This is the accepted version of the following article: Solé Ferré, Jordi et al. “Co-immobilization of P450 BM3 and glucose dehydrogenase on different supports for application as a self-sufficient oxidative biocatalyst” in Journal of Chemical Technology and Biotechnology, vol. 93 issue 7 (July 2018), which has been published in final form at DOI 10.1002/jctb.5770. This article may be used for non-commercial purposes in accordance with Wiley Terms and Conditions for Self-Archiving. 1 Co-immobilization of P450 BM3 and Glucose Dehydrogenase on different 2 supports for application as a self-sufficient oxidative biocatalyst 3 4Short title: Co-immobilization of P450 BM3 and Glucose dehydrogenase on different supports 5 6Jordi Solé1, Gloria Caminal2, Martin Schürmann3 and Gregorio Álvaro1, Marina Guillén1 7aBioprocess Engineering and Applied Biocatalysis Group, Department of Chemical Biological 8and Environmental Engineering, Universitat Autònoma de Barcelona, 08193 Bellaterra, Spain. 9bInstitut de Química Avançada de Catalunya (IQAC) CSIC, Jordi Girona 18-26, 08034, Barcelona, 10Spain 11cInnoSyn B.V., Urmonderbaan 22, 6167 RD Geleen, The Netherlands 12 13Corresponding author: 14Marina Guillén 15Department of Chemical, Biological and Environmental Engineering, Universitat Autònoma de 16Barcelona (UAB), 08193 Bellaterra, Catalonia, Spain 17Telephone: +34 93 581 4793 18Fax: +34 93 581 2013 19marina.guillen@uab.cat 20 21 1 1 2 22Abstract 23BACKGROUND. The oxy-functionalization of non-activated carbon bonds by the bacterial 24cytochrome P450 BM3 from Bacillus Megaterium, presents a promising field in biosynthesis 25and it has gained much interest in recent decades. Nevertheless, the need for the expensive 26cofactor NADPH, together with low operational stability of the enzyme have made the 27implementation of this biocatalyst unfeasible in most cases for industry. 28RESULTS. P450 BM3 and glucose dehydrogenase (GDH), as a cofactor regeneration enzyme, 29were successfully co-immobilized obtaining a bi-functional self-sufficient oxidative biocatalyst. 30Firstly, a broad screening on 13 different supports was carried out. Afterwards, five selected 31agaroses with three different functionalities (epoxy, amine and aldehyde) were studied and 32their immobilization processes optimized. Finally, P450 BM3 and GDH, were co-immobilized on 33those supports showing the best performance for P450 BM3 immobilization: Epoxy-agarose 34(Epoxy-agarose-UAB) presenting 83% and 20% retained activities respectively; AMINO-agarose 35presenting 28% and 25% and Lentikats® with which both enzymes retained 100% of the initial 36activity. Furthermore, the re-utilization of the self-sufficient immobilized derivatives was tested 37in 5 repeated cycles. 38CONCLUSIONS. P450 BM3 and GDH have been successfully immobilized on three supports and 39their re-usability has been tested in a model reaction. It represents a step-forward for future 40P450 BM3 industrial implementations. 41 42 43 44 3 2 4 45Introduction 46Cytochromes P450 (CYPs) are versatile monooxygenases able to hydroxylate non-activated 47carbon bonds with the only requirement of molecular oxygen and an electron donor. CYPs are 48found as: (i) double domain proteins, presenting a heme-containing oxidative part and a 49FAD-/FMN-containing reductase; as (ii) triple domain enzymes with a third Ferredoxin subunit 50and (iii) as fusion proteins with just one enzymatic unit 1. These enzymes have been in the 51focus of research in recent decades due to interest in their application as catalysts for the 52efficient production of fine chemicals, polymers, active pharmaceutical ingredients or 53nutritional supplements. They display a broad substrate range as well as the capability to 54catalyse a variety of oxidations including epoxidations, hydroxylation of aromatics, N-oxidation, 55deamination, dehalogenation and others 2. 56P450 BM3 (EC 1.14.14.1) from Bacillus megaterium was discovered in 1986 and excels among 57CYPs because it presents the highest turnover numbers (17000 min -1 for arachidonic acid) 3,4. In 58contrast with eukaryotic CYPs, P450 BM3 is a self-sufficient soluble protein that contains both 59oxidative and reductase domains in the same polypeptide chain. Its natural substrates are 60medium to long chain fatty acids, however, as with other CYPs, P450 BM3 is a promiscuous 61enzyme. It accepts fatty amides and alcohols, hydroxylated fatty acids and ω-oxo fatty acids. It 62also displays, not only hydroxylating activity, but olefin epoxidation, ring expansion, 63heteroatom oxidation and dealkylation, and dehydrogenation across C-O, C-N and C-C bonds, 64as well as carbon-carbon bond formation and cleavage 5. 65There are some limitations in the large-scale application of P450 BM3 such as, low operational 66stability, a dependence on the expensive NADPH electron donor and low retained activity in 67organic media 6. Thus, there have been few cases where it has been attempt to be used 7,8 68industrially . In order to tackle these bottlenecks, protein engineering, cofactor recycling 5 3 6 69systems, surface modifications, reaction media engineering or immobilization are the main 70strategies usually followed in biocatalytic processes. 71Regarding the low operational stability, immobilization has proven to be an efficient 72methodology for a broad type of oxidoreductases, not only for improving their stability but also 9,10 73for enhancing process metrics by biocatalyst recycling . In addition, immobilization can 74confer operational advantages such as the possibility to operate in continuous mode, reduction 11,12 75of foam formation, increased stability in organic solvents or simplified product purification . 76Immobilization of P450 BM3 has been pursued by other authors. For example, Maurer et al. 77tried to immobilize it on a variety of commercially available supports 13. They succeeded using 78the positively charged resins DEAE and SuperQ; and a Sol-Gel, but only the last one was found 79suitable for bioconversions. Axarli et al. immobilized a triple mutant onto Epoxy-sepharose, 80achieving retained activities of 81% and improved stabilities at 37°C 14. Weber et al. entrapped 81the heme-domain of the P450 BM3 into mesoporous molecular sieves (MCM-41 and SBA-15) 82and found a correlation between the activity observed from the derivative and pore diameter 8315. Furthermore, Zhao et al. immobilized a mutant on DEAE-650S, further entrapped into k- 84carrageenan together with catalase and zinc dust (Zn/Co(III)sep) which served as electron 85source. They could operate the reactor during 10 batch cycles with continued conversions 86above 80% 16. Other successful and more recent examples involved the fusion of the enzyme to 17–19 87linkers that enabled immobilization on solid surfaces . From all the works present in the 88literature, few are the cases where P450 BM3 has been successfully immobilized covalently for 89bioconversion purposes. Many of the above works focused on bio-sensing, in which low 20 90amounts of enzyme are required, or on the electrochemical study of the enzyme . For large 91scale applications, the attachment of large quantities of the enzyme to solid matrices by 92covalent bonds is a robust method that often confers novel properties to the enzyme and it 93does not present leaching. 7 4 8 94With regard to electron donors, there have been attempts to substitute the expensive NADPH 21–23 95co-substrate for less costly products, e.g. NADH or H2O2 by protein engineering . In a 96different way, electrochemical-driven catalysis has also proved successful to bypass the use of 97NADPH. In various works, the immobilization of P450s on electrode surfaces, that served as 24–27 98electrons sources, led to enzymatically catalysed product formation . Strategies for 99regenerating the expensive NADPH has been also studied both in enzymatic and non-enzymatic 28–30 100manners . The use of whole cell biocatalyst expressing P450 has been considered as a 101suitable and much cheaper strategy by taking advantage of the cellular metabolism for NADPH 102regeneration 31,32. However, the use of whole cells entails several drawbacks such as secondary 103undesired reactions or low reaction rates due to the substrate or product mass transfer 104limitation. Among the different strategies for cofactor regeneration, the utilization of 105secondary enzymes that make use of sacrificial substrates is one of the most attractive systems. 106The use of enzymatic NADPH regeneration systems in P450-catalyzed reactions has been 107reported by other authors. Different enzymes such as glucose dehydrogenase, formate 108dehydrogenase or phosphite dehydrogenase proved to be efficient and suitable biocatalysts for 109this purpose 13,33,34, 35. 110In this work, a bi-functional self-sufficient oxidative biocatalyst is obtained by the successful co- 111immobilization of P450 BM3 and a GDH from Thermoplasma acidophilum as cofactor- 112regenerating enzyme (Figure 1). Firstly, P450 immobilization was analysed prior to the selection 113of the best co-immobilization strategy since P450 is a less robust enzyme compared to GDH. 114Thus, deeper understanding of P450 BM3 immobilization by two different enzyme-support 115interactions (adsorption or covalent attachment) is presented. In this sense, different supports 116presenting a broad range of features were tested to get insight in the optimal immobilization 117procedure that should be followed. Finally, co-immobilization of P450 BM3 and GDH was 118performed by applying the best methodologies resulting for the P450 BM3. The encapsulation 9 5 10 119in polyvinyl alcohol particles (Lentikats®) was also studied for the confinement of both P450 120BM3 and GDH as an alternative to adsorption and covalent attachment strategies. 121The resulting-self-sufficient biocatalysts were tested in terms of operational stability in the 122hydroxylation of sodium laureate as a model reaction for proof-of-concept. 123 124Experimental 125Materials and supports 126Nicotinamide adenine dinucleotide phosphate sodium salts in its oxidized and reduced form 127(NADP+ disodium salt and NADPH tetrasodium salt) was purchased from BONTAC 128Bioengineering (Shenzhen, Guandong, China). All the other reagents were purchased from 129Sigma Aldrich® (St. Louis, MO, USA) and were of analytical grade if not stated elsewhere. 130The supports used in the screening section were donated by Purolite® Life Sciences (Bala 131Cynwyd, PA, USA) and their features are summarized in Table 2. 132The Epoxy-agarose-SIGMA was purchased from Sigma-Aldrich®. According to the supplier’s 133specifications it presented spherical shape, it had an activation degree of ≥ 20 µmol/mL and 12 134atoms spacer. 135Non-functionalized agarose (spherical beads Ø 50-150 µm) purchased from Agarose Bead 136Technologies® (ABT®, Madrid, Spain) was used for obtaining Epoxy-agarose-UAB by support 137functionalization carried out as described by Axarli et al. 14. The quantification of the epoxy 138groups present on the supports was done following the method described by Gupta 36. 139Amino functionalized agarose (AMINO-agarose) presenting amino ethyl groups and aldehyde 140functionalized agarose (GLYOXYL-agarose) with an extent of labelling of 40-60 µmol/mL both 11 6 12 141were also purchased from ABT®. All this information was always according to the supplier’s 142specifications. 143Lentikats® polymer was purchased from GeniaLab® (Braunschweig, Germany) and its 144composition consists on a mixture of polyvinyl alcohol (10%), polyethylene glycol (6%) and 145demineralized water (84%). 146Recombinant expression of P450 BM3 and Glucose dehydrogenase in E. coli 147P450 BM3 (CYP102A1 from Bacillus megaterium BM3) was recombinantly produced in 148Escherichia coli in a 20 L scale batch fermentation employing an E. coli K12 derivative and a 149pBAD/myc-HisC based expression vector. Glucose Dehydrogenase (GDH) from Thermoplasma 150acidophilum (GDH-Tac) was co-expressed from the same vector in a poly-cistronic arrangement 151of the GDH gene downstream of the P450 BM3 gene. A 500 mL pre-culture were used to 152inoculate 20 kg main culture medium with 100 µg/mL neomycin. The pre-culture was prepared 153in standard Luria-Bertani (LB) medium supplemented with 100 µg/mL neomycin. The 154fermentation was performed using terrific broth (TB) medium with glycerol. 2.5 h after 155inoculation of the fermenter as inducer, pre-sterilized L-arabinose was added to the fermenter 156to final concentration 0.02% (w/v). 24 hours after inoculation of the fermenter, the cell 157material was harvested by centrifugation. Cell free extract was prepared by adding 2 weight 158equivalents of 100 mM potassium phosphate (KP i) buffer (pH 7.0) to 1 weight equivalent of E. 159coli wet cells and sonication of this cell suspension with an ultrasound probe for 20 min with 160cooling on ice and centrifugation to remove the cell-debris. 161Total protein and enzyme content 162The cell lysate was pre-clarified by centrifugation (3220 g for 15 min.) and analysed accounting 163for total protein content by means of Bradford Protein Assay Kit (Thermo Fisher Scientific, 164Waltham, USA) using bovine serum albumin as standard 37. 13 7 14 165Enzyme content was assessed using the sodium dodecyl sulphate polyacrylamide gel 166electrophoresis (SDS-PAGE) (NuPage 12%, Invitrogen, USA) run in a Mini-PROTEAN II apparatus 167(BioRad, USA) following the protocol of Laemmli et al. 38. Low range protein markers were used 168for molecular weight determination. Gels were stained using Comassie G250 colloidal stain 169solution [34% (v/v) ethanol, 2% (v/v) H 3PO4, 17% (w/v) NH4SO4 and 0.066% Comassie G250] 170and the Image LABTM software (BioRad, USA) was used for image processing. 171Active P450 BM3 content determination 172P450 BM3 active form concentration was determined using the CO-differential spectra analysis 173described by Omura and Sato, using ε = 91 mM-1 cm-1 39,40. 174P450 BM3 activity measurement 175The activity of the P450 BM3 was measured spectrophotometrically (λ=340 nm) following the 176consumption of NADPH (ε= 6.22 mM-1 cm-1, 200 µM), using sodium dodecanoate (sodium 177laurate) as substrate (1.3 mM dissolved in sodium phosphate buffer 50 mM pH=7.5). 25 µL of 178enzyme sample were added to the total volume of the assay (500 µL). The basal consumption 179of NADPH by unspecific enzymes present in the lysate or/and by P450 uncoupling in absence of 180substrate was determined by this same test but avoiding the addition of the substrate and 181adding buffer instead. This consumption rate, that represents about 2-3%, is subtracted from 182the measurement with sodium dodecanoate. One unit of activity (U) is defined as the enzyme 183required to convert 1 µmol of NADPH per minute at those given conditions (30°C and pH = 7.5). 184The absorbance was recorded using a spectrophotometer Cary 50 Bio UV-visible (Palo Alto, 185USA). The equipment allows temperature control and stirring. Plastic micro cuvettes BRAND® 186UV (Sigma Aldrich®) were used for soluble enzyme characterization and 3.5 mL quartz cuvettes 187HELLMA® 100-QS (Hellma Analytics, Müllheim, Germany) with magnetic stirring, were used for 188the immobilized derivatives (suspension). In the first case a total volume of 500 µL was reached 189while in the second it was 2 mL. 15 8 16 190Glucose dehydrogenase activity measurement 191GDH activity is measured spectrophotometrically (λ=340 nm) following the formation of 192NADPH. Certain amount of sample (1/20 of the V T) was added to a solution containing D- 193Glucose (200 mM dissolved in sodium phosphate buffer 100 mM pH=8) and NADP + (400 µM). 194The same procedure explained above was applied for measuring the non-GDH formation of 195NADPH and the same consideration is taken to obtain the Us. Moreover, it was checked that if 196NADPH is used instead of NADP+, no consumption of cofactor by P450 is detected under the 197test conditions. The absorbance was also recorded using the aforementioned equipment and 198conditions. 199Immobilization metrics 200In every immobilization, a characterization was pursued in order to obtain the retained activity 201(equation 1) and immobilization yield (equation 2). Supernatant and suspension activities were 202analysed over time until a steady state was reached and in all cases the activity of a blank (no 203support) was also monitored to ensure that the enzyme activity was not affected by protocol’s 204conditions. A 10% w/v relation between the support and the total volume was always 205stablished and small amounts of enzyme were loaded (4-5 U of P450 BM3/g of support and 5-6 206U of GDH/g of support) so that diffusional limitations were minimized. 207 Retained activity ( % ) = Suspension activity ( Ug )−Supernatant activity( Ug ) x 100 Initial offered activity ( ) U g 208(Equation 1) ( ( g) ) U Supernatant activity 209 Immobilization yield ( % ) = 1− x 100 (Equation 2) Initial offered activity ( ) U g 17 9 18 210For the assessment of the loading capacity the supports were saturated with enzyme and the 211activity was calculated taking into account that diffusional limitations do not allow the actual 212visualization of the activity in the support (equation 3). 213 Loading capacity ( Ug )=(Offered initial activity ( Ug )−Supernatant activity ( Ug )) x Immobilization Retained activity (%) yield (% 214(Equation 3) 215Moreover, the amount of protein per gram of support was also calculated, analysing the initial 216protein content and subtracting from it, the concentration still remaining in the supernatant at 217the end of the process. In all cases protein concentration was determined by means of 218Bradford Protein Assay Kit, as already described. 219Immobilization of P450 BM3 onto Purolite® resins 220From Purolite®, a set of commercially available resins presenting a variety of matrices, pore 221sizes, linker lengths, and functional groups (Table 2) were tested for the immobilization of P450 222BM3. The conditions in which immobilizations were carried out in each case were those 223specified by the supplier’s protocols. In all cases, the relation of 10% w/v between support and 224total volume was maintained. The immobilizations were done at 25°C and mild agitation. The 225incubation times ranged from 0.5 to 5 hours depending on the speed at which the enzymes 226were attached. The epoxy containing methacrylates (ECR8204F, ECR8215F and ECR8285) were 227used in 50 mM sodium phosphate buffer pH 8 containing 0.5 M NaCl. The amino functionalized 228resins (ECR8309F, ECR8315F, ECR8409F and ECR8415F) were mixed with the enzyme in 50 mM 229sodium phosphate buffer pH 6. The supports made of a mixture of styrene and methacrylate 230(ECR1061M and ECR1030M) were utilized in 50mM sodium phosphate buffer pH 7.5; finally, 231immobilizations using the epoxy agaroses (Praesto 45, 65 and 90) were performed in 1 M 232potassium phosphate buffer pH 8. 233Immobilization of P450 BM3 onto Epoxy-agarose-SIGMA and Epoxy-agarose-UAB supports 19 10 20 234For the other epoxy-agaroses (Epoxy-agarose-SIGMA and Epoxy-agarose-UAB), the enzyme was 235dissolved in 1 M (pH 8) potassium phosphate buffer, mixed with the resin (10% w/v) and left at 23625°C with mild agitation. The Epoxy-agarose-UAB was incubated for 4 hours and Epoxy- 237agarose-SIGMA was incubated for 0.5 h. At the end of the process, in order to eliminate the 238unreacted epoxy groups, the samples were incubated with 0.2 M β-mercaptoethanol for at 239least 2 h at 25°C. 240Immobilization of P450 BM3 onto GLYOXYL-agarose support 241For the immobilization using GLYOXYL-agarose different pH’s were tested (6.5 – 8.5) using 50 242mM sodium phosphate buffers. The immobilizations lasted 6 hours and were done at 25°C and 243mild agitation. For the reduction of the Schiff base, two different reducing agents were studied: 244sodium borohydride (1 mg/mL) added at the end of the immobilization and sodium 245cyanoborohydride (0.05 mg/mL for characterization and 0.4 mg/mL for high loads of enzyme) 246added at the beginning of the immobilization. 247Immobilization of P450 BM3 onto AMINO-agarose support 248For AMINO-agarose, 50mM sodium phosphate buffer (pH 6) was used to dissolve the enzyme 249and mix it with the support. Coming up next, the enzyme was left to ionically adsorb onto the 250support for 0.5 h-1 h. After that, a 200mM stock solution of N-(3-Dimethylaminopropyl)-N'- 251ethylcarbodiimide (EDC) was prepared dissolving it in 50 mM phosphate buffer and adjusting to 252pH 6 with HCl. It was then added at different concentrations and incubated for 30 min. to 253activate the carboxylic groups and promote the covalent binding. The concentration of EDC was 254optimized in the characterization phase (1, 2, 3, 4, 5, 10, 15, 20 mM) and also when using high 255loads of enzyme (3, 6, 12, 30, 60 mM). Finally, 0.5 M NaCl was introduced to desorb all the 256protein attached non-covalently. The immobilization was carried out at 25°C and mild agitation. 257Co-immobilization of P450 BM3 and GDH in AMINO-agarose and Epoxy-agarose-UAB 21 11 22 258The conditions in which immobilizations were carried out were exactly the same as previously 259described for P450 BM3 for AMINO-agarose and Epoxy-agarose-UAB. The activity of the GDH 260present in the lysate was also monitored in this stage together with the P450 BM3. Retained 261activities and immobilization yields were obtained for GDH and P450 BM3. 262Co-Immobilization of P450 and GDH into polyvinyl alcohol lenses (Lentikats®) 263Co-immobilization of GDH and P450 BM3 by entrapment was carried out using an organic 264polymer commercialized as Lentikats®. The solid mixture was heated up to 95°C until it turned 265into a transparent liquid. Then it was cooled down to 40°C and 8 mL of the polymer were mixed 266with 2 mL of a suspension containing the desired amount of biocatalyst. Afterwards the 267mixture was poured on a petri dish and the lenses were printed using the LentiPrinter®. The 268resulting plates were left to dry at 25°C for 1 h until they lost 80% of its weight approximately. 269Once printed, the support adopted a lentil-like shape of 1-2 mm diameter and 200-300 µm 270wide where the proteins were entrapped inside. 271Re-usability of the biocatalyst for the hydroxylation of sodium laureate 272The utilization of the immobilized enzymes in repeated reaction cycles were studied following 273the consumption of sodium laurate. The reactor (V T=5 mL) contained 1.53 mM sodium laurate, 2740.2 mM NADP+ and 12.5 mM D-glucose. All dissolved in 50 mM sodium phosphate buffer pH 2757.5. The enzyme added was the minimum necessary to convert at least 90% of the initial 276substrate in 1hour. The reactors used were purchased from MultiSynTech (Witten, Germany) 277and consisted of 10mL plastic vessels with a porous plate at the bottom. It allowed the 278extraction of the whole reactor liquid content by filtering with this same porous plate without 279losing any fraction of the immobilized derivative. The reaction took place at 25°C and 1000 rpm 280using a Multi-Therm H5000-HC-E thermo-shaker (Benchmark Scientific Inc., Sayreville, NJ, 281USA). After each cycle, the immobilized enzymes were washed twice with sodium phosphate 23 12 24 282buffer 50 mM pH 7.5 then new reaction medium was added and substrate conversion was 283quantified. Experiments were carried out in duplicate. 284GC-FID analysis of sodium laurate 285Samples containing sodium laurate and its products were analysed using a 7890A gas 286chromatograph (Agilent Technologies, USA) equipped with a HP-INNOWAX 19095N-123 column 287(30 m, 0.53 mm, 1 µm, Agilent Technologies). Reaction samples were filtered (Ø 0.22µm) and 28850µL of hexanoic acid (2.8 mg/mL), that served as internal standard, were added to 300µL of 289sample. The column temperature started at 150 °C, increased to 240°C at 24 °C/min. and it was 290held at final temperature for 11 minutes. The injector temperature was kept at 300°C; for the 291flame ionization detector, the temperature was 320°C. Helium was used as a carrier gas at a 292flow rate of 8 mL/min. All analyses were carried out in duplicate. 293 294Results and discussion 295Characterization of the E. coli lysate containing P450 BM3 and GDH 296As already mentioned, P450s are enzymes that require a cofactor to perform the redox 297reactions they catalyse. One of the strategies widely used for overcoming the disadvantages of 298co-factor dependency is the use of an enzyme-coupled reaction for the regeneration of the 299expensive NADPH. In this work, GDH (EC 1.1.1.47) from the archaea Thermoplasma 300acidophilum was used as NADPH-recycling enzyme using glucose as sacrificial substrate (Figure 3011). This robust and stable tetrameric protein is reported to be an efficient enzyme for NADPH 302regeneration 34,35,41. 303In the present work, GDH was co-expressed with P450 BM3 in E. coli and both enzymes were 304present in the cell lysates utilized. 25 13 26 305The characterization of the lysate was performed regarding protein content, enzyme content 306(SDS-PAGE, Figure 2), activity and P450 BM3 active form concentration and the results are 307shown in Table 1. 308The P450 BM3 was expressed at a higher level than the GDH, however, under the conditions in 309which the activity tests were performed, GDH gave higher specific activity. Table 1, also shows 310the correlation between activity units (Us) and concentration (nmols) of active enzyme; for the 311P450 BM3 case, 2.85 U/nmol. 312Screening of supports for the immobilization of P450 BM3 – Purolite® 313The main objective of the present work was to obtain a bi-functional self-sufficient oxidative 314biocatalyst by co-immobilization of P450 BM3 and GDH. To accomplish this, supports and 315immobilization conditions that were found to be most suitable for the P450 BM3 enzyme were 316prioritised, since cytochrome is the key enzyme performing the target reaction and it has been 317described as an unfavourable enzyme for immobilizing. GDH, on the contrary, has been 318described as a robust and more stable enzyme so difficulties in its immobilisation were not 319anticipated 41. In fact, successful immobilization of GDH on different kinds of support has been 42 320reported by several authors . The results obtained for P450 BM3 immobilization, the 321materials tested and its features are summarized in Table 2. The supports purchased in the 322screening kits encompass a range of different particle sizes (45 µm to 710 µm), matrices 323(methacrylate, styrene and agarose), functional groups (amino, epoxy and aldehyde), enzyme- 324support interaction (covalent binding or hydrophobic adsorption) and linker lengths (2 to 18 325carbon atoms). All of these properties have been identified as critical for successful enzyme 326immobilization 12. 327As shown in Table 2, the immobilization yields obtained were higher than 90% in all cases, 328except for the supports that interact with the enzyme by hydrophobic adsorption: ECR1061M 329and ECR1030M. In both cases the immobilization yields only reached 21±6% and 33±1% 27 14 28 330respectively. On the other hand, no retained activities or very low values (<20%) were observed 331in all tested supports, except for Praesto 45, 65 and 90, all epoxy-agarose supports (Figure 3A). 332According to the titration performed, Praesto 45, 65 and 90 contained an epoxide 333concentration of 9, 6 and 8µmol per gram of support respectively and resulted in retained 334activities of 60-68% after 4h incubation. The maximum loading capacities of these epoxy- 335agarose resins ranged from 18 to 20 U/g and 7 to 8 mg protein/g of support. The 336immobilization profiles were very similar among the three supports and they did not show any 337significant correlation with the particle size or the initial epoxide concentration (Figure 3A). 338Regarding the methacrylate and styrene matrices, even though different ionic strengths and pH 339values were assessed, none of the tested supports showed retained activities above 20%. P450 340BM3 had adhered to the matrices in most cases, but did not remain active. Two methacrylate- 341based materials harbouring amino groups on six carbon chain linkers (ECR8409F and the 342ECR8415F) gave retained activities of 24±4% and 22±3% respectively when the enzyme was 343adsorbed onto the resins solely by ionic interactions, however once carbodiimide was added, 344to promote covalent bond formation, the activity was reduced to 6±3% and 18±1% 345respectively. Furthermore, following the supplier’s protocol, the four amino methacrylate 346based supports were further functionalized with glutaraldehyde prior to enzyme binding, but 347none resulted in successful immobilization either. 348It should be also mentioned that methacrylic matrices with the same functional groups than 349the epoxy-agarose support (ECR8204F, ECR8215F and ECR8285) did not lead to actively 350immobilized enzyme, as can be seen by the low retained activities obtained (0%, 2±1% and 3513±2% respectively). These results could indicate that for P450 immobilization the hydrophilicity 352of the matrix is a key factor. The higher hydrophobicity of the methacrylate/styrene matrices, 353together with the flexible nature of the enzyme and the presence of two active sites, are 354probably the causes behind the deactivation of the biocatalyst 43. This hypothesis would be also 29 15 30 355in accordance with the low immobilization yields and retained activities obtained with the 356styrene matrices (ECR1061M and ECR1030M). The immobilization mechanism of these 357matrices is also based on hydrophobic adsorption and both the immobilisation yield and 358retained activity are low. Other authors have also reported unsuccessful attempts to 359immobilize P450s by hydrophobic adsorption. Maurer et al. attempted to immobilize the P450 360BM3 on hydrophobic supports such as EP100, MP1000, phenylsepharose, octylsepharose and 361nutylsepharose but none of them presented final retained activity 13. 362Taking into account the obtained results in the screening which showed a strong effect of the 363hydrophilicity of the matrix on P450 immobilization, agarose was chosen to perform further 364studies. 365Immobilization of P450 BM3 onto different functionalized agaroses 366In addition to the epoxy resins already tested (Praesto 45, 65 and 90), two other agarose 367matrices with free epoxy groups and different epoxide concentrations were studied: a 368commercial lyophilized epoxy-agarose resin from Sigma-Aldrich® (Epoxy-agarose-SIGMA) and 369an in-house functionalized agarose resin (Epoxy-agarose-UAB) prepared as shown in the 370experimental section. In addition, an agarose based matrices containing aldehyde (GLYOXYL- 371agarose) and primary amine (AMINO-agarose) functional groups were tested. 372The Epoxy-agarose-SIGMA resin, according to the titration performed, contained an epoxide 373concentration of 110±7 µmol/g of support and gave an immobilization yield greater than 95% 374after 0.5 h (Figure 3B). However, a rapid decrease in activity of the suspended biocatalyst was 375noted during the immobilization procedure, indicating a strong deactivation of the enzyme 376after immobilization. Thus, after just half an hour, low retained activities (35±8%) were 377observed from this support. These results could indicate that the utilization of an epoxy- 378agarose with high epoxide concentration, such as the Epoxy-agarose-SIGMA, results in a 379progressive loss of activity over time due to the multipoint attachment of the enzyme to the 31 16 32 380support 44. The explanation can be found in the reactivity of the epoxy groups which differs 381over time: firstly, reacting with the amines present on the surface of the enzyme and secondly 382reacting with the carboxyl and thiol groups 45. At the end, if the incubation time is long enough, 383the protein can be attached by too many points and lose its whole catalytic activity. This issue 384can be stopped adding a blocking agent that removes the unreacted epoxy groups once the 385protein is immobilized. In Figure 3B, β-mercaptoethanol (0.2 M final concentration) was added 386after 0.5 h and the suspension was incubated for 3.5 h at 25°C. No significant loss of activity 387was observed during this period which validates the hypothesis of the multipoint attachment 388as the main reason for enzyme deactivation. However, due to its elevated cost (118 €/g) and 389low retained activities, compared to the Praesto supports, the Epoxy-agarose-SIGMA was 390discarded from further studies. 391On the other hand, the Epoxy-agarose-UAB displayed an epoxide concentration of 30±3 µmol/g 392of support and gave more than a 95% immobilization yield and 83±8% retained activity after 3930.5 h (Figure 3C), with a maximum loading capacity of 30±2 U/g of support (7.5±0.8 mg 394protein/g of support). A decrease in retained activity can be also seen, reaching 50±5% after 4h 395of incubation probably due to the excessive number of covalent bindings. In this case, as well 396as for Praesto supports (Figure 3A), the agarose contains a lower epoxide concentration 397compared to Epoxy-agarose-SIGMA. Therefore, the activity loss due to the multipoint 398attachment is lower for the Epoxy-agarose-UAB or even negligible for the Praesto compared to 399the Epoxy-agarose-SIGMA. Thus, when choosing the right support, a consensus should be 400found. Higher epoxide concentration may imply higher loading capacity as is seen when 401comparing Praesto (18 to 20 U/g of support) and Epoxy-agarose-UAB (30±2U/g of support); 402however, at the same time, higher epoxide concentration entails higher deactivation of the 403enzyme over time as it is the case comparing all three epoxy-agaroses: Praesto (32-40% activity 404lost), Epoxy-agarose-SIGMA (75% activity lost) and Epoxy-agarose-UAB (50% activity lost). In 405this sense, due to the loading capacity, the high retained activities reached and the low cost 33 17 34 406(2.5 €/g), the Epoxy-agarose-UAB was considered the best candidate for the immobilization of 407P450 BM3 among epoxy-agaroses. This study represents a deeper insight into the work of 408Axarli et al. 14 in the understanding of the immobilization of P450 BM3 on epoxy-agaroses and 409sets the bases for an optimal process. 410As mentioned before, other agarose-based supports with different functional groups were also 411tested such as GLYOXYL-agarose and AMINO-agarose. The GLYOXYL-agarose (aldehyde 412activated) immobilization must be pursued in alkali media 46, however, P450 BM3 is not stable 413above pH 8.5 (data not shown). After screening a range of immobilization pH’s (6.5 – 8.5), 7.5 414was the best performing one resulting in 50±2% retained activity after 6 h incubation. For the 415reduction of the Schiff base formed between the aldehyde of the resin and the amine of the 416enzyme, two reducing agents were tested: i) sodium borohydride (1 mg/mL) which is added at 417the end of the incubation and ii) sodium cyanoborohydride (0.05 mg/mL) which has slower 418kinetics and can be added at the beginning of the immobilization. Sodium borohydride resulted 419in a complete deactivation of the enzyme whilst the sodium cyanoborohydride strategy 420resulted in 36±0.1% of retained activity. Thus, sodium cyanoborohydride was selected as the 421most suitable reducing agent. Regarding the studies on maximum loading capacity, the values 422previously obtained in the characterization (36±0.1% retained activity) at low enzyme loadings 423were not extrapolated and a major inactivation occurred. This support has proved suitable for 424the immobilization of P450 BM3, nevertheless, since the covalent immobilization using high 425loading was not successful, it was discarded from further studies. 426Regarding AMINO-agarose, as for the other tested supports functionalized by amino groups 427(i.e. ECR8309F, ECR8315F, ECR8409F and ECR8415F) the immobilization is based on a three- 428stage process. In the first stage, the enzyme is adsorbed onto the support surface by ionic 429interaction of the amino groups of the support and the carboxyl groups of the enzyme, thus an 430ionic adsorption occurs. In the second stage, carbodiimide is added to the medium aiming to 35 18 36 431activate the carboxyl groups present on the enzyme making covalent bonding possible. 432Solutions with high ionic strength such as sodium chloride (0.5 M) can be finally added to 433desorb all the protein attached non-covalently (third stage). Ionic adsorption on AMINO- 434agarose was performed at pH 6 and gave a 58±6% retained activity (Figure 3D phase I). This 435value was 2.4-2.6 fold higher than the retained activities obtained during the ionic adsorption 436for the methacrylate based materials ECR8409F and ECR8415F harbouring the same functional 437group (24±4% and 22±3%), as previously discussed. The maximum loading capacity achieved 438was 112±2 U/g of support (50±4 mg protein/g of support) which represents the highest value 439among all the resins tested. Covalent binding of the amine and the carboxyl groups was carried 440out by adding EDC following ionic adsorption phase (Figure 3D phase II). Since carbodiimide 441can lead to enzyme deactivation, different concentrations and incubation times were studied to 442determine the optimal conditions for the covalent binding phase. These studies showed 443optimal values of 3 mM EDC and 30 min reaction time respectively. Following covalent 444attachment, the biocatalyst was incubated with 0.5 M NaCl for 2 h (Figure 3D, phase III) in 445order to increase the ionic strength of the media and release all the protein not adhered 446covalently. 28±1% of the initial activity offered was covalently bound to the agarose, which 447represents 30% less activity than the value obtained for the ionic adsorption phase (Figure 3D, 448phase I). On the other hand, compared to the methacrylate-based supports, the retained 449activity was 4.7-fold and 1.6-fold higher than the obtained with ECR8409F and ECR8415F 450respectively. Regarding the maximum load, 53±2 U/g of resin was reached (47±3 mg protein/g 451of support), representing a 52% decrease in immobilized activity compared to the obtained by 452ionic adsorption, likely due to enzyme deactivation caused by the carbodiimide. When testing 453for the maximum loading capacity the concentration of EDC was further optimized since more 454enzyme was added (Figure 4). It resulted in an optimal concentration between 6 and 12 mM. 455This second value was chosen due to the higher immobilization yield that it entailed. This 456support represents a step forward in comparison with the other tested agaroses and it is the 37 19 38 457first work reporting a successful covalent immobilization of P450 BM3 in an amino-based 458matrix. 459Other authors have worked with P450s and agarose based matrices before. In 1988 King et. al. 460obtained 100% immobilization yield and 66% retained activity using a cyanogen bormide- 461sepharose 4BCL and a P450 from Saccharomyces cerevisiae. Another example is the 462aforementioned case of Axarli et al. that immobilized the P450 BM3 on epoxy-agarose and 463obtained 81% retained activity 14. 464Co-immobilization of P450 BM3 and Glucose dehydrogenase 465As it has been already mentioned, for co-immobilization of both GDH and P450 BM3, it was 466considered to prioritize the supports and immobilization conditions that were found to be 467suitable for P450 BM3, i.e. Epoxy-agarose-UAB and AMINO-agarose. 468The immobilization of GDH using Epoxy-agarose-UAB resulted in poor affinity of the protein for 469the support with immobilization yields below 20%. However, of the enzyme that did adhere, a 470slight over-activation was observed compared to the same quantity of free enzyme. At the end 471of the incubation (4 h), retained activity of 20 ± 5% and a maximum loading capacity of 3±0.1 472U/g of support were achieved. Since the conditions were exactly the same as reported above 473for P450 BM3, the activity for the cytochrome was comparable and resulted in 50 ±5% retained 474activity and 30±2 U/g (7.5±0.8 mg protein/g of support). 475With AMINO-agarose, GDH presented high affinity for the support and it also showed an over- 476activation effect when adhered to the matrix. It had better tolerance to the presence of EDC 477compared with P450 BM3, with 10 mM being the optimal concentration for this enzyme 478(57±2% retained activity) in the characterization phase. Co-immobilization of both enzymes, 479initially using a low carbodiimide concentration of 3mM, resulted in a 25±4% retained activity 480for GDH and 28±1% for P450 BM3. Using a higher loading of EDC (12 mM) resulted in a loading 39 20 40 481capacity of 18 ± 1 U/g of support for GDH and 53 ± 2 U/g of support for P450 BM3. In summary, 482a self-sufficient oxidative biocatalyst was obtained. 42 483There are several examples in the literature of the immobilization of GHD . Regarding the 484immobilization on agarose based matrices, Anwar et al. immobilized a baker’s yeast GDH on 485cyanogen bromide-sepharose obtaining a retained activity of 3% 47. Persson et al. purified a 486GDH from Bacillus subtilis by immobilizing it on thiopropyl-sepharose and recovered up to 65% 487of the initial activity 48. However, as far as the authors know, this is the first time that GDH has 488been covalently co-immobilized with P450 BM3. 489Co-immobilization of P450 BM3 and GDH into polyvinyl alcohol particles - Entrapment 490Finally, as an alternative to immobilization by adsorption or by covalent attachment, an 491entrapment strategy was also investigated for co-immobilization. Both P450 BM3 and GDH 492were successfully embedded into Lentikats® (polyvinyl matrix) as described in the experimental 493section. The analysis undertaken, showed maximum retained activities of 100±0.1% for both 494P450 BM3 and GDH. 495In the sense of the loading capacity employed, the activity and concentration of protein in the 496polymer was 1.8±0.2 U/g of P450 BM3, 0.8±0.1 U/g of GDH and 1.1±0.1 mg protein/g of 497support respectively. It was not further increased due to operational issues. The capacity is 498limited by an aggregation effect observed when charging high concentrations of lysate with 499which the lenses form a compact mass that cannot be used in our reactions. This entails a 500serious drawback when looking to industrial feasibility because the enzyme should be 31 501previously purified or the specific activity of the support would remain too low . However, 502since the immobilization had been successful, albeit with low enzyme loading, and it is the first 503work that reports a successful entrapment of both P450 BM3 and GDH in Lentikats ®, it was 504decided to continue working with this immobilized derivative. 505Reusability of the biocatalyst using sodium laurate as substrate 41 21 42 506From all the supports tested, AMINO-agarose, Epoxy-agarose-UAB and Lentikats® were 507considered good candidates to be potential carriers in bio-catalysis applications. It is well- 508known that uncoupling effects on P450s produce reactive oxygen species that inactivate the 49 509enzyme (uncoupling) . This handicap is less relevant when working with native substrates 510such as sodium laureate and so this substrate was used as a model for studying the reusability 511of the self-sufficient biocatalyst. Firstly, the coupling efficiency of P450 towards this substrate 512was determined and then reusability studies were performed using the self-sufficient 513P450/GDH biocatalysts obtained with the aforementioned carriers. 514Reactions using sodium laureate (1.3 mM) and NADPH (0.2 mM) were performed, monitoring 515and comparing substrates consumption by means of GC-FID. A coupling efficiency of 90.4±8.3% 516was observed, which means that, 90.4% of the NADPH consumed was employed for sodium 517laurate hydroxylation, with the remainder directed towards uncoupled reactions. This substrate 518was considered adequate to assess the operational stability of the enzymes in repeated 519reaction cycles. 520Total turnover number (mols of substrate consumed/mols of P450 BM3) (TTN) was chosen as 521the metric to evaluate the yield of the self-sufficient biocatalyst in comparison with the soluble 522lysate. To assess it, the immobilized P450 BM3/GDH (Epoxy-agarose-UAB, AMINO-agarose and 523Lentikats®) biocatalysts were tested in sodium laurate reaction cycles of 1 hour each utilizing 524glucose as the sacrificial substrate and the cofactor in its oxidized form (NADP +) thus forcing the 525cofactor regeneration to start before the P450-catalyzed reaction. Moreover, NADP + starting 526concentration was 6.5-fold lower than the sodium laureate. The amount of P450 BM3 required 527to convert > 90% of sodium laurate in 1 hour was found to be 0.18-0.2 U/mL (0.063-0.07 528nmols/mL). GDH on the other hand, was always ensured to be present in excess so that it did 529not limit the hydroxylation reaction. The conversions obtained in each cycle for the 530immobilized systems are presented in Figure 5. The TTN for the soluble reaction was 43 22 44 53121,500±290 mols of sodium laurate consumed/mol of P450 BM3 in 1 hour of reaction. The 532immobilized derivatives allowed the re-utilisation of the self-sufficient biocatalyst boosting the 533TTN by 2.31-fold for the Epoxy-agarose-UAB, 2.98-fold for AMINO-agarose and 2.3-fold for the 534enzymes immobilized into Lentikats®. As shown in Figure 5, the conversion dropped in an 535exponential manner for epoxy-agarose (Figure 5A). On the contrary, with AMINO-agarose and 536Lentikats® the decay was more linear (Figure 5B and 5C). At the end, all three derivatives could 537be re-used successfully, thus demonstrating a significant operational stability. 538Some examples can be found in the literature about the reusability of P450s. Lee et al. 539immobilized a phasin fused P450 BM3 and re-used it 4 times with no loss of activity 17. Zhao et 540al. were able to use a P450 BM3 mutein together with catalase in 10 batch cycles with 541conversions up to 80% 16. Regarding GDH, different studies of re-usability have been reported. 542A recent example, Petrovicová et al. co-immobilized a ketoreductase and a GDH from Bacillus 543megaterium in polyvinyl alcohol particles and were able to use the derivate 18 times with 544minimal loss of activity 50. 545 546Conclusions 547P450 BM3 has gained much interest in recent years in biocatalysis for its oxy-functionalization 548characteristics, although, the implementation of this enzyme in industrial processes has lacked 549optimal metrics. Operational stability as well as its requirement of NADPH in stoichiometric 550amounts have been identified as the main bottlenecks. Aiming to overcome these two main 551drawbacks, P450 BM3 and GDH have been co-immobilized in the present work to obtain a 552robust bi-functional self-sufficient biocatalyst. 553P450 BM3 and GDH were covalently co-immobilized on epoxy-functionalized agarose and 554amino-functionalized agarose. Both enzymes were also successfully co-immobilized by 45 23 46 555entrapment in a polyvinyl alcohol/polyethylen-glycol matrix named Lentikats®. From the five 556epoxy-agaroses tested, the optimal support (Epoxy-agarose-UAB) presented an epoxide 557concentration of 30±3 µmol/g of support and retained activities of 83±8% for P450 BM3 and 55820±5% for GDH. On the other hand, AMINO-agarose showed retained activities of 28±1% for 559P450 and 25±4% for GDH. Finally, the entrapment of both enzymes in Lentikats ® lead to P450 560BM3 and GDH co-immobilization maintaining 100% of the activity initially loaded. 561P450 BM3 was also immobilized on aldehyde-functionalized agarose and presented 36±0.1% 562retained activity. However, this support did not allow high enzyme loads required for reaction 563testing, thus making this support unattractive for further studies. 564The operational stability of the obtained bi-functional self-sufficient biocatalysts (Epoxy- 565agarose-UAB, AMINO-agarose and Lentikats®) was analysed in terms of re-usability for sodium 566laureate oxidation. All three derivatives showed successful reuse in 5 cycles, entailing an 567increment in the TTN of 2.31-fold the Epoxy-agarose-UAB, 2.98-fold the AMINO-agarose and 5682.3-fold the Lentikats® compared to the soluble enzymes. 569The P450 BM3 wild type served as a model in the immobilization studies in this work. On the 570other hand, mutated P450 BM3s, usually required for oxidative reactions of non-natural 571substrates which entail high uncoupling effects, can follow the exact same principles reported 1,5,6,51 572here since the surface of the enzyme is usually unaffected . In this sense, immobilization 573can be an added value to the process, for these new enzymes, due to the many advantages 574that it confers. 575 576Acknowledgements 577The research for this work has received funding from the European Union project ROBOX 578(grant agreement n° 635734) under EU’s Horizon 2020 Programme Research and Innovation 47 24 48 579actions H2020-LEIT BIO-2014-1. This document reflects only the author's view and the Agency 580is not responsible for any use that may be made of the information it contains. The 581Department of Chemical, Biological and Environmental Engineering of Universitat Autònoma 582de Barcelona constitutes the Biochemical Engineering Unit of the Reference Network in 583Biotechnology and the research group 2017 SGR 1462, Generalitat de Catalunya. Authors also 584thank COST Action CM 1303–Systems Biocatalysis for financial support and Purolite Life 585Sciences® for the supports given. Jordi Solé acknowledges also UAB for funding his Ph.D. grant. 586 587References 588 1. Urlacher, V. B. & Girhard, M. Cytochrome P450 monooxygenases: An update on 589 perspectives for synthetic application. Trends Biotechnol. 30, 26–36 (2012). 5902. Guengerich, F. P. Common and Uncommon Cytochrome P450 Reactions Related to 591 Metabolism and Chemical Toxicity. Chem. Res. Toxicol. 14, 611–650 (2001). 5923. Narhi, L. O. & Fulco, A. J. Characterization of a catalytically self-sufficient 119,000-dalton 593 cytochrome P-450 monooxygenase induced by barbiturates in Bacillus megaterium. J. 594 Biol. Chem. 261, 7160–7169 (1986). 5954. Warman, A. J. et al. Flavocytochrome P450 BM3: an update on structure and 596 mechanism of a biotechnologically important enzyme. Biochem. Soc. Trans. 33, 747–753 597 (2005). 5985. Whitehouse, C. J. C., Bell, S. G. & Wong, L.-L. P450 BM3 (CYP102A1): connecting the dots. 599 Chem. Soc. Rev. 41, 1218–1260 (2012). 6006. Bernhardt, R. & Urlacher, V. B. Cytochromes P450 as promising catalysts for 601 biotechnological application: Chances and limitations. Appl. Microbiol. Biotechnol. 98, 49 25 50 602 6185–6203 (2014). 6037. Paddon, C. J. et al. High-level semi-synthetic production of the potent antimalarial 604 artemisinin. Nature 496, 528–532 (2013). 6058. Kaluzna, I. et al. Enabling Selective and Sustainable P450 Oxygenation Technology. 606 Production of 4-Hydroxy-??-isophorone on Kilogram Scale. Org. Process Res. Dev. 20, 607 814–819 (2016). 6089. Guzik, U. & Hupert-kocurek, K. Immobilization as a Strategy for Improving Enzyme 609 Properties-Application to Oxidoreductases. 8995–9018 (2014). 610 doi:10.3390/molecules19078995 61110. Tischer, W. & Wedekind, F. Immobilized Enzymes: Methods and Applications. Biocatal. 612 Discov. to Appl. 200, 95–126 (1999). 61311. DiCosimo, R., McAuliffe, J., Poulose, A. J. & Bohlmann, G. Industrial use of immobilized 614 enzymes. Chem. Soc. Rev. 42, 6437 (2013). 61512. Santos, J. C. S. D. et al. Importance of the Support Properties for Immobilization or 616 Purification of Enzymes. ChemCatChem 7, 2413–2432 (2015). 61713. Maurer, S. C., Schulze, H., Schmid, R. D. & Urlacher, V. Immobilisation of P450BM-3 and 618 an NADP(+) cofactor recycling system: Towards a technical application of heme- 619 containing monooxygenases in fine chemical synthesis. Adv. Synth. Catal. 345, 802–810 620 (2003). 62114. Axarli, I., Prigipaki, A. & Labrou, N. E. Engineering the substrate specificity of 622 cytochrome P450 CYP102A2 by directed evolution : production of an efficient enzyme 623 for bioconversion of fine chemicals. 22, 81–88 (2005). 62415. Weber, E. et al. Immobilization of P450 BM-3 monooxygenase on mesoporous 51 26 52 625 molecular sieves with different pore diameters. J. Mol. Catal. B Enzym. 64, 29–37 626 (2010). 62716. Zhao, L., Güven, G., Li, Y. & Schwaneberg, U. First steps towards a Zn/Co(III)sep-driven 628 P450 BM3 reactor. Appl. Microbiol. Biotechnol. 91, 989–999 (2011). 62917. Lee, J. H. et al. New platform for cytochrome P450 reaction combining in situ 630 immobilization on biopolymer. Bioconjug. Chem. 25, 2101–2104 (2014). 63118. Zernia, S. et al. Peptide-Mediated Specific Immobilization of Catalytically Active 632 Cytochrome P450 BM3 Variant. Bioconjug. Chem. 27, 1090–7 (2016). 63319. Tan, C. Y., Hirakawa, H., Suzuki, R., Haga, T. & Iwata, F. Immobilization of a Bacterial 634 Cytochrome P450 Monooxygenase System on a Solid Support. 1–6 (2016). 635 doi:10.1002/anie.201608033 63620. Ferrero, V. E. V et al. Protein and electrode engineering for the covalent immobilization 637 of P450 BMP on gold. Anal. Chem. 80, 8438–8446 (2008). 63821. Neeli, R., Roitel, O., Scrutton, N. S. & Munro, A. W. Switching pyridine nucleotide 639 specificity in P450 BM3: Mechanistic analysis of the W1046H and W1046A enzymes. J. 640 Biol. Chem. 280, 17634–17644 (2005). 64122. Cirino, P. C. & Arnold, F. H. A self-sufficient peroxide-driven hydroxylation biocatalyst. 642 Angew. Chemie - Int. Ed. 42, 3299–3301 (2003). 64323. Vidal-Limón, A., Águila, S., Ayala, M., Batista, C. V. & Vazquez-Duhalt, R. Peroxidase 644 activity stabilization of cytochrome P450BM3 by rational analysis of intramolecular 645 electron transfer. J. Inorg. Biochem. 122, 18–26 (2013). 64624. Sadeghi, S. J., Fantuzzi, A. & Gilardi, G. Breakthrough in P450 bioelectrochemistry and 647 future perspectives. Biochim. Biophys. Acta - Proteins Proteomics 1814, 237–248 (2011). 53 27 54 64825. Holtmann, D., Mangold, K. M. & Schrader, J. Entrapment of cytochrome P450 BM-3 in 649 polypyrrole for electrochemically- driven biocatalysis. Biotechnol. Lett. 31, 765–770 650 (2009). 65126. Sugihara, N., Ogoma, Y., Abe, K., Kondo, Y. & Akaike, T. Immobilization of cytochrome P- 652 450 and electrochemical control of its activity. Polym. Adv. Technol. 9, 307–313 (1998). 65327. Munge, B., Estavillo, C., Schenkman, J. B. & Rusling, J. F. Optimization of Electrochemical 654 and Peroxide-Driven Oxidation of Styrene with Ultrathin Polyion Films Containing 655 Cytochrome P450 cam and Myoglobin. 6032, 82–89 (2003). 65628. Chenault, H. K., Simon, E. S. & Whitesides, G. M. Cofactor regeneration for enzyme- 657 catalysed synthesis. Biotechnology and Genetic Engineering Reviews 6, 221–270 (1988). 65829. Hollmann, F., Hofstetter, K. & Schmid, A. Non-enzymatic regeneration of nicotinamide 659 and flavin cofactors for monooxygenase catalysis. Trends Biotechnol. 24, 163–171 660 (2006). 66130. Liu, W. & Wang, P. Cofactor regeneration for sustainable enzymatic biosynthesis. 662 Biotechnol. Adv. 25, 369–384 (2007). 66331. Lundemo, M. T. & Woodley, J. M. Guidelines for development and implementation of 664 biocatalytic P450 processes. Appl. Microbiol. Biotechnol. 99, 2465–2483 (2015). 66532. Schewe, H., Kaup, B. A. & Schrader, J. Improvement of P450BM-3whole-cell biocatalysis 666 by integrating heterologous cofactor regeneration combining glucose facilitator and 667 dehydrogenase in E. coli. Appl. Microbiol. Biotechnol. 78, 55–65 (2008). 66833. Beyer, N. et al. P450 BM3 fused to phosphite dehydrogenase allows phosphite-driven 669 selective oxidations. Appl. Microbiol. Biotechnol. (2016). doi:10.1007/s00253-016-7993- 670 7 55 28 56 67134. Lu, Y. & Mei, L. Co-expression of P450 BM3 and glucose dehydrogenase by recombinant 672 Escherichia coli and its application in an NADPH-dependent indigo production system. J. 673 Ind. Microbiol. Biotechnol. 34, 247–253 (2007). 67435. Smith, L. D., Budgen, N., Bungard, S. J., Danson, M. J. & Hough, D. W. Purification and 675 characterization of glucose dehydrogenase from the thermoacidophilic 676 archaebacterium Thermoplasma acidophilum. Biochem. J. 261, 973–7 (1989). 67736. Munishwar, N. G. Methods for Affinity Based Separations of Enzymes and Proteins. 120 67837. Bradford, M. M. A rapid and sensitive method for the quantitation of microgram 679 quantities of protein utilizing the principle of protein-dye binding. Anal. Biochem. 72, 680 248–254 (1976). 68138. Laemmli, U. K. Cleavage of structural proteins during the assembly of the head of 682 bacteriophage T4. Nature 227, 680–685 (1970). 68339. Omura, Tsuneo; Sato, R. The Carbon Monoxide-biding pigment of Liver Microsomes. J. 684 Biol. Chem. 239, 2370–2378 (1964). 68540. Guengerich, F. P., Martin, M. V, Sohl, C. D. & Cheng, Q. Measurement of cytochrome 686 P450 and NADPH – cytochrome P450 reductase. 4, 1245–1252 (2009). 68741. Weckbecker, A. & Hummel, W. Glucose Dehydrogenase for the Regeneration of NADPH 688 and NADH. Microb. Enzym. Biotransformations 17, 225–238 (2005). 68942. Srivastava, P. K. & Singh, S. Immobilization and Applications of Glucose-6-Phosphate 690 Dehydrogenase: a Review. Prep. Biochem. Biotechnol. 43, 376–384 (2013). 69143. Munro, A. W. et al. P450 BM3 : the very model of a modern flavocytochrome. 27, 250– 692 257 (2002). 69344. Álvaro, G., Fernández-Lafuente, R., Blanco, R. M. & Guisan, J. M. Immobilization- 57 29 58 694 Stabilization of Penicillin G Acylase from Escherichia Coli. 181–195 (1989). 69545. Katchalski-Katzir, E. & Kraemer, D. M. Eupergit® C, a carrier for immobilization of 696 enzymes of industrial potential. J. Mol. Catal. - B Enzym. 10, 157–176 (2000). 69746. Mateo, C. et al. Glyoxyl agarose: A fully inert and hydrophilic support for immobilization 698 and high stabilization of proteins. Enzyme Microb. Technol. 39, 274–280 (2006). 69947. Goheer, M. a, Gould, B. J. & Parke, D. V. Preparation of immobilized baker’s-yeast 700 glucose 6-phosphate dehydrogenase attached to modified sepharose and sephadex and 701 a comparison of the properties of these preparations with those of the soluble enzyme. 702 Biochem. J. 157, 289–294 (1976). 70348. Persson, M., Bülow, L. & Mosbach, K. Purification and site-specific immobilization of 704 genetically engineered glucose dehydrogenase on thiopropyl-sepharose. FEBS Lett. 270, 705 41–44 (1990). 70649. Karuzina, I. I. & Archakov, A. I. The oxidative inactivation of cytochrome P450 in 707 monooxygenase reactions. Free Radic. Biol. Med. 16, 73–97 (1994). 70850. Petroviˇ, T., Hegyi, Z., Smonou, I., Rosenberg, M. & Rebroš, M. Co-Immobilization of 709 Ketoreductase and Glucose Dehydrogenase. (2018). doi:10.3390/catal8040168 71051. Urlacher, V. B. & Girhard, M. Cytochrome P450 monooxygenases: perspectives for 711 synthetic application. Trends Biotechnol. 24, 324–330 (2006). 712 713 714 715 59 30 60 716TABLES [Protein] Specific Enzyme* P450 BM3 active [Enzyme] (mg Activity (mg form Enzyme (mg enzyme/mL protein/m (U/mg enzyme/mg concentration† lysate) L lysate) protein) protein) (µM) P450 2.8±0.6 0.52±0.01 23.8±0.6 44.9±2.1 BM3 45.4±0.8 GDH 3.6±0.4 0.14±0.01 6.64±0.16 - 717Table 1. Characterization of the cell free extract containing P450 BM3 and Glucose 718dehydrogenase. 719 720*Data obtained from SDS-PAGE analysis 721†Data obtained from CO difference spectrum assay 723 724 725 726 727 728 729 730 731 61 31 62 732Table 2. Main characteristics of the matrices tested and their results regarding immobilization and retained activity of P450 BM3. The criterion chosen to 733select a support for further studies sets a minimum for Immobilization at 90% and 25% for Retained activity. Functional Particle Immobilization Code group Matrix Interaction Pore diameter (Å) size Retained activity (%) Yield (%) (Linker) (µm) ECR8204F Epoxy Methacrylate Covalent 300-600 150-300 92±4 0 ECR8215F Epoxy Methacrylate Covalent 1200-1800 150-300 99±1 2±1 Amino ECR8309F Methacrylate Ionic/Covalent 600-1200 150-300 95±5 3±2 (C2) Amino ECR8315F Methacrylate Ionic/Covalent 1200-1800 150-300 97±2 4±2 (C2) Amino ECR8409F Methacrylate Ionic/Covalent 600-1200 150-300 100 6±3 (C6) Amino ECR8415F Methacrylate Ionic/Covalent 1200-1800 150-300 98±1 18±1 (C6) ECR8285 Epoxy (C4) Methacrylate Ionic/Covalent 400-600 300-710 100 3±2 None ECR8806F Methacrylate Hydrophobic 500-700 150-300 93±2 0 (C18) ECR1061M None Styrene/Methacrylic Hydrophobic 600-750 300-710 21±6 0 ECR1030M None Styrene/Methacrylic Hydrophobic 200-300 300-710 33±1 7±14 Praesto 45 Epoxy Agarose Covalent ND 45 92±6 63±1 Praesto 65 Epoxy Agarose Covalent ND 65 95±1 68±3 Praesto 90 Epoxy Agarose Covalent ND 90 95±1 60±11 63 32 64 734FIGURES 735 736Figure 1. Scheme of the bi-functional self-sufficient oxidative biocatalyst obtained by the co- 737immobilization of P450 BM3 from Bacillus megaterium and a GDH from Thermoplasma 738acidophilum as cofactor-regenerating enzyme. 739 740 741 742 743 744 745 746 747 748 749 750 65 33 66 751 752Figure 2. SDS-PAGE gel containing a lysate co-expressing P450 BM3 and GDH (left lane) and a 753molecular weight standard Bio-Rad Precision Plus (right lane). Numbers on both sides 754represent the molecular weight of the protein bands (kDa). 755 756 757 67 34 68 758 759Figure 3. P450 BM3 immobilization onto: A) Praesto 45 [discontinued line (-----)], 65 [dotted 760line (·····)] and 90 [combined dotted and discontinued line (-··-··-)] [Sodium Phosphate Buffer 1 761M pH 8] B) Epoxy-agarose-Sigma [Sodium Phosphate Buffer 1 M pH 8]; (I) the enzyme is bound 762to the support (II) β-mercaptoethanol 0.2 M final concentration is added to eliminate the 763unreacted epoxy groups, C) Epoxy-agarose-UAB [Sodium Phosphate Buffer 1 M pH 8] and D) 764AMINO-agarose [Sodium phosphate buffer 0.05 M pH 6]; (I) adsorption of the enzyme to the 765support (II) incubation with certain amount of carbodiimide (EDC) (III) incubation with 0.5 M 766NaCl and desorption of the unattached protein. The activity offered always ranged from 3.5 to 7676 U/g of resin. The activities of a blank (squares), the supernatant (rhombus) and the 768suspension (triangle) were continuously analysed. Error bars correspond to standard deviation 769(n=2). 770 69 35 70 771 772Figure 4. Optimization of the EDC concentration in AMINO-agarose experiments using high 773loads. The immobilization was carried out using sodium phosphate buffer 50 mM pH 6 and 774offering 92 U/g of resin. Retained activity (black dots) and immobilization yield (grey bars) are 775represented. Error bars correspond to standard deviation (n=2). 776 777 778 779 780 781 782 783 784 785 786 71 36 72 787 788Figure 5. Reaction cycles using both P450 BM3 and GDH immobilized onto: A) Epoxy-agarose- 789UAB (10 mg of agarose/mL), B) AMINO-agarose (10 mg of agarose/mL) and C) Lentikats® (50 790mg of Lentikats/mL). Reactions were carried on in 5 mL scale adding 0.18-0.2 U of total activity 73 37 74 791and using sodium phosphate buffer 50 mM pH 7.5, 1.3 mM sodium laurate, 0.2 mM NADP +, 79212.5 mM D-glucose, temperature control at 25°C and constant 1000rpm agitation using a 793Multi-Therm H5000-HC-E thermos shaker. Error bars correspond to standard deviation (n=2) 75 38 76