A Phenolic Acid Decarboxylase-Based All-Enzyme Hydrogel for Flow Reactor Technology

micromachines

Article A Phenolic Acid Decarboxylase-Based All-Enzyme Hydrogel for Flow Reactor Technology

Esther Mittmann 1 , Sabrina Gallus 1, Patrick Bitterwolf 1, Claude Oelschlaeger 2, Norbert Willenbacher 2, Christof M. Niemeyer 1 and Kersten S. Rabe 1,*

1 Institute for Biological Interfaces (IBG-1), Karlsruhe Institute of Technology (KIT), 76187 Karlsruhe, Germany; [email protected] (E.M.); [email protected] (S.G.); [email protected] (P.B.); [email protected] (C.M.N.) 2 Institute for Mechanical Process Engineering and Mechanics (MVM), Karlsruhe Institute of Technology (KIT), 76187 Karlsruhe, Germany; [email protected] (C.O.); [email protected] (N.W.) * Correspondence: [email protected]; Tel.: +49-721-608-24469

Received: 18 October 2019; Accepted: 18 November 2019; Published: 20 November 2019

Abstract: Carrier-free enzyme immobilization techniques are an important development in the field of efficient and streamlined continuous synthetic processes using microreactors. Here, the use of monolithic, self-assembling all-enzyme hydrogels is expanded to phenolic acid decarboxylases. This provides access to the continuous flow production of p-hydroxystyrene from p-coumaric acid for more than 10 h with conversions 98% and space time yields of 57.7 g (d L) 1. ≥ · · − Furthermore, modulation of the degree of crosslinking in the hydrogels resulted in a defined variation of the rheological behavior in terms of elasticity and mesh size of the corresponding materials. This work is addressing the demand of sustainable strategies for defunctionalization of renewable feedstocks.

Keywords: biocatalysis; phenolic acid decarboxylase; hydrogel; enzyme immobilization; SpyCatcher/SpyTag; microreactor

1. Introduction The current transformation of the chemical industry calls for a constant intensification of the use of biomass as feedstock for the sustainable synthesis of fine chemicals. However, the high complexity of this renewable raw material and especially the high mass fraction of oxygen in comparison to fossil carbon resources is limiting its widespread utilization. Decarboxylation is an important method for defunctionalization, which can be employed to alleviate these impediments [1–5]. In order to add value to ubiquitously available bio-derived phenolic acids [6–11], and to provide alternatives to chemical methods often relying on transition metal catalysis [2,10,12], phenolic acid decarboxylases (PADs) offer an elegant enzymatic route to catalyse the elimination of carbon dioxide from hydroxyphenacrylic acids [13–15]. The resulting styrene derivatives are of great industrial relevance since the terminal alkene provides a versatile handle for further chemical functionalization or polymerization reactions [16–18]. In contrast to other (de)carboxylases, several PADs found in bacteria [19–22] and plants [23] act on an acid-base mechanism via a quinone methide intermediate [24] and do not require a cofactor, therefore eliminating the need of a laborious and expensive cofactor regeneration system [25,26]. The versatility of PADs has been demonstrated by integration of the decarboxylation reaction into biocatalytic [6,27] and chemical [28–30] reaction cascades. However, limitations can arise due to the product inhibition [28,29] frequently observed for this class of enzymes. The transfer of the process from batch stirred tank reactors to fluidic, micro- and macrostructured reactors can provide a viable

Micromachines 2019, 10, 795; doi:10.3390/mi10120795 www.mdpi.com/journal/micromachines Micromachines 2019, 10, 795 2 of 12 solution, as the convective product removal can avoid the accumulation of high and adverse product concentrationsMicromachines 2019 [31, 10]., Therefore,x much effort is currently devoted to the development of continuous2 of 13 flow reactors employing immobilized enzymes [32–36]. In general, such reactors can be prepared continuous flow reactors employing immobilized enzymes [32–36]. In general, such reactors can be by anchoring the enzyme of interest on the reactor surface using a variety of chemical methods or prepared by anchoring the enzyme of interest on the reactor surface using a variety of chemical else by entrapment or crosslinking of the enzymes [34,37,38]. Immobilizing the enzymes on the methods or else by entrapment or crosslinking of the enzymes [34,37,38]. Immobilizing the enzymes planaron the surfaces planar of surfaces a reactor of wall a reactor can result wall incan a singleresult monolayer,in a single limitingmonolayer, the limiting catalytic the performance. catalytic Here,performance. employing immobilizationHere, employing strategies immobilization which result strategies in three-dimensional which result structures,in three‐dimensional can lead to a significantstructures, increase can lead of the to enzymatica significant activity increase per reactorof the enzymatic volume. This activity approach per reactor can be facilitatedvolume. This with so-calledapproach crosslinked can be facilitated enzyme aggregateswith so‐called or crystals crosslinked (CLEAs enzyme or CLECs) aggregates [39– 43or]. crystals Another (CLEAs advantage or of thisCLECs) strategy [39–43]. is the Another minimization advantage of of carrier this strategy materials, is the so minimization that only a small of carrier volume materials, of the valuableso that reactoronly spacea small is volume needed of for the supportingvaluable reactor structures. space is However,needed for sincesupporting the crosslinking structures. However, process usually since reliesthe on crosslinking non-specific process chemical usually crosslinkers, relies on non this‐specific method chemical can lead crosslinkers, to the inactivation this method of enzymes. can lead to theWe inactivation have recently of enzymes. demonstrated the construction of self-assembling all-enzyme hydrogels constructedWe have from recently ketoreductases demonstrated and glucose the construction dehydrogenases, of self‐ whichassembling were all genetically‐enzyme hydrogels fused with eitherconstructed the SpyTag from (ST) ketoreductases peptide or the and SpyCatcher glucose (SC)dehydrogenases, protein, respectively which were [44– 46genetically]. This highly fused e ffiwithcient autocatalyticeither the SpyTag bioconjugation (ST) peptide system or isthe based SpyCatcher on the rapid(SC) protein, formation respectively of a covalent [44–46]. isopeptide This highly bond throughefficient the autocatalytic SpyTag-SpyCatcher bioconjugation complex system that occurs is based under on the physiological rapid formation conditions of a covalent [47–50 isopeptide]. To render thebond all-enzyme through hydrogel the SpyTag immobilization‐SpyCatcher strategycomplex applicable that occurs to under a broader physiological range of biocatalysts,conditions [47–50]. we here report,To render for the the first all time,‐enzyme on thehydrogel use of recombinantimmobilization variants strategy of applicable a highly active to a homodimericbroader range PADof biocatalysts, we here report, for the first time, on the use of recombinant variants of a highly active (molecular weight of a monomer 20 kDa) obtained from Enterobacter sp. [51]. By creating SC- and homodimeric PAD (molecular weight of a monomer 20 kDa) obtained from Enterobacter sp. [51]. By ST-tagged PAD variants we demonstrate the fabrication of all-enzyme hydrogels. The functionality of creating SC‐ and ST‐tagged PAD variants we demonstrate the fabrication of all‐enzyme hydrogels. the novel biocatalytic materials is demonstrated by the continuous production of p-hydroxystyrene in The functionality of the novel biocatalytic materials is demonstrated by the continuous production a microfluidic flow reactor (Figure1). of p‐hydroxystyrene in a microfluidic flow reactor (Figure 1).

(a) (b) (c) (d)

FigureFigure 1. (1.a) (a Continuous) Continuous synthesis synthesis of of substituted substituted styrenesstyrenes from readily readily available available bio bio-derived‐derived phenolic phenolic acidsacids through through biocatalytic biocatalytic decarboxylation, decarboxylation, as as exempliﬁed exemplified here for p p-hydroxystyrene,‐hydroxystyrene, is is achieved achieved by by implementationimplementation of of a microreactor a microreactor (b ().b). This This reactor reactor is is ﬁlledfilled with an all all-enzyme‐enzyme hydrogel hydrogel (c ()c comprised) comprised of theof the dimeric dimeric Phenolic Phenolic Acid Acid Decarboxylase Decarboxylase (PAD), (PAD), crosslinked crosslinked via isopeptide bonds bonds spontaneously spontaneously generatedgenerated through through genetically genetically fused fused SpyTag SpyTag/SpyCatcher/SpyCatcher complexes complexes (d). Note (d). thatNote the that protein the hydrogelprotein hydrogel loaded into the micro reactor (b) is stained with coomassie brilliant blue. loaded into the micro reactor (b) is stained with coomassie brilliant blue.

2. Materials2. Materials and and Methods Methods AllAll chemicals chemicals were were purchased purchased from from Sigma Sigma AldrichAldrich (St. Louis, Louis, MI, MI, USA) USA) or or VWR VWR (Radnor, (Radnor, PA, PA, USA)USA) if not if not stated stated otherwise. otherwise.

2.1.2.1. Synthesis Synthesis of p-Hydroxystyreneof p‐Hydroxystyrene Via Via Wittig Wittig Reaction Reaction ForFor calibration calibration curves curves and and as as positive positive control, control,p p-hydroxystyrene‐hydroxystyrene was was synthesized synthesized as as described described previouslypreviously (Scheme (Scheme1)[ 1)52 [52].]. InIn brief, potassium potassium tert tert-butoxide‐butoxide (≥98%, ( 98%, 2.81 2.81 g; 25 g; mmol) 25 mmol) was added was added to a mixture of methyl triphenylphosphonium bromide (≥98%, ≥5.36 g; 15 mmol) in anhydrous to a mixture of methyl triphenylphosphonium bromide ( 98%, 5.36 g; 15 mmol) in anhydrous tetrahydrofurane THF (ACS grade, 20 mL) under argon atmosphere.≥ The mixture was stirred for 10 tetrahydrofurane THF (ACS grade, 20 mL) under argon atmosphere. The mixture was stirred for 10 min min at ambient temperature. Subsequently p‐hydroxybenzaldehyde (≥98%, 1.52 g; 10 mmol) at ambient temperature. Subsequently p-hydroxybenzaldehyde ( 98%, 1.52 g; 10 mmol) dissolved in dissolved in THF (10 mL) was added dropwise. The reaction was≥ stirred for 4 h and the product formation was monitored by thin layer chromatography (TLC). After quenching of the reaction

Micromachines 2019, 10, 795 3 of 12

THF (10 mL) was added dropwise. The reaction was stirred for 4 h and the product formation was Micromachines 2019, 10, x 3 of 13 monitored by thin layer chromatography (TLC). After quenching of the reaction mixture with saturated

NHmixture4Cl solution with saturated and the removal NH4Cl ofsolution THF under and the reduced removal pressure, of THF theunder crude reduced product pressure, was extracted the crude with CHproduct2Cl2 (analytical was extracted grade). with The CH organic2Cl2 (analytical layer was grade). washed The organic with a saturatedlayer was washed NaCl solution, with a saturated dried over anhydrousNaCl solution, Na2SO dried4, and over after anhydrous ﬁltration andNa2SO evaporation4, and after of filtration the solvent, and evaporation puriﬁed by chromatographyof the solvent, columnpurified (SiO by2, 5%chromatography ethyl acetate incolumnn-hexane) (SiO to2, give 5% pureethylp -hydroxystyrene,acetate in n‐hexane) 834.5 mgto (6.945give pure mmol) inp a‐hydroxystyrene, 69.5% yield as white 834.5 crystals.mg (6.945 R mmol)f = 0.44 in (16% a 69.5% ethyl yield acetate as white in n-hexane). crystals. TheRf = nuclear0.44 (16% magnetic ethyl resonanceacetate in (NMR) n‐hexane). spectra The matched nuclear magnetic with literature resonance data. (NMR) spectra matched with literature data.

Scheme 1. Synthesis of p-hydroxystyrene. Scheme 1. Synthesis of p‐hydroxystyrene. 2.2. Plasmid Construction 2.2. Plasmid Construction Plasmid construction was carried out using the isothermal recombination as described by Gibson Plasmid construction was carried out using the isothermal recombination as described by et al. utilizing oligonucleotide primers with 20–30 bp homologous overlaps [53]. The products of Gibson et al. utilizing oligonucleotide primers with 20–30 bp homologous overlaps [53]. The products the in vitro recombination were then incubated for 1 h at 37 C with DpnI (New England Bioloabs, of the in vitro recombination were then incubated for 1 h at 37 ◦°C with DpnI (New England Bioloabs, Beverly, MA, USA) to ensure the sample did not contain any methylated DNA templates from Beverly, MA, USA) to ensure the sample did not contain any methylated DNA templates from previousprevious polymerase polymerase chain chain reactions. reactions. TheThe purificationpurification of of dissolved dissolved plasmids plasmids was was carried carried out out using using thethe ZR ZR Plasmid Plasmid Miniprep-Classic Miniprep‐Classic (Zymo (Zymo Research,Research, Freiburg, Germany). Germany). Sequences Sequences were were verified verified by by commercialcommercial sequencing sequencing (LGC (LGC Genomics, Genomics, Berlin, Berlin, Germany). Germany). The The sequences sequences of theof the primers primers can can be found be infound Table1 in. pET_PAD-His Table 1. pET_PAD was‐His used was as aused template as a template for further for further cloning cloning steps [ steps30]. [30]. pET_PAD‐SC‐His: The backbone encoding for a N‐terminal PAD and C‐terminal 6× His‐Tag separated by a glycine spacer wasTable amplified 1. Primers using used primers in this EM01 study. and EM02 with pET_PAD‐His as the template. This backbone was then recombined with a SC encoding insert, which had been Primer DNA-Sequence generated by polymerase chain reaction (PCR) using the primers EM03 and EM04 with pTF16_Lpp‐ EM01OmpA‐SC GGTCATCATCACCATCATCATTAAGATCC as the template [48]. pET_PAD‐ST‐His: The C‐terminal ST sequence were inserted into EM02plasmid pET_PAD GCTACCACCACCACCTTTCAGATTATC‐His by PCR using the primers EM05 and EM06. pET_ST‐PAD‐ST‐His: The EM03plasmid pET_PAD CCGGATAATCTGAAAGGTGGTGGTGGTAGCGTTGATACCCTGAGCGGTCTGAG‐ST‐His was amplified by PCR using the primers EM07 and EM08 resulting in a linearized plasmid backbone which was then complemented to the full‐length PAD with a DNA‐ EM04 CGGATCTTAATGATGATGGTGATGATGACCAATATGTGCATCACCTTTGGTTGCTTTACC fragment containing an additional N‐terminal ST‐GGGGS and matching overlaps (synthesized via EM05“GeneArt TGGTTGATGCATATAAACCGACCAAAGGTCATCATCACCATCATCATTAAGATCCGene Synthesis” by ThermoFisher, Waltham, MA, USA) using Gibson assembly, as EM06described above. CGGTTTATATGCATCAACCATAACAATATGTGCGCTACCACCACCACCTTTCAGATTATC EM07 CACAATTCCCCTATAGTGAGTCGTATTAATTTC Table 1. Primers used in this study. EM08 GTGAAAGCATCTATAAAATCAGCTGGACC Primer DNA‐Sequence EM01 GGTCATCATCACCATCATCATTAAGATCC pET_PAD-SC-His: The backbone encoding for a N-terminal PAD and C-terminal 6 His-Tag EM02 GCTACCACCACCACCTTTCAGATTATC × separated by a glycine spacer was amplified using primers EM01 and EM02 with pET_PAD-His as the EM03 CCGGATAATCTGAAAGGTGGTGGTGGTAGCGTTGATACCCTGAGCGGTCTGAG template. This backbone was then recombined with a SC encoding insert, which had been generated EM04 CGGATCTTAATGATGATGGTGATGATGACCAATATGTGCATCACCTTTGGTTGCTTTACC by polymerase chain reaction (PCR) using the primers EM03 and EM04 with pTF16_Lpp-OmpA-SC EM05 TGGTTGATGCATATAAACCGACCAAAGGTCATCATCACCATCATCATTAAGATCC as the template [48]. pET_PAD-ST-His: The C-terminal ST sequence were inserted into plasmid EM06 CGGTTTATATGCATCAACCATAACAATATGTGCGCTACCACCACCACCTTTCAGATTATC pET_PAD-His by PCR using the primers EM05 and EM06. pET_ST-PAD-ST-His: The plasmid EM07 CACAATTCCCCTATAGTGAGTCGTATTAATTTC pET_PAD-ST-His was amplified by PCR using the primers EM07 and EM08 resulting in a linearized EM08 GTGAAAGCATCTATAAAATCAGCTGGACC plasmid backbone which was then complemented to the full-length PAD with a DNA-fragment 2.3. Gene Expression and Protein Production

Micromachines 2019, 10, 795 4 of 12 containing an additional N-terminal ST-GGGGS and matching overlaps (synthesized via “GeneArt Gene Synthesis” by ThermoFisher, Waltham, MA, USA) using Gibson assembly, as described above.

2.3. Gene Expression and Protein Production Chemically competent Escherichia coli BL21 (DE3) were transformed with the corresponding expression vector. E. coli cells containing the expression vector were selected overnight at 37 ◦C on lysogeny broth (LB)/agar plates containing 100 µg/mL ampicillin. For expression, a single colony was transferred to a suitable volume of LB medium containing 100 µg/mL ampicillin and the culture was incubated overnight at 37 ◦C and 180 rpm in a shaking incubator. The overnight culture was then used to inoculate a 20 times larger volume of 25 ◦C warm ampicillin containing LB medium. This culture was then incubated for approximately 1.5–2 h up to an OD600 of 0.6–0.9 at 37 ◦C and 180 rpm. After reaching the appropriate cell density, the culture was cooled down to 25 ◦C for at least 15 min and subsequently induced with isopropyl β-D-1-thiogalactopyranoside (IPTG) to a final concentration of 0.5 mM. The induced culture was then incubated overnight at 25 C and then harvested by centrifugation (10,000 g, 10 min, 4 C). The cells ◦ × ◦ were then resuspended in cold sodium phosphate buffer containing 10 mM imidazole (NPI 10; 50 mM NaH PO , 500 mM NaCl, 10 mM Imidazole, pH 8) and frozen at 80 C. 2 4 − ◦ 2.4. Protein Puriﬁcation

The cell pellets were thawed at 25 ◦C in a water bath and then incubated with DNaseI and lysozyme for 30 min at 25 ◦C. Further cell disruption was then carried out using ultrasonification. The lysate was subsequently centrifuged for 1 h at 45,000 rcf and 4 ◦C. The pellet was discarded and the supernatant sterile-filtered through a 0.45 µm Durapore PVDF membrane with Steriflip® Filter Units (Merck, Darmstadt, Germany). For protein purification, the cleared lysate was loaded on 2 5 mL His Ni Superflow Cartridge × 60 (Clontech, Palo Alto, CA, USA) mounted on an Äkta Pure liquid chromatography system (GE Healthcare, Freiburg, Germany). The purification was carried out with NPI 10 as running buffer and sodium phosphate buffer containing 500 mM imidazole (NPI 500;50 mM NaH2PO4, 500 mM NaCl, 500 mM Imidazole, pH 8) as elution buffer. After applying the lysate onto the column, the column was washed with 2% NPI 500 and the protein was eluted using a linear gradient (2% to 100% NPI 500). The column outflow was collected in 900 µL fractions and protein containing fractions (detected at 280 nm) were pooled. Subsequently,the buffer was exchanged to phosphate buffered saline (PBS) using Vivaspin 20, 5000 MWCO concentrators (Sartorius, Göttingen, Germany).

2.5. Sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) Analysis Samples were mixed with 4 SDS-PAGE loading buﬀer (200 mM Tris-Cl, pH 6.8, 400 mM DTT, 8% × SDS, 0.4% bromophenol blue, 40% glycerol), boiled at 95 ◦C for 10 min and loaded onto a 1 mm thick SDS-PAGE gel containing 16% acrylamide. The gels were run at 100 V and stained with coomassie Brilliant Blue G-250. PageRuler (Plus) Prestained (Thermo Fisher Scientiﬁc) was used as molecular weight reference marker.

2.6. Determination of Decarboxylase Activity

Fifty µL of an enzyme solution in KPi buffer (25 mM KPi, pH 6) were transferred in an ultraviolet (UV) transparent 96 well microtiter plate and 150 µL of p-coumaric acid to a final concentration of 1 mM in KPi buffer were added. The consumption of p-coumaric acid was recorded at 294 nm using a Synergy MX microplate reader (BioTek, Winooski, VT, USA) over a period of 10 min at 25 ◦C. The activity was determined by calculating the slope in the linear range of the decrease of the absorption intensity (OD/min).

2.7. Hydrogel Prepration

Protein solutions of PAD-SC and PAD-ST/ST-PAD-ST were diluted in KPi buffer to a final total protein concentration of 1 mM in 20 µL for analysis of ST/SC complex formation. For microrheology Micromachines 2019, 10, 795 5 of 12 measurements a total volume of 200 µL was used and 0.2 mg/mL “dragon” green fluorescent polystyrene microspheres (200 nm diameter; Bangs Laboratories, Fishers, IN, US) were added. Polymerization was carried out for 1 h at 30 ◦C, 1000 rpm in a thermoshaker. Subsequently, the buffer was evaporated under constant centrifugation at 2200 g for overnight at 30 C. × ◦ 2.8. Microrheology Measurements Prior to multiple particle tracking (MPT)-analysis, the dried hydrogel samples were swollen by adding 25 µL KPi buffer for 30 min under continuous shaking at 25 ◦C, 500 rpm. To perform MPT experiments, we used green fluorescent polystyrene microspheres of 200 nm diameter as tracer particles. Images of the beads were recorded onto a personal computer via a sCMOS camera Zyla X (Andor Technology, 50 f/s, Belfast, Northern Ireland) mounted on an inverted fluorescence microscope (Axio Observer D1, Zeiss, Kohen, Germany) equipped with a Fluar 100 , N.A. 1.3, oil-immersion lens. × Movies of more than 100 fluctuating microspheres were analysed using the software Image Processing System (iPS, Visiometrics, Terrassa, Spain) and a self-written Matlab code, based on the widely used Crocker and Grier tracking algorithm [54]. Tocharacterize sample heterogeneities, we examined the distribution of mean square displacements (MSDs), known as Van Hove correlation function [55] and calculated the non-Gaussian parameter α [56]: x4(τ) α = h i 1 (1) 3 x2(τ) 2 − h i where x is the distance of particle center of mass along the x coordinate. This quantity is zero for a Gaussian distribution which is expected for a homogeneous and uniform sample, while higher α values reflect the presence of spatial heterogeneities.

2.9. High Performance Liquid Chromatography (HPLC) Analysis All high performance liquid chromatography (HPLC) analyses were performed on an Agilent Technologies (Palo Alto, CA, USA) 1260 Bioinert Series with autosampler and diode array detector (DAD). P-hydroxystyrene and p-coumaric acid were detected and quantified by reverse phase HPLC using an Eclipse XDB C18 column (5 µm, Agilent) with a precolumn of the same material. The separation was realized at 10 ◦C with an isocratic mobile phase consisting of 60% acetonitrile and 40% ddH2O with 0.1% (v/v) trifluoroacetic acid. The flow rate was 1.5 mL/min. Absorption was detected simultaneously at 260, 275 and 370 nm and a suitable wavelength for calibration and evaluation was chosen for each substrate. For calibration, dilutions of p-hydroxystyrene and p-coumaric acid in the range of 0.5–5 mM were prepared. For measurements, 150 µL of each sample were diluted with 150 µL acetonitrile additionally containing 40 µM p-hydroxyazobenzene as internal standard and 250 mM HCl prior to application.

2.10. Microfluidic Setup and Analysis of the Hydrogels Under Continuous Flow The microfluidic reactor was prepared as previously described [44]. The upper part containing the reaction channel was prepared by pouring of polydimethylsiloxane (PDMS) (Sylgard 184, Dow Corning, Midland, MI, USA) over molds to generate replicas. The channel architecture featured a straight channel which was 3 mm wide, 1 mm high and 54 mm long and a total volume of 150 µL. To serve as spacers for later fluidic connectors, cannulas (Sterican, B. Braun Melsungen AG, Melsungen, Germany) were integrated in the brass replica molds. The PDMS was cured at 60 ◦C for at least 3 h. The microreactor was typically mounted inside an incubator (set to 30 ◦C), filled with 150 µL 1 mM protein solution (PAD-SC/ST-PAD-ST 2:1) and incubated for 60 min. The PDMS chips were then sealed with a polyolefin foil (HJ-BIOANALYTIK GmbH, Erkelenz, Germany). Low pressure syringe pumps (neMESYS 290N) equipped with 10 mL syringes were connected to a manual switching valve that was connected to the reactor for perfusion of reaction media at a flowrate of 10 µL/min. The syringes were filled with 10 mL substrate solution containing 5 mM of p-coumaric acid in KPi buffer and 0.01% Micromachines 2019, 10, 795 6 of 12

(v/v) sodium azide to avoid fouling. The reactor outflow was connected to the Compact Positioning System rotAXYS360 (CETONI, Korbussen, Germany) to allow for automatic fractioning into 96-well Micromachines 2019, 10, x 6 of 13 plates, previously loaded with 150 µL 250 mM HCl in acetonitrile to stop all enzymatic reactions and 40 µMto pstop-hydroxyazobenzene all enzymatic reactions used as and standard 40 μM for p‐ internalhydroxyazobenzene calibration. Theused flow as standard was generated for internal using a CETONIcalibration. neMESYS The flow Base was module. generated Syringe using pump a CETONI and positioning neMESYS Base system module. were Syringe controlled pump by and the QmixElements-Software.positioning system were The flowcontrolled modules by were the connectedQmixElements via tubing‐Software. (inlets: The silicone flow modules Tygon tubing were R3603,connected ID = 1.6 mm,via tubing Saint-Gobain, (inlets: silicone Cavaillon, Tygon France; tubing outlets: R3603, conventional ID = 1.6 mm, PTFE Saint tubing,‐Gobain, ID =Cavaillon,0.5 mm). France; outlets: conventional PTFE tubing, ID = 0.5 mm). 3. Results and Discussion 3. Results and Discussion In order to develop a PAD-based all-enzyme hydrogel, we initially generated expression plasmids for PAD variantsIn order that to weredevelop genetically a PAD‐ fusedbased withall‐enzyme polypeptide hydrogel, domains we ofinitially SC or ST.generated Overexpression expression in E. coliplasmidsand purification for PAD variants via an additional that were 6geneticallyHis-Tag fused allowed with to polypeptide obtain the biocatalyticallydomains of SC or active ST. × recombinantOverexpression proteins in in E. near coli homogeneousand purification purity, via an as judgedadditional from 6× SDS-PAGE His‐Tag allowed analysis to (Figureobtain 2thea). This highbiocatalytically purity allowed active us recombinant to later produce proteins materials in near homogeneous with precisely purity, defined as properties.judged from All SDS variants‐PAGE expressedanalysis with (Figure very high2a). This expression high purity yields allowed100 mg us /toL culturelater produce in a non-optimized materials with process. precisely To defined assess ≥ the catalyticproperties. performance All variants of expressed these PAD with species, very a high first expression benchmark yields test was ≥100 performed mg/L culture following in a non the‐ decreaseoptimized of substrate process. by To monitoring assess the its catalytic absorption performance at 294 nm. of these The results PAD species, in Figure a first2b show benchmark the direct test was performed following the decrease of substrate by monitoring its absorption at 294 nm. The comparison of the activity of the various enzyme variants to convert the model substrate p-coumaric results in Figure 2b show the direct comparison of the activity of the various enzyme variants to acid. It is clearly evident that all variants show comparable activities, however, the addition of one or convert the model substrate p‐coumaric acid. It is clearly evident that all variants show comparable two tags slightly decreased the activity as compared to the enzyme without an additional tag (dark activities, however, the addition of one or two tags slightly decreased the activity as compared to the blue bar, Figure2b). enzyme without an additional tag (dark blue bar, Figure 2b).

(a) (b)

FigureFigure 2. Variants 2. Variants of phenolic of phenolic acid acid decarboxylase decarboxylase (PAD) (PAD) from fromEnterobacter Enterobactersp. sp. containing containing either either one one or two SpyTagor two SpyTag (ST) or one(ST) SpyCatcher or one SpyCatcher (SC) were (SC) expressed were expressed in Escherichia in Escherichia coli, purified coli, purified and tested and tested for their for activitytheir towards activityp-coumaric towards p acid.‐coumaric Note acid. that all Note proteins that all carry proteins a C-terminal carry a 6 C‐His-Tag.terminal 6× (a) His Denaturing‐Tag. (a) × 16% SDS-PAGEDenaturing analysis 16% SDS of‐PAGE the purified analysis PAD of the constructs. purified PAD The PAD constructs. variants The used PAD here variants are homodimers used here are in homodimers in their native form. Lane 1: PAD (20 kDa), Lane 2: PAD‐ST (22 kDa); Lane 3: ST‐PAD‐ their native form. Lane 1: PAD (20 kDa), Lane 2: PAD-ST (22 kDa); Lane 3: ST-PAD-ST (24 kDa); Lane ST (24 kDa); Lane 4: PAD‐SC (32 kDa); Lane 5: PageRuler Prestained Protein Ladder (Thermo 4: PAD-SC (32 kDa); Lane 5: PageRuler Prestained Protein Ladder (Thermo Scientific). (b) Enzymatic Scientific). (b) Enzymatic activity at 25 °C in μmolpCA/(μmolPAD min) of the four differently tagged activity at 25 ◦C in µmolpCA/(µmolPAD min) of the four differently tagged PAD variants against 1 mM PAD variants against 1 mM p‐coumaric acid (pCA) as substrate, determined by an absorbance based p-coumaric acid (pCA) as substrate, determined by an absorbance based depletion assay. depletion assay. Since the PAD enzyme naturally occurs as a homodimer we generated variants with one (PAD-ST) Since the PAD enzyme naturally occurs as a homodimer we generated variants with one (PAD‐ or two SpyTags (ST-PAD-ST) in order to study self-assembly properties with the complementary ST) or two SpyTags (ST‐PAD‐ST) in order to study self‐assembly properties with the complementary partner containing one SpyCatcher (PAD-SC) per monomeric subunit. This strategy should allow us to partner containing one SpyCatcher (PAD‐SC) per monomeric subunit. This strategy should allow us vary theto vary degree the ofdegree crosslinking of crosslinking and thus and mesh thus size mesh of thesize hydrogel of the hydrogel by adjusting by adjusting the stoichiometric the stoichiometric molar ratiosmolar of the ratios different of the enzyme different variants, enzyme as variants, illustrated as illustrated in Figure3 a.in SeveralFigure 3a. types Several of PAD-hydrogels types of PAD‐ (H1–H4,hydrogels Figure (H1–H4,3) were produced.Figure 3) were For produced. type H1 hydrogel, For type onlyH1 hydrogel, PAD-SC only and ST-PAD-STPAD‐SC and were ST‐PAD mixed‐ST in a 2:1were molar mixed ratio, in a 2:1 which molar should ratio, resultwhich inshould the highest result in possible the highest degree possible of crosslinking. degree of crosslinking. For type H2, preparedFor type withH2, prepared a ratio of 4:2:1with ofa ratio PAD-SC:PAD-ST:ST-PAD-ST, of 4:2:1 of PAD‐SC:PAD a‐ST:ST lower‐PAD degree‐ST, of a crosslinkinglower degree was of crosslinking was expected. Likewise, type H3, formed from a molar ratio of 6:4:1 of PAD‐SC:PAD‐ ST:ST‐PAD‐ST, should display a further reduced degree of crosslinking. In type H4, where no double‐

Micromachines 2019, 10, x 7 of 13

tagged ST‐PAD‐ST was used and the PAD‐SC:PAD‐ST ratio was adjusted to 1:1, linear chains of crosslinked proteins should be formed without any branching points. MicromachinesTo evaluate2019, 10 ,the 795 ST/SC coupling in these systems, the formation of the covalent isopeptide bond7 of 12 was monitored by SDS‐PAGE analysis (Figure 3b). Indeed, the expected dimeric and trimeric PAD conjugates were clearly formed. As expected, the ST‐PAD‐ST variant, which contains two expected. Likewise, type H3, formed from a molar ratio of 6:4:1 of PAD-SC:PAD-ST:ST-PAD-ST, should crosslinking handles, formed trimeric conjugates with monomers of the PAD‐SC, as indicated by the displaydominant a further band reducedof approximately degree of 87 crosslinking. kDa in lane InH1. type This H4, result where also no proves double-tagged the accessibility ST-PAD-ST of both was usedthe andN‐ as the well PAD-SC:PAD-ST as C‐terminal ratio SpyTag was peptide adjusted in to the 1:1, ST linear‐PAD chains‐ST construct. of crosslinked In contrast, proteins no shouldhigher be formedconjugates without were any observed branching when points. single‐tagged PAD‐ST was used (lane H4).

(a) (b)

FigureFigure 3. 3.(a ()a Di) Differentfferent types types of of PAD-HydrogelsPAD‐Hydrogels (H1–H4) (H1–H4) with with different different degrees degrees of of crosslinking crosslinking can can be be preparedprepared by by using using di differentfferent stoichiometries stoichiometries of the SC/ST SC/ST-tagged‐tagged PAD PAD variants. variants. The The various various different different hydrogelshydrogels are are depicted depicted here here as simplified as simplified and idealizedand idealized 2D-representations 2D‐representations with thewith individual the individual building blocksbuilding approximately blocks approximately drawn to scale.drawn The to scale. molar The ratios molar of ratios PAD-SC:PAD-ST:ST-PAD-ST of PAD‐SC:PAD‐ST:ST‐PAD were‐ST 2:0:1 were (H1), 4:2:12:0:1 (H2), (H1), 6:4:1 4:2:1 (H3), (H2), 1:1:0 6:4:1 (H4). (H3), For 1:1:0 each (H4). assembly, For each one assembly, repeating one unitrepeating within unit the within gel is highlighted.the gel is highlighted. (b) Denaturing 16% SDS‐PAGE analysis of the various types of hydrogels (H1–H4). The (b) Denaturing 16% SDS-PAGE analysis of the various types of hydrogels (H1–H4). The assembly assembly reaction was carried out for 30 min at 30 °C and 1000 rpm and was quenched by the addition reaction was carried out for 30 min at 30 ◦C and 1000 rpm and was quenched by the addition of 4 of 4× SDS‐PAGE loading dye and incubation at 95 °C for 10 min. × SDS-PAGE loading dye and incubation at 95 ◦C for 10 min.

ToMicrorheological evaluate the ST /experimentsSC coupling were in these carried systems, out to confirm the formation the polymeric of the covalent nature of isopeptide the hydrogels bond and to quantitatively determine the degree of crosslinking in the various H1–H4 materials in order was monitored by SDS-PAGE analysis (Figure3b). Indeed, the expected dimeric and trimeric PAD to identify the most stable gel for the subsequent flow experiments (Figure 4). All hydrogels were conjugates were clearly formed. As expected, the ST-PAD-ST variant, which contains two crosslinking obtained as monolithic, elastic materials with rheological properties comparable to hydrogels handles, formed trimeric conjugates with monomers of the PAD-SC, as indicated by the dominant prepared from different enzymes, as reported previously by our group [44,46]. Figure 4a shows the band of approximately 87 kDa in lane H1. This result also proves the accessibility of both the N- as variation of mean square‐displacements (MSDs) as a function of lag time for polystyrene particles wellwith as a C-terminal diameter of SpyTag 200 nm peptide dispersed in the in the ST-PAD-ST hydrogel construct. H1. All MSDs In contrast, are almost no higherindependent conjugates of time were observed(slope ≈ when0) for single-taggedtimes  < 1 s with PAD-ST a narrow was used distribution (lane H4). of absolute values characterized by a non‐ GaussianMicrorheological parameter α experiments 1. This result were is carried consistent out towith confirm an elastic the polymerictrapping of nature tracer of particles the hydrogels in a andhomogeneous to quantitatively gel‐like determine network. the Similar degree results of crosslinking were obtained in the for various all hydrogels H1–H4 investigated. materials in The order to identify the most stable gel for the subsequent flow experiments (Figure4). All2 hydrogels were calculation of the elastic shear modulus G0 using the relation G0 = 2kBT/3παΔr where kB is the obtainedBoltzmann as monolithic, constant, T elastic the temperature, materials with α the rheological tracer particle properties radius comparable and Δr2 the to time hydrogels‐independent prepared from different enzymes, as reported previously by our group [44,46]. Figure4a shows the variation of average MSD exhibits a gradual decrease of G0 from 33 ± 5 via 17 ± 2 and 16 ± 1 to 11 ± 1 Pa for H1, meanH2, square-displacementsH3 and H4, respectively. (MSDs) The higher as a function elasticity of obtained lag time for for H1 polystyrene is presumably particles due to with the fact a diameter that of 200 nm dispersed in the hydrogel H1. All MSDs are almost independent of time (slope 0) for H1 contains the highest amount of divalent ST‐PAD‐ST whereas H4 is built solely from fusion≈ timesproteinsτ < 1 with s with only a narrow two binding distribution valences, of absolutewhich suggests values characterizedin the latter case by athe non-Gaussian occurrence of parameter much α less1. crosslinking This result isevents. consistent Finally, with we an directly elastic determined trapping of the tracer mesh particles size  of the in a network homogeneous according gel-like to ≈ network.the classical Similar theory results of rubber were obtained elasticity for with all 𝐺 hydrogels and investigated. we found a significant The calculation  increase of the from elastic  shear modulus G using the relation G = 2k T/3πα∆r2 where k is the Boltzmann constant, T the 50 ± 3 nm for H1 0to 73 ± 3 nm for H4 (Figure0 4b),B which corresponds Bwith a decrease of the crosslinking r2 temperature,proportion.α Hence,the tracer the significantly particle radius larger and elasticity∆ the time-independent and smaller mesh averagesize of H1 MSD reflects exhibits the greater a gradual decrease of G from 33 5 via 17 2 and 16 1 to 11 1 Pa for H1, H2, H3 and H4, respectively. connectivity0 that originates± from ±the higher ±amounts ±of ST‐PAD‐ST providing four valences as The higher elasticity obtained for H1 is presumably due to the fact that H1 contains the highest amount of divalent ST-PAD-ST whereas H4 is built solely from fusion proteins with only two binding valences, which suggests in the latter case the occurrence of much less crosslinking events. Finally, we directly determined the mesh size ξ of the network according to the classical theory of rubber elasticity with G = kBT and we found a significant ξ increase from 50 3 nm for H1 to 73 3 nm for H4 (Figure4b), 0 ξ3 ± ± Micromachines 2019, 10, 795 8 of 12

whichMicromachines corresponds 2019, with 10, x a decrease of the crosslinking proportion. Hence, the significantly8 largerof 13 elasticityMicromachines and smaller 2019, 10, mesh x size of H1 reflects the greater connectivity that originates from the8 of higher 13 amountsbranching of ST-PAD-ST points. providingHydrogel fourH1, valencesas the most as branching tightly connected points. Hydrogel material, H1,was as selected the most as tightly the connectedbranchingformulation material, points. for was the Hydrogel subsequent selected asH1, experiments the as formulation the most in the tightly for flow the reactor.connected subsequent Altogether, material, experiments this was study inselected thealso flow illustrates as reactor.the that the rheological properties of the novel all‐enzyme hydrogels can be rationally modulated by Altogether,formulation this study for the also subsequent illustrates experiments that the rheological in the flow propertiesreactor. Altogether, of the novel this study all-enzyme also illustrates hydrogels thatadjusting the rheological the stoichiometric properties crosslinker of the novel ratio. all ‐enzyme hydrogels can be rationally modulated by can be rationally modulated by adjusting the stoichiometric crosslinker ratio. adjusting the stoichiometric crosslinker ratio.

(a) (b) (a) (b) FigureFigure 4. Multiple 4. Multiple particle particle tracking tracking (MPT) (MPT) microrheology microrheology analysis analysis of of locallocal viscoelastic properties properties of of differentFiguredifferent types 4. Multiple oftypes PAD of hydrogels particlePAD hydrogels tracking (H1–H4, (MPT)(H1–H4, specified microrheology specified in Figure in 3analysisFigure). ( a) Mean3). of (locala) squareMean viscoelastic square displacements displacementsproperties (MSDs) of (blackdifferent(MSDs) curves) (blacktypes of individual ofcurves) PAD fluorescentofhydrogels individual (H1–H4, marker fluorescent beadsspecified marker of diameterin Figurebeads 200of3). diameter ( nma) Mean dispersed 200 square nm in disperseddisplacements the hydrogel in the H1. hydrogel H1. The red curve is the ensemble‐average MSD. (b) Comparison between mesh size values The red(MSDs) curve (black is the curves) ensemble-average of individual MSD.fluorescent (b) Comparison marker beads between of diameter mesh 200 size nm values dispersedξ as deducedin the  as deduced from MPT experiments for the different hydrogels. All experiments were carried out in fromhydrogel MPT experiments H1. The red for curve the is di thefferent ensemble hydrogels.‐average All MSD. experiments (b) Comparison were between carried mesh out insize duplicates. values  duplicates.as deduced Note from that MPT the experiments experimentally for the determined different hydrogels.mesh correlate All experiments with the fraction were ofcarried crosslinking out in Note that the experimentally determined mesh correlate with the fraction of crosslinking proposed for duplicates.proposed forNote hydrogel that the samples experimentally H1–H4. determined mesh correlate with the fraction of crosslinking hydrogel samples H1–H4. proposed for hydrogel samples H1–H4. To shed light on the assembly kinetics, the ST/SC‐mediated gelation of PAD‐SC with the double‐ To shed light on the assembly kinetics, the ST/SC-mediated gelation of PAD-SC with the taggedTo shedST‐PAD light‐ST on was the assemblyfurther investigated kinetics, the by ST/SC SDS‐PAGEmediated analysis gelation (Figure of PAD 5). ‐TheSC with results the indicated double‐ double-taggedtaggedthat the ST coupling‐PAD ST-PAD-ST‐ST reaction was further was is almost further investigated completed investigated by SDSwithin‐ byPAGE the SDS-PAGE first analysis few minutes(Figure analysis 5).after (Figure The mixing. results5). In Theindicated contrast, results indicatedthatno polymerizationthe that coupling the coupling reaction occurred reaction is almost even after is completed almost prolonged completed within incubation the first within timesfew the minutes in first control fewafter reactions minutes mixing. with In after contrast, variants mixing. In contrast,nolacking polymerization no complementary polymerization occurred binding even occurred aftersites prolonged(lane even 10 after in incubationFigure prolonged 5). Hence,times incubation in these control results times reactions clearly in control with confirm variants reactions and withlacking variantsemphasize complementary lacking the excellent complementary binding coupling sites binding capabilities(lane 10 sites in Figure of (lane the 5). 10 PAD Hence, in Figure variants, these5). results Hence, thereby clearly these emphasizing resultsconfirm clearly and the confirmemphasizespecificity and emphasize andthe highexcellent thereaction excellent coupling rates couplingof capabilities the ST/SC capabilities system of the [47]. PAD of the variants, PAD variants, thereby thereby emphasizing emphasizing the the specificityspecificity and and high reaction reaction rates rates of of the the ST/SC ST/SC system system [47]. [47 ].

Figure 5. Kinetic analysis of the ST/SC‐conjugation of ST‐PAD‐ST and PAD‐SC analyzed by coomassie FigureFigurestained 5. Kinetic 5. 16%Kinetic analysisSDS analysis‐PAGE. of theof Stoichiometric the ST ST/SC/SC-conjugation‐conjugation amounts of (30of ST-PAD-STST pmol)‐PAD of‐ST both and and enzymes PAD PAD-SC‐SC were analyzed analyzed incubated by by coomassie coomassiefor up to stainedstained2 16%h at SDS-PAGE.16%30 °C. SDS The‐PAGE. reaction Stoichiometric Stoichiometric was stopped amounts amounts by (30the (30 pmol)addition pmol) of of bothof both4× enzymesSDS enzymes‐PAGE were were loading incubated incubated buffer for forand up up heat to to 2 h at 30 2C.inactivation, h Theat 30 reaction °C. Thewhich was reaction leads stopped wasto bydenaturation stopped the addition by theof of nonaddition 4 ‐covalentSDS-PAGE of 4× protein SDS loading‐PAGE interactions. bu loadingﬀer and Asbuffer heat expected, inactivation,and heat the ◦ × whichinactivation,covalent leads to bond denaturation which between leads ST ofto‐ PAD non-covalentdenaturation‐ST and PADof protein non‐SC‐covalent is interactions. rapidly protein formed As interactions. upon expected, mixing, As the expected,leading covalent to the bondthe betweencovalentformation ST-PAD-ST bond of dimeric between and PAD-SC and ST trimeric‐PAD is‐ rapidlyST conjugates and formedPAD with‐SC upon sizesis rapidly of mixing, 55 kDaformed leading and 87upon kDa, to themixing, respectively. formation leading Note of to dimeric thethat and trimericformationPAD‐SC conjugates (32of dimeric kDa) does with and not sizestrimeric react of 55conjugates with kDa PAD and with(20 87 kDa,kDa) sizes respectively.evenof 55 afterkDa anda prolonged Note 87 kDa, that respectively. incubation PAD-SC (32 ofNote kDa)120 that min does (lane 10). not reactPAD with‐SC (32 PAD kDa) (20 does kDa) not even react after with a prolongedPAD (20 kDa) incubation even after of a 120 prolonged min (lane incubation 10). of 120 min (lane 10).

Micromachines 2019, 10, 795 9 of 12

Micromachines 2019, 10, x 9 of 13 We then used the polymerization of the PAD variants to produce biocatalytic hydrogels that were mountedWe then into used miniaturized the polymerization flow reactors of the PAD in order variants to facilitate to produce the biocatalytic continuous hydrogels synthesis that of p-hydroxystyrenewere mounted (Figure into miniaturized6). To this end, flow syringes reactors containing in order a to 5 mMfacilitate solution the ofcontinuousp-coumaric synthesis acid were of connectedp‐hydroxystyrene to PDMS microreactors (Figure 6). To loaded this end, with syringes hydrogel containing H1 and thea 5 substratemM solution solution of p‐coumaric was perfused acid at a flowwere rate connected of 10 µ Ltomin PDMS1 with microreactors automated loaded sampling with of hydrogel the outflow H1 and (Figure the 6substratea). The microreactor solution was · − architectureperfused featured at a flow a reactorrate of volume10 μL∙min of− 1501 withµL automated with dimensions sampling of 3of the1 outflow54 mm. (Figure The PAD-based 6a). The × × all-enzymemicroreactor hydrogel architecture showed featured a stable a reactor and almost volume complete of 150 μ conversionL with dimensions for time of 3 periods × 1 × 54 ofmm.>10 The h (FigurePAD6b)‐based with only all‐enzyme traces of hydrogel the substrate showed remaining a stable inand the almost product complete samples, conversion as determined for time by reverseperiods phaseof HPLC >10 h (Figure (inset in 6b) Figure with only6b). traces The high of the and substrate constant remaining conversion in the rates product even samples, after longer as determined runtimes indicateby reverse that no phase significant HPLC enzyme(inset in Figure leaching 6b). takes The high place, and which constant would conversion otherwise rates lead even to after a constant longer runtimes indicate that no significant enzyme leaching takes place, which would otherwise lead to a1 decline in turnover. Without further optimization, a typical space time yield (STY) of 57.7 g (d L)− constant decline in turnover. Without further optimization, a typical space time yield (STY)· · of with conversions 98% was obtained. Hence, this initial proof-of-concept study clearly indicated 57.7 g∙(d∙L)−1 with≥ conversions ≥98% was obtained. Hence, this initial proof‐of‐concept study clearly that the novel PAD-based all-enzyme hydrogels hold a great potential for further optimization and indicated that the novel PAD‐based all‐enzyme hydrogels hold a great potential for further application in biocatalysis. optimization and application in biocatalysis.

(a) (b)

FigureFigure 6. Continuous 6. Continuous flow, flow, biocatalytic biocatalytic production production of of p-hydroxystyrene p‐hydroxystyrene using using PAD PAD hydrogel-loaded hydrogel‐loaded microreactors.microreactors. (a )(a Microfluidic) Microfluidic set-up set‐up used used in in this this study. study. (b(b)) TimeTime dependentdependent synthesissynthesis ofof p-hydroxystyrene,p‐hydroxystyrene, determined determined from from the the outflow outflow ofof the enzyme enzyme-loaded‐loaded reactor reactor that that was was perfused perfused with 5 mM substrate at a flowrate of 10 μL/min at 25 °C. The PDMS reactor had a total volume of 150 μL. with 5 mM substrate at a flowrate of 10 µL/min at 25 ◦C. The PDMS reactor had a total volume of 150 µL. The shownThe shown data data was was determined determined by reverse by reverse phase phase HPLC HPLC (inset) (inset) from from two two separate separate experiments. experiments.

4. Conclusions4. Conclusions In conclusion,In conclusion, we hereinwe herein describe describe a novel a novel material, material, a self-assemblinga self‐assembling all-enzyme all‐enzyme decarboxylase decarboxylase hydrogel,hydrogel, that that was was used used for for biocatalytic biocatalytic flow flow synthesis synthesis of ofp-hydroxystyrene, p‐hydroxystyrene, which which inin turnturn can be usedused for thefor the synthesis synthesis of of fine fine chemicals chemicals andand pharmaceuticals. We We here here demonstrate demonstrate for the for first the time, first time,that that a aphenolic phenolic acid acid decarboxylase decarboxylase from from EnterobacterEnterobacter sp. cansp. be can genetically be genetically fused with fused either with ST eitheror SC polypeptide domains without compromising the enzyme’s catalytic performance. We also ST or SC polypeptide domains without compromising the enzyme’s catalytic performance. We also demonstrate that the biocatalytic hydrogels can be adjusted in terms of their rheological properties demonstrate that the biocatalytic hydrogels can be adjusted in terms of their rheological properties and mesh size by simply varying the stoichiometric molar ratio of the interconnecting fusion partners. and mesh size by simply varying the stoichiometric molar ratio of the interconnecting fusion partners. Further in‐depth studies of the modular composition and the resulting changes in polymer properties Further in-depth studies of the modular composition and the resulting changes in polymer properties will be used to optimize the biocatalytic performance and compare it with other immobilization will bemethods. used to The optimize proof‐of the‐concept biocatalytic results performancepresented here and already compare clearly it with show other that immobilizationthe PAD‐based methods.hydrogels The are proof-of-concept excellently suited results for applications presented in here flow already biocatalysis. clearly This show enabled that us the to PAD-based generate a hydrogelsconversion are excellently ≥98% in a suitedstraight for channel applications reactor with in flow a volume biocatalysis. of 150 μ ThisL and enabled a space us‐time to generate yield of a conversion57.7 g∙(d∙L)98%−1. Therefore, in a straight our study channel does reactor not only with expand a volume the scope of 150of allµ‐Lenzyme and a hydrogels space-time by yield adding of 1 ≥ 57.7 ga (dnewL) −enzyme. Therefore, with high our study application does not relevance only expand to the thetoolbox. scope The of all-enzyme present work hydrogels also suggests by adding that · · a newour enzyme approach with holds high great application potential relevance for the toestablishment the toolbox. of The advanced present workmicrofluidic also suggests manufacturing that our approachprocesses holds that great include potential numbering for the establishmentup of chips to ofimprove advanced productivity microfluidic [46] manufacturingor the implementation processes of that includebiocatalytic numbering and chemoenzymatic up of chips to improve cascades productivity in machine [‐46assisted] or the compartmentalized implementation of biocatalyticproduction and chemoenzymaticprocesses [32]. cascades in machine-assisted compartmentalized production processes [32].

Micromachines 2019, 10, 795 10 of 12

Author Contributions: Conceptualization, E.M., S.G., C.M.N. and K.S.R.; Data curation, E.M., S.G., C.O., C.M.N. and K.S.R.; Formal analysis, E.M.; Funding acquisition, S.G., P.B., C.M.N. and K.S.R.; Investigation, E.M., S.G., and C.O.; Methodology, E.M., S.G., P.B. and C.O.; Project administration, C.M.N. and K.S.R.; Resources, N.W., C.M.N. and K.S.R.; Software, E.M., C.O. and N.W.; Supervision, C.M.N. and K.S.R.; Validation, E.M., C.O., C.M.N. and K.S.R.; Visualization, E.M.; Writing—original draft, E.M.; Writing—review & editing, S.G., P.B., C.O., C.M.N. and K.S.R. Funding: This work was supported by the Helmholtz programme “Bio-Interfaces in Technology and Medicine”. Sabrina Gallus and Patrick Bitterwolf thank “Fonds der Chemischen Industrie (FCI)” for a Kekulé fellowship. Acknowledgments: We thank Martin Peng for providing the expression plasmid for PAD-His and Anke Dech and Miriam Wolters for experimental help. Conﬂicts of Interest: The authors declare no conﬂict of interest.

References

1. Dawes, G.J.S.; Scott, E.L.; Le Notre, J.; Sanders, J.P.M.; Bitter, J.H. Deoxygenation of biobased molecules by decarboxylation and decarbonylation-A review on the role of heterogeneous, homogeneous and bio-catalysis. Green Chem. 2015, 17, 3231–3250. [CrossRef] 2. Cadot, S.; Rameau, N.; Mangematin, S.; Pinel, C.; Djakovitch, L. Preparation of functional styrenes from biosourced carboxylic acids by copper catalyzed decarboxylation in PEG. Green Chem. 2014, 16, 3089–3097. [CrossRef] 3. Robinson, A.M.; Hensley, J.E.; Medlin, J.W. Bifunctional Catalysts for Upgrading of Biomass-Derived Oxygenates: A Review. ACS Catal. 2016, 6, 5026–5043. [CrossRef] 4. Gunawardena, D.A.; Fernando, S.D. Methods and Applications of Deoxygenation for the Conversion of Biomass to Petrochemical Products. In Biomass Now-Cultivation and Utilization; Matovic, M.D., Ed.; InTech: Croatia, Yugoslavia, 2013. [CrossRef] 5. Xiao, X.; Bergstrom, H.; Saenger, R.; Johnson, B.; Sun, R.C.; Peterson, A. The role of oxygen vacancies in biomass deoxygenation by reducible zinc/zinc oxide catalysts. Catal. Sci. Technol. 2018, 8, 1819–1827. [CrossRef] 6. Ni, J.; Wu, Y.T.; Tao, F.; Peng, Y.; Xu, P. A Coenzyme-Free Biocatalyst for the Value-Added Utilization of Lignin-Derived Aromatics. J. Am. Chem. Soc. 2018, 140, 16001–16005. [CrossRef][PubMed] 7. Leisch, H.; Grosse, S.; Morley, K.; Abokitse, K.; Perrin, F.; Denault, J.; Lau, P.C.K. Chemicals from agricultural biomass: Chemoenzymatic approach for production of vinylphenols and polyvinylphenols from phenolic acids. Green Process. Synth. 2013, 2, 7–17. [CrossRef] 8. Morley, K.L.; Grosse, S.; Leisch, H.; Lau, P.C.K. Antioxidant canolol production from a renewable feedstock via an engineered decarboxylase. Green Chem. 2013, 15, 3312–3317. [CrossRef] 9. Zakzeski, J.; Bruijnincx, P.C.; Jongerius, A.L.; Weckhuysen, B.M. The catalytic valorization of lignin for the production of renewable chemicals. Chem. Rev. 2010, 110, 3552–3599. [CrossRef] 10. Liu, D.J.; Sun, J.; Simmons, B.A.; Singh, S. N-Heterocyclic Carbene Promoted Decarboxylation of Lignin-Derived Aromatic Acids. ACS Sustain. Chem. Eng. 2018, 6, 7232–7238. [CrossRef] 11. Tinikul, R.; Chenprakhon, P.; Maenpuen, S.; Chaiyen, P. Biotransformation of Plant-Derived Phenolic Acids. Biotechnol. J. 2018, 13, e1700632. [CrossRef] 12. Liu, J.; Liu, Q.; Yi, H.; Qin, C.; Bai, R.; Qi, X.; Lan, Y.; Lei, A. Visible-light-mediated decarboxylation/oxidative amidation of alpha-keto acids with amines under mild reaction conditions using O(2). Angew. Chem. Int. Ed. Engl. 2014, 53, 502–506. [CrossRef][PubMed] 13. Aleku, G.A.; Prause, C.; Bradshaw-Allen, R.T.; Plasch, K.; Glueck, S.M.; Bailey, S.S.; Payne, K.A.P.; Parker, D.A.; Faber, K.; Leys, D. Terminal Alkenes from Acrylic Acid Derivatives via Non-Oxidative Enzymatic Decarboxylation by Ferulic Acid Decarboxylases. ChemCatChem 2018, 10, 3736–3745. [CrossRef] [PubMed] 14. Payer, S.E.; Faber, K.; Glueck, S.M. Non-Oxidative Enzymatic (De)Carboxylation of (Hetero)Aromatics and Acrylic Acid Derivatives. Adv. Synth Catal. 2019, 361, 2402–2420. [CrossRef][PubMed] 15. Rosazza, J.P.; Huang, Z.; Dostal, L.; Volm, T.; Rousseau, B. Review: Biocatalytic transformations of ferulic acid: An abundant aromatic natural product. J. Ind. Microbiol. 1995, 15, 457–471. [CrossRef] 16. Ferré-Filmon, K.; Delaude, L.; Demonceau, A.; Noels, A.F. Catalytic methods for the synthesis of stilbenes with an emphasis on their phytoalexins. Coordin. Chem. Rev. 2004, 248, 2323–2336. [CrossRef] Micromachines 2019, 10, 795 11 of 12

17. Savio, M.; Ferraro, D.; Maccario, C.; Vaccarone, R.; Jensen, L.D.; Corana, F.; Mannucci, B.; Bianchi, L.;

Cao, Y.; Stivala, L.A. Resveratrol analogue 4,40-dihydroxy-trans-stilbene potently inhibits cancer invasion and metastasis. Sci. Rep. 2016, 6, 19973. [CrossRef] 18. Nasrullah, J.M.; Raja, S.; Vijayakumaran, K.; Dhamodharan, R. A practical route for the preparation of poly(4-hydroxystyrene), a useful photoresist material. J. Polym. Sci. Part A Polym. Chem. 2000, 38, 453–461. [CrossRef] 19. Landete, J.M.; Rodriguez, H.; Curiel, J.A.; de las Rivas, B.; Mancheno, J.M.; Munoz, R. Gene cloning, expression, and characterization of phenolic acid decarboxylase from Lactobacillus brevis RM84. J. Ind. Microbiol. Biotechnol. 2010, 37, 617–624. [CrossRef] 20. Rodriguez, H.; Landete, J.M.; Curiel, J.A.; de Las Rivas, B.; Mancheno, J.M.; Munoz, R. Characterization of the p-coumaric acid decarboxylase from Lactobacillus plantarum CECT 748(T). J. Agric. Food Chem. 2008, 56, 3068–3072. [CrossRef] 21. Sheng, X.; Lind, M.E.; Himo, F. Theoretical study of the reaction mechanism of phenolic acid decarboxylase. FEBS J. 2015, 282, 4703–4713. [CrossRef] 22. Rodriguez, H.; Angulo, I.; de Las Rivas, B.; Campillo, N.; Paez, J.A.; Munoz, R.; Mancheno, J.M. p-Coumaric acid decarboxylase from Lactobacillus plantarum: Structural insights into the active site and decarboxylation catalytic mechanism. Proteins 2010, 78, 1662–1676. [CrossRef][PubMed] 23. Gao, S.; Yu, H.N.; Wu, Y.F.; Liu, X.Y.; Cheng, A.X.; Lou, H.X. Cloning and functional characterization of a phenolic acid decarboxylase from the liverwort Conocephalum japonicum. Biochem. Biophys. Res. Commun. 2016, 481, 239–244. [CrossRef][PubMed] 24. Sheng, X.; Himo, F. Theoretical Study of Enzyme Promiscuity: Mechanisms of Hydration and Carboxylation Activities of Phenolic Acid Decarboxylase. ACS Catal. 2017, 7, 1733–1741. [CrossRef] 25. Gu, W.; Yang, J.; Lou, Z.; Liang, L.; Sun, Y.; Huang, J.; Li, X.; Cao, Y.; Meng, Z.; Zhang, K.Q. Structural basis of enzymatic activity for the ferulic acid decarboxylase (FADase) from Enterobacter sp. Px6-4. PLoS ONE 2011, 6, e16262. [CrossRef][PubMed] 26. Schweiger, A.K.; Rios-Lombardia, N.; Winkler, C.K.; Schmidt, S.; Moris, F.; Kroutil, W.; Gonzalez-Sabin, J.; Kourist, R. Using Deep Eutectic Solvents to Overcome Limited Substrate Solubility in the Enzymatic Decarboxylation of Bio-Based Phenolic Acids. ACS Sustain. Chem. Eng. 2019, 7, 16364–16370. [CrossRef] 27. Busto, E.; Simon, R.C.; Kroutil, W. Vinylation of Unprotected Phenols Using a Biocatalytic System. Angew. Chem. Int. Ed. Engl. 2015, 54, 10899–10902. [CrossRef][PubMed] 28. Pesci, L.; Baydar, M.; Glueck, S.; Faber, K.; Liese, A.; Kara, S. Development and Scaling-Up of the Fragrance Compound 4-Ethylguaiacol Synthesis via a Two-Step Chemo-Enzymatic Reaction Sequence. Org. Process Res. Dev. 2017, 21, 85–93. [CrossRef] 29. Gomez Baraibar, A.; Reichert, D.; Mugge, C.; Seger, S.; Groger, H.; Kourist, R. A One-Pot Cascade Reaction Combining an Encapsulated Decarboxylase with a Metathesis Catalyst for the Synthesis of Bio-Based Antioxidants. Angew. Chem. Int. Ed. Engl. 2016, 55, 14823–14827. [CrossRef] 30. Peng, M.; Mittmann, E.; Wenger, L.; Hubbuch, J.; Engqvist, M.K.M.; Niemeyer, C.M.; Rabe, K.S. 3D-Printed phenacrylate decarboxylase flow reactors for the chemoenzymatic synthesis of 4-hydroxystilbene. Chemistry 2019.[CrossRef] 31. Lindeque, R.M.; Woodley, J.M. Reactor Selection for Effective Continuous Biocatalytic Production of Pharmaceuticals. Catalysts 2019, 9, 262. [CrossRef] 32. Rabe, K.S.; Muller, J.; Skoupi, M.; Niemeyer, C.M. Cascades in Compartments: En Route to Machine-Assisted Biotechnology. Angew. Chem. Int. Ed. Engl. 2017, 56, 13574–13589. [CrossRef][PubMed] 33. Sheldon, R.A.; Pereira, P.C. Biocatalysis engineering: The big picture. Chem. Soc. Rev. 2017, 46, 2678–2691. [CrossRef][PubMed] 34. Tamborini, L.; Fernandes, P.; Paradisi, F.; Molinari, F. Flow Bioreactors as Complementary Tools for Biocatalytic Process Intensification. Trends Biotechnol. 2018, 36, 73–88. [CrossRef][PubMed] 35. Peschke, T.; Skoupi, M.; Burgahn, T.; Gallus, S.; Ahmed, I.; Rabe, K.S.; Niemeyer, C.M. Self-Immobilizing Fusion Enzymes for Compartmentalized Biocatalysis. ACS Catal. 2017, 7, 7866–7872. [CrossRef] 36. Woodley, J.M. New frontiers in biocatalysis for sustainable synthesis. Curr. Opin. Green Sustain. Chem. 2020, 21, 22–26. [CrossRef] 37. Britton, J.; Majumdar, S.; Weiss, G.A. Continuous flow biocatalysis. Chem. Soc. Rev. 2018, 47, 5891–5918. [CrossRef] Micromachines 2019, 10, 795 12 of 12

38. Zdarta, J.; Meyer, A.S.; Jesionowski, T.; Pinelo, M. Developments in support materials for immobilization of oxidoreductases: A comprehensive review. Adv. Colloid Interface Sci. 2018, 258, 1–20. [CrossRef] 39. Sheldon, R.A. CLEAs, Combi-CLEAs and ‘Smart’ Magnetic CLEAs: Biocatalysis in a Bio-Based Economy. Catalysts 2019, 9, 261. [CrossRef] 40. Cui, J.D.; Jia, S.R. Optimization protocols and improved strategies of cross-linked enzyme aggregates technology: Current development and future challenges. Crit. Rev. Biotechnol. 2015, 35, 15–28. [CrossRef] 41. Velasco-Lozano, S.; López-Gallego, F.; Mateos-Díaz, J.C.; Favela-Torres, E. Cross-linked enzyme aggregates (CLEA) in enzyme improvement—A review. Biocatalysis 2016, 1, 7626. [CrossRef] 42. Yan, E.K.; Cao, H.L.; Zhang, C.Y.; Lu, Q.Q.; Ye, Y.J.; He, J.; Huang, L.J.; Yin, D.C. Cross-linked protein crystals by glutaraldehyde and their applications. RSC Adv. 2015, 5, 26163–26174. [CrossRef] 43. Honda, T.; Miyazaki, M.; Nakamura, H.; Maeda, H. Facile preparation of an enzyme-immobilized microreactor using a cross-linking enzyme membrane on a microchannel surface. Adv. Synth. Catal. 2006, 348, 2163–2171. [CrossRef] 44. Peschke, T.; Bitterwolf, P.; Gallus, S.; Hu, Y.; Oelschlaeger, C.; Willenbacher, N.; Rabe, K.S.; Niemeyer, C.M. Self-Assembling All-Enzyme Hydrogels for Flow Biocatalysis. Angew. Chem. Int. Ed. Engl. 2018, 57, 17028–17032. [CrossRef][PubMed] 45. Peschke, T.; Bitterwolf, P.; Hansen, S.; Gasmi, J.; Rabe, K.S.; Niemeyer, C.M. Self-Immobilizing Biocatalysts Maximize Space-Time Yields in Flow Reactors. Catalysts 2019, 9, 164. [CrossRef] 46. Bitterwolf, P.; Gallus, S.; Peschke, T.; Mittmann, E.; Oelschlaeger, C.; Willenbacher, N.; Rabe, K.S.; Niemeyer, C.M. Valency engineering of monomeric enzymes for self-assembling biocatalytic hydrogels. Chem. Sci. 2019, 10, 9752–9757. [CrossRef] 47. Zakeri, B.; Fierer, J.O.; Celik, E.; Chittock, E.C.; Schwarz-Linek, U.; Moy, V.T.; Howarth, M. Peptide tag forming a rapid covalent bond to a protein, through engineering a bacterial adhesin. Proc. Natl. Acad. Sci. USA 2012, 109, E690–E697. [CrossRef] 48. Sutherland, A.R.; Alam, M.K.; Geyer, C.R. Post-translational Assembly of Protein Parts into Complex Devices by Using SpyTag/SpyCatcher Protein Ligase. Chembiochem 2019, 20, 319–328. [CrossRef] 49. Hatlem, D.; Trunk, T.; Linke, D.; Leo, J.C. Catching a SPY: Using the SpyCatcher-SpyTag and Related Systems for Labeling and Localizing Bacterial Proteins. Int. J. Mol. Sci. 2019, 20, 2129. [CrossRef] 50. Peschke, T.; Rabe, K.S.; Niemeyer, C.M. Orthogonal Surface Tags for Whole-Cell Biocatalysis. Angew. Chem. Int. Ed. Engl. 2017, 56, 2183–2186. [CrossRef] 51. Gu, W.; Li, X.; Huang, J.; Duan, Y.; Meng, Z.; Zhang, K.Q.; Yang, J. Cloning, sequencing, and overexpression in Escherichia coli of the Enterobacter sp. Px6-4 gene for ferulic acid decarboxylase. Appl. Microbiol. Biotechnol. 2011, 89, 1797–1805. [CrossRef] 52. Demidoﬀ, F.C.; de Souza, F.P.; Netto, C.D. Synthesis of Stilbene-Quinone Hybrids through Heck Reactions in PEG-400. Synthesis-Stuttgart 2017, 49, 5217–5223. [CrossRef] 53. Gibson, D.G.; Young, L.; Chuang, R.Y.; Venter, J.C.; Hutchison, C.A., 3rd; Smith, H.O. Enzymatic assembly of DNA molecules up to several hundred kilobases. Nat. Methods 2009, 6, 343–345. [CrossRef][PubMed] 54. Crocker, J.C.; Grier, D.G. Methods of digital video microscopy for colloidal studies. J. Colloid Interface Sci. 1996, 179, 298–310. [CrossRef] 55. Van Hove, L. Correlations in Space and Time and Born Approximation Scattering in Systems of Interacting Particles. Phys. Rev. 1954, 95, 249–262. [CrossRef] 56. Weeks, E.R.; Crocker, J.C.; Levitt, A.C.; Schoﬁeld, A.; Weitz, D.A. Three-dimensional direct imaging of structural relaxation near the colloidal glass transition. Science 2000, 287, 627–631. [CrossRef]

© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).