Biocatalytic Production of Catechols Using a High Pressure Tube-in

Article pubs.acs.org/OPRD

Biocatalytic Production of Catechols Using a High Pressure Tube-inTube Segmented Flow Microreactor Bartłomiej Tomaszewski, Andreas Schmid, and Katja Buehler* Laboratory of Chemical Biotechnology, Department of Biochemical and Chemical Engineering, TU Dortmund University, Emil‑Figge Straße 66, 44227 Dortmund, Germany S Supporting Information *

ABSTRACT: This study reports the synthesis of 3-phenylcatechol at the preparative scale using a continuous segmented flow tube-in-tube reactor (TiTR). 2-Hydroxybiphenyl 3-monooxygenase (HbpA) was applied as a biocatalyst for the hydroxylation reaction, which is dependent on the substrate 2-hydroxybiphenyl, NADH, and oxygen. While the regeneration of the cofactor NADH was guaranteed by formate dehydrogenase (FDH), oxygen was supplied via the membrane surface from the outside of the reactor system. The oxygen transfer rate through the membrane of the TiTR was determined to be 24 μmol O2 min−1 mL−1 emphasizing the potential of the TiTR as promising technology for realizing gas-dependent enzymatic reactions. Residence time and total turnover number have been identified as key limiting parameters. It was possible to scale-up this system by extending the TiTR by additional residence time units. This allowed synthesis of 1 g of 3-phenylcatechol at a high space time yield of 14.5 g L−1 h−1.

■

INTRODUCTION Oxyfunctionalization of unactivated carbons is since decades a very important research topic in academia and industry. Among various methods, enzymes may be applied as potent catalysts for such reactions especially when regio-, chemo-, and/or stereoselectivity are required.1 Although the importance of enzymes has long since been recognized,2−4 the key problem with applying them is the necessity of point-by-point optimization of reaction conditions by identifying and intensifying the key parameters limiting their catalytic activity. Intensive research has been conducted in this field addressing issues like product/substrate inhibition and cofactor regeneration, thermodynamic equilibria, or ratio of enzymes in multienzymatic systems.5−7 But despite the progress achieved in understanding these interesting catalysts, the in vitro application of monooxygenases for oxyfunctionalization reactions is still rather challenging. One of the often encountered bottlenecks is the insufficient supply of oxygen for O2-dependent reactions. The most common way of supplying this substrate is to bubble the gas through the respective reaction liquid. However, if doing so, it is necessary to consider possible enzyme deactivation and gas stripping of substrate and product. One solution to this problem is a bubble-free aeration system via a membrane.8 Its applicability was demonstrated for cyclohexane monooxygenase9,10 and laccase in a conventional batch setup.11 A shortcoming of this system was the low membrane surface area to liquid volume ratio, which limited the oxygen transfer and thus the overall reaction rate. A similar concept has later been utilized for the development of a tube-in-tube reactor (TiTR), where the membrane surface area to volume ratio could be maximized. In this tubular system the gaseous substrates entered the reaction compartment over the whole length of the tubing.12 The potential of this system was recognized for multiphasic chemical reactions in flow where © XXXX American Chemical Society

different phases, e.g., liquid/gas, are separated by means of a semipermeable perfluorocarbon membrane.13−21 It has already been successfully implemented for chemical multigram-scale synthesis of the anti-inflammatory compound fanetizole.20 In comparison to plug-flow reactors, the gaseous substrates can be supplied much more effectively, providing a constant gas concentration during the entire residence time.14 So far, only one example of utilizing the TiTR for an enzymatic reaction has been reported.22 In this example the enzyme 2-hydroxybiphenyl 3-monooxygenase (HbpA) was used for catalyzing the regioselective ortho-hydroxylation of 2hydroxybiphenyl, yielding 3-phenylcatechol on a nonpreparative scale (Figure 1).22 There, HbpA was applied together with formate dehydrogenase (FDH) for the constant recycling of the cofactor NADH in a two liquid phase system comprising decanol as a carrier phase and a potassium phosphate buffer (KPi) as aqueous reaction phase for the enzymes. In the referred study the reaction conditions were optimized with respect to substrate loadings and flow rate. Even though the space time yields measured were among highest published for this type of reaction, the system suffered from low substrate conversions related to a short residence time. In addition oxygen supply was assumed to limit the productivity at atmospheric pressures. In the present contribution, this reaction was taken to the gram-scale using the TiTR concept. In the course of scaling-up, major reaction parameters like gas permeation (oxygen transfer rate), pressure, catalyst loading, and residence time were investigated, and process constraints for the gram scale system were defined. According to these findings, the existing TiTR methodology was extended, and it was possible to purify nearly Special Issue: Continuous Processes 14 Received: June 27, 2014

A

dx.doi.org/10.1021/op5002116 | Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development

Article

Figure 1. (A) Scheme of 2-hydroxybiphenyl 3-monooxygenase (HbpA) catalyzed hydroxylation of 2-hydroxybiphenyl to 3-phenylcatechol with concomitant NADH regeneration by formate dehydrogenase (FDH). (B) Schematic representation of a tube-in-tube reactor (TiTR) with an aqueous−organic two liquid segmented flow. Enz = enzyme.

Figure 2. Gas (air) transport through the Teflon AF-2400 membrane in a 1 m TiTR (0.8 mL total volume, 0.5 mL volume inside the membrane) determined by the buret method. (A) Influence of the back pressure on the final gas titer in KPi buffer (flow rate 0.8 mL·min−1). (B) Gas transport at 6.9 bar back pressure into KPi buffer and into a 1:1 decanol−KPi buffer system. Total flow rate 0.8 mL·min−1. (C) Effect of the flow rate on the amount of transported gas at 6.9 bar back pressure in a 1:1 decanol−KPi buffer system. (D) Dependence of the final gas titer in a 1:1 decanol−KPi buffer system on the residence time at 6.9 bar back pressure.

In all experiments, the liquid in the membrane upstream of the back pressure regulator was homogeneous (free of gas bubbles) to ensure that the gas collected and quantified in the buret outgassed exclusively from the liquids. Not surprisingly, the final gas titer was dependent on the back pressure applied as the back pressure determined the gas solubility in the liquids. When only KPi buffer was fed into the reactor, the highest gas titer was determined at 6.9 bar back pressure and 7 bar air pressure around the membrane (Figure 2A). Apart from the back pressure, the feed composition had a significant influence on gas titers (Figure 2B). As expected, higher gas titers were possible in a two liquid phase system comprising KPi buffer and decanol due to the higher solubility of oxygen in organic solvents.23 Also the flow rate is an important parameter influencing oxygen concentration in the liquid (Figure 2C). Higher flow rates are beneficial for oxygen transfer; however, this leads to short residence times, which may impair conversion yields of a given reaction.

1 g of pure 3-phenylcatechol by applying HbpA in a sequential TiTR setup.

■

RESULTS Factors Influencing Gas Transfer through the Teflon AF-2400 Membrane. In order to apply a TiTR for gasdependent enzymatic reactions on a preparative scale, the gas permeation through the membrane needed to be quantified and adjusted to the consumption rate for establishing the process operational boundaries. Increasing the partial pressure of oxygen either by increasing the total pressure of the gas or by using oxygen enriched air (or pure oxygen) was expected to have a positive impact on oxygen transfer. Thus, experiments have been conducted at various pressures, and the amount of O2 passing through the membrane at the respective conditions was quantified. In addition parameters influencing its transport and its effect on HbpA performance were evaluated. Results of these experiments are summarized in Figure 2. B



Article

bar)/high air (7 bar) pressure also had no further effect. In both cases the product formation rate and final product titer were comparable (Figure 3B). Finally pure oxygen was supplied instead of air, but no further change in the reaction rate could be observed. From these findings it was concluded that the reaction was not limited by the gas transport and oxygen availability; therefore, further experiments were conducted to check whether the biotransformation rate was limited by catalyst loading. Influence of the Biocatalyst Loading and Increased Residence Time on the Reaction Rate and Product Titer. From an engineering point of view, reactions should operate in a kinetically controlled regime in order to use the enzymes efficiently. To verify if the product formation rate would benefit from increased enzyme loadings, their concentration was varied. Reaction conditions were set to 6.9 bar back pressure and 7 bar of the surrounding air. HbpA and FDH initial volumetric activities in the aqueous phase were raised stepwise, while keeping the unit ratio of the enzymes at 0.81:1 FDH−HbpA. At a low enzyme loading (2 U·mL−1 HbpA), the reaction was limited by the enzyme amount (kinetically controlled) increasing the enzyme concentration further (up to 8 U· mL−1) resulted in increased product formation rate. Above 8 U· mL−1 the impact of the enzyme concentration was negligible, and applying more catalyst did not have an impact on the product formation rate, indicating that the reaction was mass transfer controlled (Figure 4A). However, for the preparative scale with a longer residence time, higher HbpA amounts were applied to account for any activity loss (Table 1). Similarly to results obtained earlier and presented in Figure 3A, the product formation rate increased with the total flow rate. Thus, it was assumed that the product formation rate was dependent on the mass transfer of the organic substrate between the phases and not on the oxygen availability or the catalyst amount applied. It is known that with increasing flow rate the mass transfer increases in two phase flow microreactors24 as a direct consequence of faster internal circulations in the segments.25,26 From the experiments above, two key conclusions could be drawn. It was possible to increase the oxygen concentration in the pressurized TiTR system, and under the given conditions the reaction in the 1 m TiTR was clearly not oxygen limited. Moreover, the flow rates had a significant impact on the reaction rate. Thus, the reactor length had to be significantly

Quantifying the gas outgassing from the liquids at different flow rates allowed to determine the final gas titer (mL gas dissolved per L of liquid), which could be reached inside the TiTR (Figure 2D). Almost full saturation was reached within 40 s. Similar findings were reported for hydrogen passing Teflon AF-2400 membranes, where almost complete saturation of a dichloromethane phase with H2 was reached within 10 s.14 Impact of Elevated Pressure in the TiTR on HbpA and FDH Stability and Activity. Little is known about the effect of elevated pressure on enzyme structure and activity. As a TiTR is a pressurized system, it was necessary to evaluate the effect of increased pressure on enzymes’ stabilities and investigate the influence of the higher oxygen concentration on the biotransformation. In terms of enzyme stability pressure did not seem to have any impact on the respective enzymes as can be deduced from the data presented in Table 1. Table 1. Influence of the back pressure in the reactor (6.9 bar at the outlet) on the residual enzymes’ activity after passing through a 1 m TiTRa initial

after passage

enzyme

(U·mL−1)

0 bar (U·mL−1)

6.9 bar (U·mL−1)

HbpA

2.3 ± 0.1 8.7 ± 0.1 21.6 ± 1.5 1.7 ± 0.1 5.6 ± 0.1 13.1 ± 0.8

1.9 ± 0.1 6.7 ± 0.4 23.3 ± 0.5 1.4 ± 0.0 5.9 ± 0.1 15.8 ± 1.6

1.9 ± 0.0 7.8 ± 0.1 24.2 ± 1.5 1.5 ± 0.0 6.7 ± 0.1 14.6 ± 0.4

FDH

Total flow rate 0.1 mL·min−1, respective residence time 8.9 min. 0 bar indicates no back pressure applied additionally (n = 2). a

The results of the abiotic experiments aiming at quantifying the oxygen permeability at different flow conditions established process operational boundaries. The oxygen concentration in the aqueous phase increased proportionally to its pressure (Henry’s law). It was expected that this would positively influence the oxygen dependent HbpA catalyzed hydroxylation reaction. Surprisingly, only a slight impact of the pressure on the reaction rate was observed, not correlating with the available oxygen concentration (Figure 3A). Varying the differential gas (oxygen) pressure across the membrane, e.g., low liquid (1.4 bar)/high air pressure (8 bar) or high liquid (6.9

Figure 3. (A) Influence of the back pressure on the product formation rate in a 1 m TiTR (0.8 mL total volume). Outer membrane pressure of the air set to 0.0 bar (atm), 1, 2, 4, and 5 bar at subsequent 0.0 bar (atm), and 1.4, 2.8, 5.2, and 6.9 bar back pressure, respectively. (B) Influence of changing differential pressure across the membrane and applying pure oxygen instead of air. The product formation rate was calculated for 1 L of reactor working volume. C



Article

Figure 4. (A) Influence of enzyme loading and flow rates on the product formation rate in a 1 m TiTR (0.8 mL total volume) run at 6.9 bar back pressure and 7 bar air pressure in a two liquid phase system. Respective residence times reached were 8.9, 4.5, 2.2, and 1.1 min for increasing flow rates. Standard reaction conditions applied as described in Experimental Section. (B) Product formation rate and final product titer in the system aerated with pure oxygen instead of air at various residence times. Flow rate: 0.8 mL·min−1 with a residence time of 2.1, 3.1, 4.1, and 5.1 min, respectively, for increasing reactor length (errors are presented as a standard deviation, n = 3). The product formation rate is given for 1 L of reactor working volume.

Figure 5. (A) Schematic representation of a TiTR used in the previous study22 on a nonpreparative scale. (B) Sequential tube-in-tube reactor used in this study on a preparative scale. Total volume 8 mL, residence time 11.5 min at a flow rate of 0.7 mL·min−1. BPR = back pressure regulator.

increased. Conducting the reaction in a TiTR composed exclusively of Teflon AF-2400 tube longer than 4 m was not possible due to the nature of the Teflon AF-2400 material. Longer tubes are fragile and more susceptible to snapping while handling. Nevertheless, to take advantage of this technology for oxygen supply, the existing TiTR was modified by connecting two TiTR modules for oxygen supply and installing residence time units behind them as depicted in Figure 5B. Thereby, the residence time could be increased significantly. Before deciding for the final setup of the sequential TiTR, the optimal residence time (hence tube length) was investigated (Figure 4B). Depending on the tube length, the conversion of 2-hydroxybiphenyl increased until it stayed nearly constant at 5.1 min, while the product formation rate decreased gradually. As a compromise between the product formation rate and the final product titer, the length of the residence time units was set to 3 m (volume 2.4 mL), resulting in a total residence time of 3.1 min per residence time unit and 11.5 min in the total system at a flow rate of 0.7 mL·min−1. This setting was the basis for an experiment on a preparative scale applying HbpA in a sequential TiTR for the synthesis of 3-phenylcatechol. Synthesis of 3-Phenylcatechol on a Preparative Scale Using a Pressurized Sequential Tube-in-Tube Reactor. The system introduced above was operated in a semicontinuous mode. During the first 7 cycles freshly prepared

enzyme solutions were used, whereas for the eighth cycle 50 mL of the recycled enzyme solution supplemented with sodium formate (4 mmol) and NAD+ (20 μmol) was used instead, to verify if the enzymes could be reused (Figure 6). Before the eighth cycle, the activity of the recycled enzymes were measured in an independent UV assay and compared to the initial ones revealing almost 90% residual HbpA activity, while FDH did not experience any activity loss (Table 2). Surprisingly conversion in the TiTR was significantly lower, when using recycled enzyme solution. This could be attributed to the composition of the recycled aqueous phase, which also contained residual organic solvent and the dissolved product. This could inhibit the HbpA in the TiTR. Activity measured in a spectrophotometric assay would not be affected thereby, as the respective samples were diluted 300-fold prior to the measurement. During the whole experiment the reactor outlet was placed in a bottle continuously flushed with nitrogen in order to prevent the hydroxylation reaction continuing outside of the reactor. After every cycle, the phases were separated, and the collected organic phase was eventually used for product recovery. Purification of the Product 3-Phenylcatechol Was Successful and 740 mg of Product Were Isolated. The sequential TiTR for hydroxylation of 2-hydroxybiphenyl made it possible to synthesize 12 mmol of product and isolate 4 D



Article

Figure 7. Hydroxylation of 2-hydroxybiphenyl to 3-phenylcatechol applying 2-hydroxybiphenyl 3-monooxygenase (HbpA) and formate dehydrogenase (FDH) on a preparative scale.

Figure 6. Synthesis of 3-phenylcatechol in a sequential TiTR reactor. Product concentrations have been determined in the sample collected at the reactor outlet during consecutive cycles and are given for a total volume (combined aqueous and organic phase). The eighth cycle was conducted with 50 mL of the recycled aqueous phase supplemented with sodium formate (4 mmol) and NAD+ (20 μmol). Total residence time in the system (8 mL) ca. 11.5 min at total flow rate of 0.7 mL· min−1. For cycle 5, 6, and 7, n = 1.

gas is supplied at the outside of that membrane. A similar wellknown system is the Vapourtec gas/liquid membrane reactor, which operates in a reverse configuration with the gas flowing inside the membrane.30 The reverse configuration provides a higher liquid/gas contact area and should consequently lead to higher liquid saturation levels at the outlet. In addition it provides better temperature control and could be a well-suited alternative to the Gastropod. Quantification of Oxygen Permeability through the Teflon AF-2400 Membrane in a TiTR. The air used initially as a gas phase for aerating the liquids was composed of 21% O2 and 78% nitrogen. Because Teflon AF-2400 is more selective for oxygen than for nitrogen, oxygen is transported faster. By comparing the fluxes of oxygen and nitrogen streams through the membrane, it was possible to determine their ratio. Under the assumptions that the driving force was equal to the initial partial pressure difference (Δpi) and that it was constant, two flux equations were possible (eqs 1 and 2).

mmol (740 mg) of pure (>99% by HPLC and H NMR) 3phenylcatechol (Figure 7). The collected sample was centrifuged in order to achieve separation of the aqueous and organic phase. From the 400 mL organic phase present, only 360 mL was removed as a precaution to not contaminate the sample by accidentally transferring precipitated proteins or aqueous liquid. This organic sample was directly loaded onto a neutral alumina column. The overall product isolation yield after all purification steps was 35%, and the product was confirmed by the H NMR (see Supporting Information). To increase this low purification yield, a couple of measures are possible. The in-line membrane based phase separation unit recently introduced by Jensen and co-workers is an interesting option in this respect, however it proved challenging for liquids with low interfacial tension (6.8 mN·m−1, like n-butyl acetate and water),27 thus limiting its operation window for the here applied water/decanol system with an interfacial tension of 2.2 mN·m−1 (see Supporting Information, Table S1). Nevertheless, it could improve the recovery of the organic phase, which contributes significantly to the low purification yield27 and shorten the overall procedure. In addition, the temperature applied for vacuum sublimation could be lowered from 120 to 40 °C as shown previously.28 This would prevent product losses by degradation but also slow down this step. As an alternative to sublimation, recrystallization of 3-phenylcatechol from n-hexane as shown before29 could be evaluated.

JO =

Permeability × ΔpO 2 δmembr

(1)

JN =

Permeability × ΔpN 2 δmembr

(2)

2

2

For a membrane thickness of δ = 0.01 mm and permeability for oxygen and nitrogen of 990 and 490 Barrer, respectively (1 Barrer = 10−10 cm3 cm (s·cm2·cm Hg)−1 and is equivalent to 3.348 × 10−19 kmol·m (m2·s·Pa)−1 in SI units), the theoretical fluxes of oxygen and nitrogen were calculated as follows: 990

■

JO =

DISCUSSION This study reports on an enzyme catalyzed hydroxylation reaction using the Gastropod TiTR for gas/liquid reactions, in which liquid(s) flow inside a semipermeable membrane, while

2

10−10 cm 3O2·cm membr s·cm 2 membr·cm Hg

0.01 cm membr

= 1.58 × 10−4

× 0.21 × 76 cm Hg

cm 3O2 s·cm 2 membr

Table 2. Activity of 2-hydroxybiphenyl 3-monooxygenase (HbpA) and formate dehydrogenase (FDH) measured in a spectrophotometric assay before and after the reaction in the sequential tube-in-tube reactor systema

a

enzyme

initial volumetric activity (U·mL−1)

final volumetric activity (U·mL−1)

residual activity (%)

HbpA FDH

10.26 ± 0.21 6.76 ± 0.14

9.16 ± 0.19 7.51 ± 0.16

89.3 111.2

Residence time 11.5 min. (Errors are presented as a standard deviation, n = 2.) E



Article

Table 3. Oxygen concentration and transport rate in liquid/liquid two-phase flow (potassium phosphate buffer−decanol at 1:1 ratio) at 30 °C and 6.9 bar back pressure nature of gas

total flow rate (mL·min−1)

O2 concentration in reactor (mM)

O2 transport ratec [μmol·(min·mLreactor)−1]

0.8 0.4 0.2 0.1 0.8

5.3 5.7 5.9 6.0 15.0

8.4

a

air

pure oxygenb

24.1

a

Volume fraction of oxygen in the air after passing through the membrane is equal to 0.35. bAssuming that molar fraction of oxygen in the air after passing through the membrane is equal to 1, and all transported gas is oxygen, it is possible to calculate approximate oxygen concentration and its transport rate. cBecause the experiment was conducted in the range close to the oxygen saturation levels, only a theoretical oxygen transport rate is given. Calculations are based on the assumption that oxygen (air) is transported steadily over the residence time and the saturation is not reached earlier.

Table 4. Data for the enzymatic conversion of the 2-hydroxybiphenyl to 3-phenylcatechol by 2-hydroxybiphenyl 3monooxygenase (HbpA) in a sequential tube-in-tube reactor in a preparative scale experiment enzymea

total enzyme (U)

residual activityb (%)

enzyme usedc (mmol)

product (mmol)

enzyme usedd (U·gproduct−1)

TTN (molprod·molenz−1)

HbpA

3591

89

0.0017

10.5

197

6217

a

Data are not given for formate dehydrogenase (FDH) as no activity loss for that enzyme was recorded (Table 2). bEnzyme was recovered after 11.5 min total residence time. cBased on the activity of 3.5 U mg−1 of pure HbpA and molecular mass of 65000 g mol−1. dActual consumption of the enzyme based on the amount of the catalyst that lost activity.

490 JN = 2

10−10 cm 3 N2·cm membr s·cm 2 membr·cm Hg

0.01 cm membr

= 2.90 × 10−4

× 0.78 × 76 cm Hg

cm 3 N2 s·cm 2 membr −4

Jtotal = JO + JN = 4.48 × 10 2

2

trend was observed for the transport of ammonia through the Teflon AF-2400 membrane where the amount of ammonia quantified in the feed was different in methanol, DME, or toluene (concentration increased with increasing dielectric constants).20 Enzyme Consumption and Strategies for Maximizing the TTN. The TTN of the here utilized HbpA was fairly low and was calculated to be approximately 6 × 103 for the preparative scale experiment, resulting in a rather huge amount of enzyme which needed to be applied (Table 4). Although enzyme-to-product ratio is 2 orders of magnitude higher than, e.g., for the industrial process of tert-leucine production (0.9 Uenzyme·gproduct−1),3 it compares well with the epoxidation of styrene derivatives in a stirred tank reactor. In that example enzyme consumption was between 200 and 300 Uenzyme· gproduct−1 with a TTN in the range of 2000−3000 molproduct· molenzyme;31 however, in the here presented work the product was synthesized faster due to a higher STY. Compared to the nonpreparative scale system, the TTN of HbpA could be increased 3-fold.22 This was realized with the sequential TiTR system in this study (see above), which allowed a more efficient usage of the enzymes and prolonged residence times. Another approach for improving the TTN that was evaluated was recycling of the biocatalysts. Despite the fact that HbpA only lost approximately 10% of its initial activity after one passage through the sequential TiTR system, as determined via an UV spectrophotometric assay, activities calculated from the product amount formed in the reactor in the recycling run (8th cycle, Figure 6) were much lower (65%), although the system was supplemented with fresh cofactors. Most probably enzyme activity in the TiTR was hampered due to residual solvents/ product accumulated in the aqueous phase and not properly separated during work up of the enzyme fraction for recycling, whereas for UV spectrophotometric measurements samples were strongly diluted. In order to further maximize the TTN, an integrated process design could be a solution. Thereby the TTN may be increased, downstream processing simplified, and the number of process unit steps reduced. A further development of the sequential TiTR would be a membrane TiTR that would allow for in situ

cm 3gas s·cm 2 membr

JN2 is the flux of nitrogen, and JO2 is the flux of oxygen. By dividing the oxygen flux by the total flux, one could calculate the ratio of oxygen in the total gas flux to be 0.35 (or 0.35 cmO23·cmtotal−3). Results are summarized in Table 3. Quantifying gas transfer through the membrane was important as it was possible to ultimately exclude oxygen limitation. It also allowed for establishing the maximum operational boundaries of the reaction system. A maximal oxygen transport rate through the Teflon AF-2400 membrane at a flow rate of 0.8 mL·min−1 was calculated to be 8.4 μmol· min−1·mLreactor−1 for air, and theoretically 24.1 μmol·min−1· mLreactor−1 for oxygen (vide infra). Compared to the reported STY of 26.6 g·L−1·h−1 (overall volumetric productivity 3.4 μmol·min−1·mLreactor−1) for the synthesis of tert-leucine in a process running at Evonik3 in an enzyme membrane reactor, the potential of the here proposed reactor setup for oxygen limited reactions becomes clear. However, this potential was not utilized for the applied reaction here, as a rate of only 1.3 μmol·min−1·mLtotal−1 (1.3 mM·min−1) was reached, leaving room for future improvements. Despite these shortcomings, the increased concentrations of dissolved oxygen in the liquid resulting from the pressurized TiTR system could be efficiently utilized in the sequential TiTR setup, and the residence time could be extended to 11.5 min, without running into an oxygen limitation. Among other factors, the feed composition and its nature are important, since higher oxygen titers in the liquid were reached for two-phase flow as opposed to single phase flow. A similar F



Article

Table 5. Overall volumetric mass transfer coefficients (kLa) in various liquid/liquid two phase segmented flow microreactorsa system

phases (product)

reaction

capillary diameter (mm)

overall mass transfer coefficient kLa (s−1)

ref

segmented flow tube-in-tube microreactor segmented flow microreactor

decanol/KPi(3phenylcatechol) hexadecane/KPi (heptaldehyde)

enzymatic

0.8

0.104

this work

enzymatic

n-butanol/aqueous (succinic acid)

nonreacting system

segmented flow microreactor

kerosene/aqueous

acid base reaction

0.042 0.025 0.009 0.12−0.31 0.07−0.15 0.02−0.09 0.5

36, 37

segmented flow (extractor)

0.5 1.0 2.15 0.5 0.75 1.0 0.38

a

24

38

Compounds given in brackets are extracted. Adapted and modified from Karande (2011).36

Table 6. Selection of reactions catalyzed by soluble oxidoreductases with enzymatic cofactor regenerationa substrate 1-phenyl-2-propanone (15 mM); continuous membrane reactor 2-oxo-4-phenyl-butyric acid (200 mM); continuous membrane reactor trimethylpyruvic acid (500 mM); repetitive batch 2-octanone (7 mM); continuous membrane reactor ethyl-4-chloro-3-oxobutanoate (100 mM); batch

(R,S)-ethyl 2-(2-oxocyclohexyl)acetate (20 mM); batch 4-methylcyclohexanone (40 mM); batch (±)-bicyclo[3.2.0]hept-2-en-6-one (1 mM); batch methylstyrene (69 mM); batch 2-hydroxybiphenyl (110 mM); batch 2-hydroxybiphenyl (200 mM); continuous

product

enzyme

Oxygen Independent Oxidoreductases 1-phenyl-propanol alcohol dehydrogenase R. erythropolis (ADH)b 2-hydroxy-4-phenyl-butyric D-lactate-dehydrogenase S. epidermidis acid (LDH)c L-leucine-NAD oxidoreductase B. tert-leucine sphaericus (LeuDH)d (S)-2-octanol carbonyl reductase from C. parapsilosis (CPCR)e (S)-4-chloro-3NADPH-dependent ketoreductase hydroxybutyrate (KRED)f Oxygen Dependent Oxidoreductases (R)-ethyl 2-(7-oxooxepan- cyclohexanol monooxygenase 2-yl)acetate (CHMO)g 5-methyloxepan-2-one cyclohexanol monooxygenase (CHMO)h various regioisomeric (BVMOs) from Pseudomonas putida lactones NCIMB 10007i methylstyrene oxide styrene monooxygenase (StyA)j 3-phenylcatechol 2-hydroxybiphenyl 3-monooxygenase (HbpA)k 3-phenylcatechol 2-hydroxybiphenyl 3-monooxygenase (HbpA)l

CV (%)

RT (h)

PV (g·L−1·h−1)

72

0.33

2.6

3

91

4.6

6.9

39

74

2

26.6

3, 40

97

1

0.9

7

100

8

20

41

100

2

0.8

10

48

10

0.4

9

100

1

0.1

42

95.3 23

11 8

0.9 0.5

31 43

15

0.2

14.5

this work

ref

Abbreviations: CV, conversion; RT, residence time; PV, productivity. bADH (0.95 U·mL−1), 50 mL scale, single phase, UF-membrane reactor in combination with hydrophobic membrane reactor. cLDH and FDH (4 U·mL−1), 220 mL scale, substrate (200 mM), NAD+ (0.2 mM), continuous enzyme membrane reactor, single phase. dLeuDH (15 U·mL−1), FDH (5 U·mL−1), substrate (500 mM), NAD+ (0.2 mM), repetitive batch reactor with UF membrane, scale not known. eCPCR (1.5 U·mL−1), FDH (3.2 U·mL−1), 100 mL scale, substrate (7 mM), NAD+ (0.5 mM), single phase. f KRED (0.9 g), GDH (0.6 g), 3 L scale, multiphase reaction triethanolamine:butyl acetate. gCHMO (0.2 U·mL−1), FDH (0.6 U·mL−1), 30 mL scale, single phase, substrate (20 mM), NADP+ (0.4 mM), bubble free aeration. hCHMO (0.16 U·mL−1), FDH (0.32 U·mL−1), 30 mL scale, single phase, substrate (40 mM), NADP+ (0.25 mM), bubble free aeration. iBVMO (1.6 U·mL−1), FDH (1.0 U·mL−1), 2 mL scale, single phase, substrate (1 mM), NADH (0.16 mM). jStyA (2 U·mL−1), formate dehydrogenase (FDH) (8 U·mL−1), 200 mL scale, 1:1 phase ratio dodecane−buffer, substrate (50 mM), NAD+ (1 mM) kHbpA (0.2 U·mL−1), FDH (0.3 U·mL−1), 200 mL scale, 4:1 phase ratio decanol−buffer, substrate (110 mM) in decanol. l HbpA (9.2 U·mL−1), FDH (7.5 U·mL−1), NAD+ (1.6 mM), 8 mL scale, 1:1 phase ratio decanol−buffer, substrate (200 mM) in decanol, residence time 11.5 min, segmented flow microreactor with bubble free aeration. a

intermediate used for synthesis of hypertension and angina drug,34 as comparison to productivity of 70 g·m−2membr·day−1 in the present work (14.5 g·L−1·h−1, where 1 L equals 5 m2membr). Mass Transfer. To determine if the mass transfer could limit the reaction rate, the overall volumetric mass transfer coefficient (kLa) was calculated and compared with other processes. The kLa for the transport of the substrate 2hydroxybiphenyl from the decanol to the aqueous phase was calculated on the basis of the product formation rate to be 0.104 s−1 (Supporting Information). This value measured for a 0.8 mm i.d. Teflon AF-2400 capillary used in this study agrees well with published kLa values measured for the extraction of succinic acid using a comparable capillary and is slightly higher than the one calculated for the enzymatic process in

separation of aqueous and organic phase. Application of membrane reactors for homogeneous single phase reactions is well-established on the industrial scale in (C)STR.32 Ultrafiltration membranes are often applied to keep enzymes in a confined reaction compartment allowing for continuous substrate addition and product removal in a stationary (batch mode) or mobile (continuous) aqueous phase. After reaction, respective enzymes are retained and can be reused in another batch. Similarly, membranes are applied for biphasic aqueous/ organic reaction systems where they prevent emulsion formation (by separating the phases), simplify product recovery, and allow for repeated use of the biocatalysts.33 A reactor of this type has been successfully applied on industrial scale with a productivity of 145 g·m−2membr·day−1 of a chiral G



Article

for application of some oxidoreductases, especially dehydrogenases. Among immobilized oxidoreductases, which do not need oxygen, a benchmarking reaction is the gas phase reaction with alcohol dehydrogenase from Lactobacillus brevis immobilized onto glass beads. A STY of 4.45 g·L−1·h−1 was reported in a reactor continuously operated up to 80 h.46 Other examples applying hydrolytic enzymes are the hydrolysis of lactose by immobilized thermophilic beta-glycosidase with an impressive space time yield for glucose of 500 g·L−1·h−1,47 or the chemoenzymatic epoxidation of 1-methylcyclohexene by an immobilized lipase Novozym 435 in a packed bed reactor, which reports space time yields of 26.9 g·L−1·h−1.48 Enzyme immobilization could be an option to improve the performance of the enzymes presented here. Various methods of enzyme immobilization in microreactors exist and have been reviewed previously.49 For the reaction described here catalyzed by HbpA and FDH, entrapment of the enzymes in a gelatin/ chitosan matrix has been evaluated. However, enzyme loading and specific activities were very poor, leading in the end to significantly lower STY as compared to the approach using soluble enzymes (data not shown). Therefore, we did not investigate this further. It is very possible that by using a different immobilization technique overall reaction performance with respect to the TTN and enzyme cost would be improved.

microreactor (Table 5). There are various parameters influencing mass transfer in the system, e.g., flow rate, capillary inner diameter, interfacial tension, and viscosity of the liquid. Therefore, kLa values measured even in the same reactor but with liquids of different properties might differ; hence, comparison between different systems is difficult. Additionally, as the kLa is calculated based on the product formation rates, the mass transfer for enzymatic systems might be underestimated, since the molecules not only need to pass the mass transfer barrier but also have to diffuse into the active site of the enzyme. To conclude, the mass transfer calculated in the present work is in the range of the previously published values for mass transfer limited reactions (Table 5); therefore, it could be the reaction rate limiting factor. In order to improve mass transfer, one could increase the flow rate; however, this would in turn reduce residence times and thereby the conversion yield and TTN. It should be also be possible to use a Teflon AF-2400 membrane with a lower internal diameter, which will decrease diffusion distance (tube i.d.) and diffusion time in the tube. By lowering the tube internal diameter, the diffusion mass transfer becomes more important than advection (mass transfer resulting from the bulk movement of the liquid), mixing occurs faster, and thus shorter tubes are required.35 In addition the surface area to volume ratio is enlarged and thus the interfacial mass transfer. Although the higher fluid velocity would increase the Reynolds number, it is to be expected that no transition into the turbulent flow would take place as Reynolds numbers are normally low in microreactor systems (Re = 15 in the current system). Comparison to Other Oxidoreductase Catalyzed Reactions. To discuss the here reported findings in a broader context, a selection of oxidoreductases is given in Table 6. Respective enzymes were grouped according to their oxygen dependency. Among the oxygen independent reactions the highest STY for this class of enzymes was reported for the production of tert-leucine. Interfacial mass transfer limitations were not relevant here as only a single phase was needed. Among the oxygen-dependent reactions the HbpA catalyzed hydroxylation of 2-hydroxybiphenyl presented in this study shows the highest STY. In general oxygen-dependent reactions perform with lower STY than oxygen-independent ones. Selection of an appropriate reactor format is therefore crucial to allow oxygen dependent reactions perform comparably. Since the example presented here shows a STY matching that of oxygen independent reactions, it underlines the applicability of the TiTR technology and its potential for reaction intensification of fast and oxygen dependent reactions catalyzed by oxidoreductases. As this study reports on the use of soluble enzymes in a microreactor system, only examples using soluble enzymes have been included in the comparison presented above (Table 6). A recently published detailed review summarizes the application of enzymes in continuous flow reactors.44 One of the examples is the application of cellobiose dehydrogenase and laccase in an enzyme membrane reactor with a bubble-free aeration for the production of lactobionic acid from lactose, where STY of 3.1 g·L−1·h−1 was reached in a reactor continuously operated for 3 days.45 Molecular oxygen was necessary for the regeneration enzyme; therefore, this reaction was not mentioned in Table 6. Nevertheless, continuous enzyme membrane reactors with biocatalyst entrapped within ultrafiltration membranes seem to be a most effective method

■

CONCLUSIONS The here presented tube-in-tube reactor is a suitable tool for conducting enzymatic gas dependent reactions at a preparative scale. The applicability of the system was shown for hydroxylation of 2-hydroxybiphenyl to 3-phenylcatechol by means of an enzyme catalyst coupled to an enzymatic regeneration of NADH by FDH. Determination of the oxygen transfer rates proved that the reaction was not limited by the oxygen transfer over the membrane or its availability in the aqueous phase. Safety issues related to the use of pure oxygen and organic solvent should be considered before further scale-up. Although it was shown that the oxygen transfer rate can be higher for pure oxygen instead of air, it might be more feasible from economic and safety point of view to use the latter. Using pure oxygen at elevated pressures poses a risk of explosion, and although the tube-in-tube reactor used here is designed in a way to minimize the volume of oxygen in the system, the safety aspect should not be ignored. Other parameters, which need optimization before scale-up, are the initial substrate concentration and residence time that drives the reaction to completion. As shown previously for the same reaction, its rate in liquid/liquid segmented flow microreactor decreases with decreasing initial substrate concentration.22 Therefore, lowering the initial substrate concentration from 200 mM to 100 mM, while keeping the residence time, would only increase the conversion by around 40%.22 In order to drive the reaction to completion, at compromised reaction rate, a very low substrate concentration would need to be applied and residence time adjusted accordingly for scale-up. To summarize, it is envisioned that the concept presented here can be used as a platform for future development of fast enzymatic reactions seeking more efficient gas supply and/or aqueous organic two-liquid phase system reactions dependent on high mass transfer rates. H



Article

Figure 8. Exemplary tube-in-tube reactor setup with the peripheral units.

■

HbpA Activity Determined by Product Formation: RPHPLC Analysis. Monitoring of the HbpA catalyzed biotransformation reaction was achieved by determining the concentration of the product 3-phenylcatechol. Measurements were done using a Hitachi Elite LaChrom HPLC system (VWR, Darmstadt, Germany) equipped with a Hitachi Elite LaChrom L-2450 diode array detector. All measurements were performed on an XTerra RP 18 3.5 μm column (Waters, Dublin, Ireland). An isocratic method at a total flow rate of 1 mL·min−1 50:50 (% v/v) ACN−water (both supplemented with 0.1% TFA) was used for elution. The product 3-phenylcatechol as well as the substrate 2-hydroxybiphenyl were detected at 2.69 and 3.19 min retention times. Concentrations of 3-phenylcatechol were calculated as a peak area ratio of 2-hydroxybiphenyl to 3phenylcatechol. Tube-in-Tube Reactor Setup. The tube-in-tube reactor Gastropod was obtained from Cambridge Reactor Design (Cambridge, UK). A sketched assembly is shown in Figure 5. A basic unit was 1 m long (volume 0.5 mL). The inner tube of TiTR was made of Teflon AF-2400 (i.d. 0.8 mm, o.d. 1.0 mm). The membrane was kept in a 1/8 in. Tefzel tube. Back pressure was controlled by a back pressure regulator (BPR) installed on the outlet of the membrane. To avoid pulsation in the system caused by the different liquid viscosities, a coiled 1 m long 1/8 in. PFA tube was connected by a T-connector (Upchurch Scientific, Göhler HPLC, Chemnitz, Germany) to the BPR. After equilibrating the system, the air, which was initially present in this tube, pressurized and exerted a stable back pressure on the system. TiTR assembly allowed independent control of the gas pressure and partial pressure difference across the membrane. Two Chemyx Nexus 6000 syringe pumps (KR Analytical, Sandbach, UK) were used to pump aqueous and organic phase streams. Liquids were united in a 1 mm internal diameter T-connector forming a stable segmented flow as sketched out in Figure 1.

EXPERIMENTAL SECTION Chemicals. NADH disodium salt and 1-decanol were obtained from AppliChem (Darmstadt, Germany), sodium formate, NAD+ disodium salt and FAD from Fluka (Buchs, Switzerland), 2-hydroxybiphenyl, activated aluminum oxide (neutral, CAMAG 507-C-I, Brockman grade I) from SigmaAldrich (Taufkirchen, Germany), and 3-phenylcatechol from Wako Chemical (Neuss, Germany). All other chemicals have been bought either at Fluka, AppliChem, or Sigma-Aldrich at the highest purity required and used as received. Enzymes. The production and purification of HbpA and FDH were based on previously published procedures43,50 and were used directly as 50% (v/v) glycerol stocks stored at −20 °C. HbpA Activity Assays. Unless indicated otherwise, reactions were conducted at 30 °C, which was the temperature used for the initial characterization of HbpA.51 HbpA-Activity Determined by NADH Consumption in an UV Spectrophotometer. HbpA activity was determined by measuring the consumption of NADH (ε = 6220 M−1·cm−1) at 340 nm using a Cary 300 Bio UV−vis spectrophotometer (Varian, Darmstadt, Germany). One unit of enzyme activity (U) refers to 1 μmol of NADH consumed in 1 min at the 30 °C. A standard assay was conducted as follows: potassium phosphate (KPi) buffer (960 μL of 100 mM, pH 7.5) supplemented with NADH (12 μL of 25 mM, 0.3 μmol) and FAD (10 μL of 2 mM, 0.02 μmol) was mixed in a 1.5 mL PMMA cuvette and thermostatted at 30 °C for 5 min; HbpA solution (10 μL, ca. 0.03 U) was added and mixed by pipet, and the background activity, i.e., a futile endogenous NADH oxidation in the absence of a substrate, was determined. Subsequently the reaction was started by the addition of methanolic solution of 2-hydroxybiphenyl (8 μL of 25 mM, 0.2 μmol). The reaction was monitored for 1 min. Substrate specific activity was corrected for endogenous NADH oxidation. I



Article

the HPLC, and most prominent fractions were pooled (ca. 70 mL, 99% purity by HPLC) and dried on a rotary evaporator (initially 40 °C and 250 mbar reduced to 28 mbar over 3.5 h; final drying step was done at 35 °C for 1 h). A total of 4 g of residue was collected and successfully sublimed in batches on a “coldfinger” (1.5−1.2 mbar, cooling water at 2 °C). The coldfinger was heated up uniformly either by a heat gun or in an oil bath set to 120 °C. The white powder was removed from the finger; the finger was washed with methanol, and the procedure was repeated. A total of 740 mg of product was collected. A portion of 11 mg of powder was resuspended in 0.75 mL of CDCl3 and analyzed on 300 MHz H NMR. The spectrum was compared to that of the standard of product and substrate, and the white powder was ultimately assigned as 3phenylcatechol. Buret Method. To quantify the amount of air passing through the membrane, the buret method adapted from O’Brien et al. (2011)14 was used (Figure 9). A glass buret

Air was delivered from an in-house pressurized air system (nominal pressure 8 bar). Oxygen was delivered to the outer membrane directly from the pressurized oxygen tank (200 bar). When longer residence times were required either a PTFE tube (1/16 in. i.d.) was connected to the reactor outlet before the BPR or a 4 m (2 mL) version of a TiTR was used, which was assembled on site with commercially available interconnects (Swagelok, Kings Langley, UK). The reaction temperature was controlled by keeping the reactor tubing in a water bath. An exemplary setup with a Gastropod TiTR and all the peripherals is presented in Figure 8. Reaction in TiTR. If not stated otherwise, all biotransformation reactions were conducted in the TiTR. Relative flow rates of aqueous (KPi buffer) and organic phase (decanol) were always kept constant at a 1:1 ratio. The aqueous phase was complemented by adding 50% glycerol solution of HbpA and FDH (for final volumetric activities of 2−20 U·mL−1), and depending on the required reaction conditions 100 or 166 mM sodium formate, 1 or 1.6 mM NAD+, 20 or 100 μM FAD, and 0.6 mg·mL−1 Tween 20 (for facilitating flow stability and preventing enzyme inactivation). The reaction mixture was filled up to the respective volume for reaching the required volumetric activities with KPi buffer (100 mM, pH 7.5). The aqueous phase was filtered through a 0.22 μm Milex CP filter with PES membrane to remove any particles that might otherwise deposit in the reactor and result in flow instabilities. 2-Hydroxybiphenyl (200 or 400 mM in organic phase) dissolved in decanol formed a separate organic phase. If not stated otherwise, the reactor was thermostatted at 30 °C in a water bath. Gram Scale Production of 3-Phenylcatechol in a Sequential TiTR. For gram scale production of 3-phenylcatechol, a reactor setup presented in Figure 5B was used. It consisted of two connected TiTR units, in which each single unit was followed by a coiled 3 m long PTFE tube (i.d. 1 mm, 0.8 mL·min−1) for residence time extension. Oxygen was delivered from an oxygen tank separately to each TiTR unit. Product was collected in a glass flask kept on ice and continuously flushed with nitrogen gas to avoid product degradation by overoxidation. The reaction temperature in the water bath was set to 35 °C/temperature in reaction mixture 30 °C. Product Purification. Product purification was achieved by applying flash chromatography method adapted from Schmid et al.50 and scaled up for our purposes. Shortly, 150 g of activated alumina was suspended in methanol and poured into a glass column. A portion of 0.5 cm of sea-sand was deposited on top. The column was washed with methanol. A portion of 360 mL of the decanol containing a mixture of 2-hydroxybiphenyl (10.4 g) and 3-phenylcatechol (2.0 g) was added and pushed through by applying a 0.5 bar overpressure. Fractions of 200−300 mL were collected. When decanol was almost eluted, methanol was applied on top. Elution was monitored by TLC (stationary phase = neutral alumina, mobile phase = 0.5 M HCl in methanol). A total of 0.9 L of methanol was used to elute decanol and residual substrate. When no more substrate or decanol could be detected, the eluent was changed to 0.5 M HCl in methanol for eluting the product. A sample of 1 L of eluent was used. After the first 150−200 mL, the fraction volume was changed to 20 mL. Product eluted from the column as a green liquid of neutral pH and was followed by an acidic brownish/red liquid that contained only residual amounts of product. Green-colored samples were checked on

Figure 9. Picture and schematic view of the setup used for buret method quantification of the gas passing through the Teflon AF-2400 membrane. A: Displaced liquid. B: Bubbles of the gas after decompression. C: Entry of the reactor outlet tube (BPR: back pressure regulator).

filled with liquid was inverted and submerged in a cylinder. The buret was closed at the top, and the liquid from the reactor was delivered to the buret from the bottom. When the liquid decompressed, gas travelled to the top and caused the displacement of the liquid in the buret. The volume of the gas was quantified, and the volume of the oxygen in the collected gas was calculated using Henry’s law.

■

ASSOCIATED CONTENT

■

AUTHOR INFORMATION

S Supporting Information *

Calculation of mass transfer; physical properties of the fluids used in this study; H NMR spectrum of 2-hydroxybiphenyl and 3-phenylcatechol standards; H NMR spectrum of 3-phenylcatechol synthesized in this study; selection of catalytically fast enzymes that could profit from enhanced mass transfer and high oxygen (or another gas) transfer rates in a TiTR. This material is available free of charge via the Internet at http:// pubs.acs.org. Corresponding Author

*Address: Dr. Katja Buehler, Chair of Chemical Biotechnology, TU Dortmund University, 44221 Dortmund, Germany. E-mail: [email protected]. J



Article

Notes

(21) Skowerski, K.; Czarnocki, S. J.; Knapkiewicz, P. ChemSusChem 2013, 7, 536. (22) Tomaszewski, B.; Lloyd, R. C.; Warr, A. J.; Buehler, K.; Schmid, A. ChemCatChem 2014, 6, 2567. (23) Battino, R.; Rettich, T. R.; Tominaga, T. J. Phys. Chem. Ref. Data 1983, 12, 163. (24) Kashid, M. N.; Harshe, Y. M.; Agar, D. W. Ind. Eng. Chem. Res. 2007, 46, 8420. (25) Thulasidas, T. C.; Abraham, M. A.; Cerro, R. L. Chem. Eng. Sci. 1997, 52, 2947. (26) Zaloha, P.; Kristal, J.; Jiricny, V.; Völkel, N.; Xuereb, C.; Aubin, J. Chem. Eng. Sci. 2012, 68, 640. (27) Adamo, A.; Heider, P. L.; Weeranoppanant, N.; Jensen, K. F. Ind. Eng. Chem. Res. 2013, 52, 10802. (28) Kohler, H. P.; Kohler-Staub, D.; Focht, D. D. Appl. Environ. Microbiol. 1988, 54, 2683. (29) Held, M.; Schmid, A.; Kohler, H. P.; Suske, W.; Witholt, B.; Wubbolts, M. G. Biotechnol. Bioeng. 1999, 62, 641. (30) Yang, L.; Jensen, K. F. Org. Process Res. Dev. 2013, 17, 927. (31) Hofstetter, K.; Lutz, J.; Lang, I.; Witholt, B.; Schmid, A. Angew. Chem., Int. Ed. 2004, 43, 2163. (32) Woltinger, J.; Karau, A.; Leuchtenberger, W.; Drauz, K. Adv. Biochem. Eng./Biotechnol. 2005, 92, 289. (33) Rios, G. M.; Belleville, M. P.; Paolucci, D.; Sanchez, J. J. Membr. Sci. 2004, 242, 189. (34) Lopez, J. L.; Matson, S. L. J. Membr. Sci. 1997, 125, 189. (35) Günther, A.; Jensen, K. F. Lab Chip 2006, 6, 1487. (36) Karande, R. PhD thesis. TU Dortmund University, Dortmund, Germany, 2012. (37) Karande, R.; Schmid, A.; Buehler, K. Adv. Synth. Catal. 2011, 353, 2511. (38) Burns, J. R.; Ramshaw, C. Lab Chip 2001, 1, 10. (39) Coronas, J.; Santamaría, J. Catal. Today 1999, 51, 377. (40) Lütz, S.; Rao, N. N.; Wandrey, C. Chem. Eng. Technol. 2006, 29, 1404. (41) Ma, S. K.; Gruber, J.; Davis, C.; Newman, L.; Gray, D.; Wang, A.; Grate, J.; Huisman, G. W.; Sheldon, R. A. Green Chem. 2010, 12, 81. (42) Dittmeyer, R.; Höllein, V.; Daub, K. J. Mol. Catal. A: Chem. 2001, 173, 135. (43) Lutz, J.; Mozhaev, V. V.; Khmelnitsky, Y. L.; Witholt, B.; Schmid, A. J. Mol. Catal. B: Enzym. 2002, 19−20, 177. (44) Yuryev, R.; Strompen, S.; Liese, A. Beilstein J. Org. Chem. 2011, 7, 1449. (45) Van Hecke, W.; Haltrich, D.; Frahm, B.; Brod, H.; Dewulf, J.; Van Langenhove, H.; Ludwig, R. J. Mol. Catal. B: Enzym. 2011, 68, 154. (46) Ferloni, C.; Heinemann, M.; Hummel, W.; Daussmann, T.; Büchs, J. Biotechnol. Prog. 2004, 20, 975. (47) Thomsen, M. S.; Nidetzky, B. Biotechnol. J. 2009, 4, 98. (48) Wiles, C.; Hammond, M. J.; Watts, P. Beilstein J. Org. Chem. 2009, 5, 27. (49) Asanomi, Y.; Yamaguchi, H.; Miyazaki, M.; Maeda, H. Molecules 2011, 16, 6041. (50) Schmid, A.; Vereyken, I.; Held, M.; Witholt, B. J. Mol. Catal. B: Enzym. 2001, 11, 455. (51) Suske, W. A.; Held, M.; Schmid, A.; Fleischmann, T.; Wubbolts, M. G.; Kohler, H. P. J. Biol. Chem. 1997, 272, 24257.

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS We would like to thank Dr. Rohan Karande for his helpful suggestions during the experimental work and for the critical reading of the manuscript. This project was supported by the BIOTRAINS Marie Curie Initial Training Network, financed by the European Union through the seventh Framework People Program (grant agreement number 238531).

■

ABBREVIATIONS

■

REFERENCES

TTN, total turnover number (μmolprod·μmolenz−1), aka turnover frequency; HbpA, 2-hydroxybiphenyl 3-monooxygenase; FDH, formate dehydrogenase; PFA, perfluoroalkoxy alkane; Tefzel, modified ETFE (ethylene-tetrafluoroethylene); Teflon AF-2400, amorphous fluoroplastic, a copolymer of 2,2bistrifluoromethyl-4,5-difluoro-1,3-dioxole and tetrafluoroethylene; KPi, potassium phosphate buffer; kLa, overall volumetric mass transfer coefficient, s−1, min−1; TiTR, tube-in-tube reactor

(1) Li, Z.; van Beilen, J. B.; Duetz, W. A.; Schmid, A.; de Raadt, A.; Griengl, H.; Witholt, B. Curr. Opin. Chem. Biol. 2002, 6, 136. (2) Schmid, A.; Dordick, J. S.; Hauer, B.; Kiener, A.; Wubbolts, M.; Witholt, B. Nature 2001, 409, 258. (3) Liese, A.; Seelbach, K.; Buchholz, A.; Haberland, J. Industrial Biotransformations; Wiley-VCH Verlag GmbH & Co. KGaA: Weinheim, Germany, 2006. (4) Bornscheuer, U. T.; Huisman, G. W.; Kazlauskas, R. J.; Lutz, S.; Moore, J. C.; Robins, K. Nature 2012, 485, 185. (5) Wichmann, R.; Wandrey, C.; Bückmann, A. F.; Kula, M.-R. Biotechnol. Bioeng. 1981, 23, 2789. (6) Bossow, B.; Wandrey, C. Ann. N.Y. Acad. Sci. 1987, 506, 325. (7) Liese, A.; Zelinski, T.; Kula, M. R.; Kierkels, H.; Karutz, M.; Kragl, U.; Wandrey, C. J. Mol. Catal. B: Enzym. 1998, 4, 91. (8) Hambach, B.; Biselli, M.; Runstadler, P. W.; Wandrey, C. In Animal Cell Technology; Spier, R. E., Griffiths, J. B., MacDonald, C., Eds.; Butterworth-Heinemann: London, 1992; p 381. (9) Rissom, S.; Schwarz-Linek, U.; Vogel, M.; Tishkov, V. I.; Kragl, U. Tetrahedron: Asymm. 1997, 8, 2523. (10) Schwarz-Linek, U.; Krödel, A.; Ludwig, F.-A.; Schulze, A.; Rissom, S.; Kragl, U.; Tishkov, V. I.; Vogel, M. Synthesis 2001, 2001, 0947. (11) Pilz, R.; Hammer, E.; Schauer, F.; Kragl, U. Appl. Microbiol. Biotechnol. 2003, 60, 708. (12) O’Brien, M.; Baxendale, I. R.; Ley, S. V. Org. Lett. 2010, 12, 1596. (13) Koos, P.; Gross, U.; Polyzos, A.; O’Brien, M.; Baxendale, I.; Ley, S. V. Org. Biomol. Chem. 2011, 9, 6903. (14) O’Brien, M.; Taylor, N.; Polyzos, A.; Baxendale, I. R.; Ley, S. V. Chem. Sci. 2011, 2, 1250. (15) Polyzos, A.; O’Brien, M.; Petersen, T. P.; Baxendale, I. R.; Ley, S. V. Angew. Chem., Int. Ed. 2011, 50, 1190. (16) Browne, D. L.; O’Brien, M.; Koos, P.; Cranwell, P. B.; Polyzos, A.; Ley, S. V. Synlett 2012, 23, 1402. (17) Cranwell, P. B.; O’Brien, M.; Browne, D. L.; Koos, P.; Polyzos, A.; Pena-Lopez, M.; Ley, S. V. Org. Biomol. Chem. 2012, 10, 5774. (18) Petersen, T. P.; Polyzos, A.; O’Brien, M.; Ulven, T.; Baxendale, I. R.; Ley, S. V. ChemSusChem 2012, 5, 274. (19) Bourne, S. L.; O’Brien, M.; Kasinathan, S.; Koos, P.; Tolstoy, P.; Hu, D. X.; Bates, R. W.; Martin, B.; Schenkel, B.; Ley, S. V. ChemCatChem 2013, 5, 159. (20) Pastre, J. C.; Browne, D. L.; O’Brien, M.; Ley, S. V. Org. Process Res. Dev. 2013, 17, 1183. K


Biocatalytic Production of Catechols Using a High Pressure Tube-in

Recommend Documents