDa number does not account for substrate inhibitionof the enzyme further improvement to integrate thislimit is necessary.

The interconnection between the experimental pa-rameters studied and their impact on the overall per-formance of the SFER can now be visualized in theternary diagram. As this figure defines the reactor op-eration, the windows of operation may be used asguidance for strategic improvement in reactor design.As an example a new reactor set-up was developed tomaximize the product formation at higher substrateconcentration (Figure 5). This was based on the movefrom line A to B and from B to C in the ternary dia-gram [Figure 7 (B)]. Technically this was obtained byassembling the 2.15 mm, 1 mm and 0.5 mm capillariesin series and starting the conversion at a substrateconcentration of 300 mM. In this system it was possi-ble to produce 85 mM of the product 1-heptanol in105 min, which was 1.3-fold higher than the maximumreached in a SFER set-up using a 2.15 mm inner di-ameter single capillary (Table 2). One of the criticalparameters during this approach was the adjustmentof the optimal surfactant concentration. Tween 20 isnecessary to stabilize the enzyme but it also decreasesthe surface tension and thus destabilizes the segment-ed flow at a certain concentration. The amount of sur-factant added is depending on the surface area, whichchanges with the capillary inner diameter and thus isdifferent throughout the reactor if using a coupled ap-proach (Figure 5). Nevertheless, the surfactant con-centration applied stays constant and cannot be ad-justed to the different diameters. As a consequencethe enzyme becomes inactivated at some point. If thesystem is running for a longer time, the inactivationof the enzyme is significant, leading to an overalllower production rate. In this case it is better to use alarger sized capillary, where the optimal surfactantconcentration is present throughout the reactor reduc-ing the enzyme inactivation (Table 2). For short resi-dence times the coupled approach performs better.




Conclusions

The segmented flow system is a promising tool to en-hance enzymatic reactions limited in rate by masstransfer. The results presented in this work point outthe system boundaries that affect the product forma-tion rates in the SFER. The segmented flow elimi-nates emulsion formation and thus significantly re-duces work-up and of downstream phase separation.The synthesis of 1-heptanol using TADH and FDH inthe segmented flow system sets a benchmark to per-form enzymatic transformations in segmented flowmicrosystems. The presented work is the basis for im-plementing enzymatic reactions in microsystems to ascalable level using the numbering technique.

Experimental Section

Chemicals

All chemicals used in this study were purchased either fromSigma–Aldrich (Steinheim, Germany) or Carl Roth GmbH(Karlsruhe, Germany), unless indicated otherwise. All aque-ous solutions were prepared using deionized water obtainedthrough a Seralpur Pro 90 CN (Seral, Ransbach-Baumbach,Germany). The buffer composition in all experiments was100 mM Bis-Tris, pH 6.5 at 45 8C, unless stated otherwise.

Segmented Flow Enzymatic Microreactors (SFER)Set-Up

The set-up of the SFER was similar to that described previ-ously,[7] and shown in Figure 8. A two-channel peristalticpump (Ismatec REGLO, Glattbrugg, Switzerland) fittedwith 0.7 mm or 1 mm inner diameter solvent resistant tubing(Ismatec Tygon MHLL, Glattbrugg, Switzerland) was usedto pump aqueous and organic liquids. Segmented flow wasattained through a 1 mm inner diameter T-piece connector(Roland Vetter RTA-TB6, Ammerbuch, Germany), and in-troduced into 0.5 mm, 1 mm or 2.15 mm inner diameterPTFE tubing (VICI, Macherey-Nagel, Dueren, Germany).The PTFE capillaries were kept in a thermo bath to main-tain the reaction temperature (45 8C). Prior to starting thebiotransformation, the flow was stabilized for 10 min at thedesired flow rate. A standard experiment contained hexade-cane as organic phase with 100 mM 1-heptaldehyde as sub-strate, and 100 mM Bis-Tris buffer as aqueous phase con-taining 0.5 mM NADH, 500 mM ammonium formate, andTADH as well as FDH for cofactor regeneration. The ratioof TADH:FDH was always 2:1. The Tween20 concentrationvaried with the capillary diameter applied. For 2.15 and1 mm i.d. capillaries 0.11 mgmL�1, and for 0.5 mm i.d. capil-lary 0.55 mgmL�1 have been added to the aqueous phase.

The reactions were started by contacting and passing or-ganic phase containing the substrate and aqueous phase in-cluding the enzymes and cofactors through the capillaries.The residence time of the segments was controlled either byvarying the flow rates or by changing the capillary length.The segments were collected from the tube outlet in an Ep-pendorf tube for every 2 min. Aqueous phase and organic

phase were separated based on the density difference. Sub-strate consumption and product formation were analyzedusing gas chromatography (GC), while residual enzyme ac-tivity was determined via an UV spectrophotometer basedassay (see below).

Preparation of TADH

Thermophilic alcohol dehydrogenase (TADH) was purifiedfrom recombinant E. coli BL21 (DE3) pLysS pASZ2 as de-scribed in literature.[7,8] Enzyme aliquots were stored at�20 8C. A final purity of 90 to 95% (based on PAGE analy-sis) was achieved with no contaminating activities present.

Preparation and Purification of FDH

The formate dehydrogenase (FDH) C23S mutant from Can-dida boidinii was used for cofactor regeneration.[20] It waspurified from recombinant E. coli JM110 cultivated in a 3-Lconical flask with baffles using 300 mL of Terrific Broth[21]

supplemented with ampicillin (100 mg mL�1). At anOD450 nm of 0.5–0.6, cells were induced by the addition of0.5 mM IPTG and the cultivation was continued for another5 h. Subsequently, cells were harvested by centrifugation at4618 � g, 4 8C for 20 min in a Sorvall RC-5B centrifuge(Thermo Electron Corporation, Langenselbold, Germany)and stored at �20 8C. Enzyme purification was performedby resuspension of the cell pellet to 30% (w/v) in 10 mMphosphate buffer (pH 7.5 at room temperature) and passingit through a French press unit at 1050 psi for two times(Aminco SLM Instruments, Urbana, IL). Insoluble celldebris was removed by ultracentrifugation for 30 min at91,500 � g. The supernatant was loaded onto an anion ex-change XK 16/20 column filled with 24 mL of DEAEstreamline material at a flow rate of 2 mL min�1 in 10 mM

Figure 8. Schematic view of the liquid-liquid segment flowset-up. 1: aqueous phase reservoir; 2: organic phase reser-voir; 3 and 4: solvent resistant tubing; 5 and 6: two channelperistaltic pump; 7: T-shaped mixer; 8: aqueous phase seg-ments; 9: organic phase segments; 10: 0.5 mm, 1 mm or2.15 mm inner diameter PTFE capillary; 11: samples collect-ed for off line analysis; 12: water bath to control tempera-ture; 13: magnified sketch of segment showing mass transferand reaction scheme.




phosphate buffer. Elution was performed at the same flowrate by applying a linear gradient of 2 M NaCl min�1 in10 mM phosphate buffer pH 7.5. Fractions were concentrat-ed 5-fold by filtration (Centricon, 10 KDa, Millipore Corpo-ration, Schwalbach, Germany) at 3990 � g (4 8C), aliquotedin 1-mL Eppendorf tubes and stored at �20 8C. A finalpurity of 60 to 70% was reached (based on PAGE analysis)and no contaminating activities were found in the controlexperiments.

Determination of TADH and FDH Activity, andProtein Concentration

TADH and FDH activities were measured by UV absorp-tion of NADH at 340 nm using a Cary 300 Bio UV-VISspectrophotometer (Varian, Darmstadt, Germany). Theassay mixture for TADH contained 160 mM 3-methylcyclo-hexanone and 0.2 mM NADH in 100 mM Bis-Tris to a totalvolume of 200 mL, pH 6.5 at 45 8C. This mixture was thermo-statted at 45 8C and the reaction was started by addingTADH.

The assay mixture for FDH contained 200 mM ammoni-um formate and 0.5 mM NAD in 100 mM phosphate bufferpH 6.5 at 30 8C to a total volume of 200 mL. This mixturewas thermostatted at 30 8C and the reaction was started byadding FDH.

For TADH activity the decrease of NADH was followedat 340 nm for 60 sec, while for FDH the increase of theNADH signal was monitored. Enzyme activities were calcu-lated using a specific absorption coefficient of e=6220 M�1 cm�1 for NADH at 340 nm. 1 unit of enzyme activi-ty was defined as 1 mmol of NADH consumed or producedper minute.

Protein concentration was measured using the quick startBradford dye (Bio-Rad, Munich, Germany), with a standardcurve prepared using bovine serum albumin.

Product Recovery

After the enzymatic transformation, phase separation wasbased on the density difference. 1-Heptanol was purified byfractional distillation at 7 mbar at elevated temperature. Theorganic phase was heated to 180 8C and the temperature ofthe distillate increased to 65 8C. Two distillate fractions andthe final residue were collected from the still. The first distil-late fraction contained the major portion of 1-heptaldehyde(70% based on GC using a flame ionization detector),whereas the rest of the material was 1-heptanol. The secondfraction contained 1-heptanol with a purity of 97–98% asdetermined by GC-FID, with small traces of 1-heptaldehyde.The final residue contained 1-hexadecane at a purity of99.4% (GC-FID analysis) with minor amounts of 1-hepta-nol.

Sample Preparation and Analysis

Aqueous phase samples were extracted with an equalamount of ice-cold diethyl ether containing 0.2 mM of do-decane as an internal standard, and mixed in on an Eppen-dorf thermo mixer (1400 rpm, 10 8C) for 2 min to extract thecompounds into the ether phase. After quick spinning of thesamples in a Heraeus Fresco 17 Micro-centrifuge (ThermoElectron Corporation, Langenselbold, Germany), the ether

phase was separated and dried by adding sodium sulphateprior to analysis on a Focus gas chromatography (Focus GC,Thermo Electron Corporation, Dreieich, Germany) using achiral RT-ßDex-sm column (30 m� 0.25 mm � 0.25 mm;Restek GmbH, Bad Homburg, Germany). The organicphase samples were diluted 1:10 times in ice-cold diethylether, dried with sodium sulphate, and analyzed as describedabove.

Acknowledgements

We thank Prof. Martina Pohl (Forschungszentrum J�lich) forthe kind gift of plasmids containing formate dehydrogenasegenes. Furthermore, we thank Dr. Jonathan Collins for hisexpert support during vacuum distillation. This work was fi-nanced by the ZACG (Zentrum f�r Angewandte ChemischeGenomik), the European Union (EFRE) and by the Ministryof Innovation, Science, Research and Technology of NorthRhine-Westphalia.

References

[1] a) P. Watts, S. Haswell, Chem. Soc. Rev. 2005, 34, 235–246; b) F. Benito-Lopez, R. Tiggelaar, K. Salbut, J.Huskens, R. Egberink, D. Reinhoudt, H. Gardeniers,W. Verboom, Lab Chip 2007, 7, 1345–1351; c) W. Ehr-feld, V. Hessel, H. Lçwe, Microreactors, New Technolo-gy for Modern Chemistry, Wiley, New York, 2000.

[2] a) T. Wirth, Microreactors in organic synthesis and cat-alysis, Wiley-VCH, Weinheim, 2008, pp 283; b) D. Ro-berge, B. Zimmermann, F. Rainone, M. Gottsponer, M.Eyholzer, N. Kockmann, Org. Process Res. Dev. 2008,12, 905–910.

[3] M. Kashid, Y. Harshe, D. Agar, Ind. Eng. Chem. Res.2007, 46, 8420–8430.

[4] a) J. Kinkel, E. Tomlinson, Int. J. Pharm. 1980, 6, 261–275; b) B. Zheng, J. Tice, R. Ismagilov, Anal. Chem.2004, 76, 4977–4982.

[5] a) L. Nord, K. B�ckstorm, LG. Danielsson, F. Ingman,B. Karlberg, Anal. Chim. Acta 1987, 194, 221–233; b) J.Burns, C. Ramshaw, Lab Chip 2001, 1, 10–15; c) J.Burns, C. Ramshaw, Chem. Eng. Commun. 2002, 189,1611–1628.

[6] J. Bolivar, J. Wiesbauer, B. Nidetzky, Trends Biotech-nol. 2011, 29, 333–342.

[7] R. Karande, A. Schmid, K. Buehler, Langmuir 2010,26, 9152–9159.

[8] V. Hçllrigl, F. Hollmann, AC. Kleeb, K. Buehler, A.Schmid, Appl. Microbiol. Biotechnol. 2008, 81, 263–273.

[9] a) B. B�hler, I. Bollhalder, B. Hauer, B. Witholt, A.Schmid, Biotechnol. Bioeng. 2003, 81, 683–694; b) J.Woodley, M. Lilly, Chem. Eng. Sci. 1990, 45, 2391–2396.

[10] D. Janasek, J. Franzke, A. Manz, Nature 2006, 442,374–380.

[11] R. Andrew Ramelmeier, H. Blanch, Biotechnol.Bioeng. 1989, 2, 97–102.

[12] K. Samant, KM. Ng. AIChE J. 1998, 44, 2212–2228.




[13] J. Jovanovic, E. Rebrov, T. Nijhuis, V. Hessel, J. Schout-en, Ind. Eng. Chem. Res. 2010, 49, 2681–2687.

[14] R. Mathys, Dissertation, Swiss Federal Institute ofTechnology, 1997.

[15] a) W. Swarts, P. Vossenberg, M. Meerman, A. E. M.Janssen, Biotechnol. Bioeng. 2008, 99, 855–861; b) K.Koch, R. Van den Berg, P. Nieuwland, R. Wijtmans, M.Wubbolts, H. Schoemaker, F. Rutjes, J. Hest, Chem.Eng. Journ. 2008, 135S, S89-S92.

[16] R. Pompano, W. Li, R. Ismagilov, Biophys. J. 2008, 95,1531–1543.

[17] M. Kashid, A. Gupta, A. Renken, L. Kiwi-Minsker,Chem. Eng. J. 2010, 158, 233–240.

[18] N. Sayar, B. Chenb, G. Lyea, J. Woodley, Biochem.Eng. J. 2009, 47, 10–18.

[19] a) J. Woodley, N. J. Titchener-Hooker, Bioprocess Eng.1996, 14, 263–268; b) B. Chen, F. Baganz, J. Woodley,Chem. Eng. Sci. 2007, 62, 3178–3184; c) M. Hogan, J.Woodley, Chem. Eng. Sci. 2000, 55, 2001–2008.

[20] H. Slusarczyk, S. Felber, M. R. Kula, M. Pohl, Eur. J.Biochem. 2000, 267, 1280–1289.

[21] J. Sambrook, W. D. Russell, Molecular Cloning: A Lab-oratory Manual, Cold Spring Harbor Laboratory Press,New York, 2001.




miniaturizing biocatalysis: enzyme-catalyzed reactions in an aqueous/organic segmented flow...

Documents